Like many of the readers of this group, I've spent a lot of time pondering
the popularity and evolution of lisp. I've also spent a lot of time
playing with the various implementations of lisp out there and I've come to
a few tentative conclusions:
1. Althouth Lisp users often criticize Scheme for it's overly conservative
and incomplete definition, in many respects Lisp now suffers from the same
incompleteness for modern application development. Many of the tools
needed to develop typical modern-day applications are non-standard among
implementations, crude and often only partially functional, or simply
unavailable. FFI, database connectivity, network protocol bindings, gui
toolkit bindings, xml tools etc. are fragmented and unstandardized among
all CL implementations.
2. Ironically, the diversity of CL implementations seems to exacerbate the
problem of extending CL. The implementations I've seen of some of the
above tools clearly expend a lot of effort trying to support them all, with
code which is a mess of implementation-specific switches and parallel FFI
code. I have to imagine that time spent dealing with cross-implementation
compatibility is time not spend adding robustness, functionality, or
documentation. The irony here, of course, is that in some ways it is the
very standardization of Lisp that allows for this kind of forking because
there is a reference standard upon which to base new and otherwise
divergent implementations.
3. As a result of this fragmentation, Lisp is actually a poor choice for
many important kinds of application development. More and more it is
pushed into the role of a central application intelligence interfaced with
more mainstream applications which provide connectivity to the rest of the
world. Only its superb core design has allowed to retain even this
position. But for many other applications, the burden of maintaining this
kind of homogeneous system is too high and it really is easier to develop
in a more gregarious language.
To me, this is truly depressing. The more I study Lisp the more I
appreciate the genius of its design. As others have recently commented in
this group, mainstream computer hardware has finally evolved to the point
where all the old objections to Lisp are irrelevant and other popular
languages are much LESS efficient. In the distributed, networked world of
modern software development, Lisp should be king, but it languishes in
niche applications.
A survey of the handful of languages that are growing reveals some obvious
commonalities. Living languages today are either funded and marketed by
huge corporations with vast advertising and development budgets or are
open-source languages supported by volunteers. Since a Java or C# push
isn't likely for Lisp, the only alternative seems to be a strong
open-source development effort. My impression of the popular open-source
languages (Perl, Python, Ruby, Tcl and, to a lesser extent, Ocaml) is that
they all have the following in common:
1. A single implementation. This means any extensions or add-on modules
are developed only once and that their authors need only worry about
architecture-related portability issues. All developer effort goes toward
maintaining and improving a single common implementation.
In general, the authors of language extensions and external bindings are
less experienced than the gurus who develop the core language, so it's
crucial that the techniques for doing this are straightforward and
well-documented. I think that, in the long run, the success of an
open-source langauge depends on these people more than on the core
designers. Most of these people would run in terror if they saw how
difficult it is to write GTK bindings that work in CMUCL, CLISP, and ACL,
for example.
2. A simple and straightforward means of bootstrapping the language on any
sane platform with a c compiler. All of these languages are very portable.
3. No langauge standardization bureaucracy. The langauge designers are
free to evolve the langauge when necessary. These langauges have evolved
quickly in response to user requests and technological shifts.
Although there are high-quality free implementations of Lisp, the
open-source Lisp community has so fair failed to build any of the momentum
of these lesser langauges. I believe it's due to the failure to recognize
their strengths. I may be heading a new and fairly large-scale development
effort soon which will involve a fair amount of the sophisticated logic and
intelligence for which Lisp is such a superb tool, but I don't think I can
choose Lisp in good conscience for two reasons. First, because I know I
will waste a lot of time patching together code to interface to the rest of
the system and second because none of my developers are going to know Lisp.
Why? Because the kinds of programs on which most younger programmers cut
their teeth are a PITA to write in any open-source Lisp.
What can be done about this? I think it's imperative that the open-source
Lisp community should learn from the Pythonistas:
1. Pick a single open-source Lisp and focus all development effort on that
implementation. We have more than our share of brilliant implementors, but
splitting our efforts among CLISP, CMUCL and SBCL is lethal.
2. Do what it takes to make porting this implementation as simple as
possible. This implementation should compile and run on any new
posix-compliant platform with nothing more than a C compiler. It should
also be trivial to make executables with this Lisp and it should function
as an executable script interpreter with no extra gymnastics.
3. Take advantage of having a single reference implementation to define a
clean and straightforward CPAN-like system and a well-documented and
straightforward C FFI. I know there are efforts to define something like
this now, but, to my eyes, they are all severely limited by the difficulty
of supporting so many implementations.
4. Build full-featured and reliable interfaces to common toolkits and
protocols on this infrastructure. Lisp should talk to databases as
fluently as Perl, parse XML as well as Python, speak SNMP, HTTP, LDAP,
IMAP, and SOAP, and couple with C as easily as Ruby.
5. Take advantage of having a single implementation to extend and change,
when necessary, the Common Lisp standard. Obviously this should be done
extremely carefully, but I think it's a mistake at this point to consider
the CL spec to be either complete or sacrosanct.
Speaking as an open-source hacker that would love to do what he can to help
promote and grow lisp, I think these kinds of changes are necessary. As it
stands now, trying to get anything done in the current Babel is even more
frustrating than watching Ruby, Perl, and Python slowly reinvent Lisp the
hard way.
--
miles egan
·····@puzl.pixar.com (Miles Egan) writes:
>
> Like many of the readers of this group, I've spent a lot of time pondering
> the popularity and evolution of lisp. I've also spent a lot of time
> playing with the various implementations of lisp out there and I've come to
> a few tentative conclusions:
>
> 1. Althouth Lisp users often criticize Scheme for it's overly conservative
> and incomplete definition, in many respects Lisp now suffers from the same
> incompleteness for modern application development. Many of the tools
> needed to develop typical modern-day applications are non-standard among
> implementations, crude and often only partially functional, or simply
> unavailable. FFI, database connectivity, network protocol bindings, gui
> toolkit bindings, xml tools etc. are fragmented and unstandardized among
> all CL implementations.
While there's some truth to what you say here, I'm not convinced that
most of these need to be available through the implementation. I
think your conclusions here would come out at least a bit different if
you would separate things that "must come from the implementor" from
things that "could come from anywhere". I admit that as implementors
start to USE some of these things, it gets more blurred, but
"thankfully" (and I use the term with a kind of grim humor) they
mostly haven't. Still, FFI is primitive, I think. db connectivity might
or might not be. But network protocol bindings are not really primitive.
We just need a little foothold support (which could come through the ffi),
and libraries could be portable. Maybe also GUIs. Definitely XML tools
don't need to come from vendors. So let's keep those separated.
I mention this because I think one of the problems is that the Lisp community
as a whole is used to relying on the VENDORS for answers and they must,
as the Java community has, start relying on the USERS to provide some stuff,
leaning only on the vendors for what they cannot make themselves. There are
or ought to be more users than vendors, so it's the users who are empowered
to keep up. They must do their fair share and blame must be attributed
in proper proportions.
> 2. Ironically, the diversity of CL implementations seems to exacerbate the
> problem of extending CL. The implementations I've seen of some of the
> above tools clearly expend a lot of effort trying to support them all, with
> code which is a mess of implementation-specific switches and parallel FFI
> code. I have to imagine that time spent dealing with cross-implementation
> compatibility is time not spend adding robustness, functionality, or
> documentation. The irony here, of course, is that in some ways it is the
> very standardization of Lisp that allows for this kind of forking because
> there is a reference standard upon which to base new and otherwise
> divergent implementations.
Again I'd point to the users for not having indicated where they insist that
vendors agree. Users could demand standards for the foothold components,
and then lots more portable things could arise.
I don't disagree with the thrust of your observation, really, but I think
to be productive it must say who's guilty and what they must do to change.
I believe in pointing fingers on things like this. Not to be mean. Just to
create concrete dialog about who needs to do what.
> 3. As a result of this fragmentation, Lisp is actually a poor choice for
> many important kinds of application development.
This seems an overbroad conclusion. Most applications aren't developed for
deployment in all lisps initially, but in one. And in individual lisps the
support is there to do a lot.
> More and more it is
> pushed into the role of a central application intelligence interfaced with
> more mainstream applications which provide connectivity to the rest of the
> world. Only its superb core design has allowed to retain even this
> position. But for many other applications, the burden of maintaining this
> kind of homogeneous system is too high and it really is easier to develop
> in a more gregarious language.
This is personal taste and I doubt can be supported. I don't find it easier
in so uniform a fashion. It's apples and oranges, really.
> To me, this is truly depressing.
You've obviously thought hard about this issue. I would encourage you to
think harder about where to apply pressure and resources to fix it.
What's depressing is assuming it will remain static. There are enegetic,
interested people stillin the mix. Things can change. Not everyone can figure
out what to do to contribute. Leaders are called for.
> The more I study Lisp the more I
> appreciate the genius of its design. As others have recently commented in
> this group, mainstream computer hardware has finally evolved to the point
> where all the old objections to Lisp are irrelevant and other popular
> languages are much LESS efficient. In the distributed, networked world of
> modern software development, Lisp should be king, but it languishes in
> niche applications.
>
> A survey of the handful of languages that are growing reveals some obvious
> commonalities. Living languages today are either funded and marketed by
> huge corporations with vast advertising and development budgets or are
> open-source languages supported by volunteers. Since a Java or C# push
> isn't likely for Lisp, the only alternative seems to be a strong
> open-source development effort.
I really reject the notion that open-source is the answer to anyone's problems,
though I don't have any problem with someone doing it. (I'm assuming by open
source you mean like gnu style, and not where everything is still under
trade-secret-do-not-use source access.) Giving away code is fine, and might
bolster certain ailing sections of the community, but an economic model must
be created to sustain us. Saying "others are giving stuff away, we must too"
isn't an economic model. Where are Lisp people expected to make their money?
> My impression of the popular open-source
> languages (Perl, Python, Ruby, Tcl and, to a lesser extent, Ocaml) is that
> they all have the following in common:
>
> 1. A single implementation. This means any extensions or add-on modules
> are developed only once and that their authors need only worry about
> architecture-related portability issues. All developer effort goes toward
> maintaining and improving a single common implementation.
IMO, you're not going to get a single implementation even with open source.
Lispers LIKE having their own little playgrounds. They aren't going to agree
to a single core that is no longer theirs. There are already several
free implementations. There is no motion to merge them except probably
by outsiders who wish they would merge. But the people who maintain them
seem content with their separate little worlds.
> In general, the authors of language extensions and external bindings are
> less experienced than the gurus who develop the core language, so it's
> crucial that the techniques for doing this are straightforward and
> well-documented. I think that, in the long run, the success of an
> open-source langauge depends on these people more than on the core
> designers. Most of these people would run in terror if they saw how
> difficult it is to write GTK bindings that work in CMUCL, CLISP, and ACL,
> for example.
>
> 2. A simple and straightforward means of bootstrapping the language on any
> sane platform with a c compiler. All of these languages are very portable.
Most existing implementations are already built on C cores. That doesn't
mean they're always portable. I have always assumed that was because C
is not as portable as it claims. It relies on libraries that are often
different, plus other machine differences that are obscure and hard to
track down. I don't really know what to make of this point.
> 3. No langauge standardization bureaucracy. The langauge designers are
> free to evolve the langauge when necessary. These langauges have evolved
> quickly in response to user requests and technological shifts.
I'm not big on bureaucracy. I've quit J13 for this reason. I think it
moves too slowly.
You should not assume, however, that I don't believe in standards. I do.
Quick evolution is not the opposite of bureaucracy, though. Whether or
not you have lots of red tape, you can still want standards. I've been
giving thought to this issue but am not ready to offer a solution myself
on this. Still, I think there are solutions to be had.
Incidentally, nothing about the present standard keeps anyone from
experimenting, other than market insistence on standards compliance. To the
degree you think you can convince the market not to comply, there is no
actual obstacle now.
> Although there are high-quality free implementations of Lisp, the
> open-source Lisp community has so fair failed to build any of the momentum
> of these lesser langauges. I believe it's due to the failure to recognize
> their strengths. I may be heading a new and fairly large-scale development
> effort soon which will involve a fair amount of the sophisticated logic and
> intelligence for which Lisp is such a superb tool, but I don't think I can
> choose Lisp in good conscience for two reasons. First, because I know I
> will waste a lot of time patching together code to interface to the rest of
> the system and second because none of my developers are going to know Lisp.
> Why? Because the kinds of programs on which most younger programmers cut
> their teeth are a PITA to write in any open-source Lisp.
You can always amke yet another open implementation of Lisp besides the
many already there and try to get people to believe it's the one and only.
I doubt you will succeed.
> What can be done about this? I think it's imperative that the open-source
> Lisp community should learn from the Pythonistas:
>
> 1. Pick a single open-source Lisp and focus all development effort on that
> implementation. We have more than our share of brilliant implementors, but
> splitting our efforts among CLISP, CMUCL and SBCL is lethal.
I'll just sit back and watch on this. Maybe you'll win. I don't know.
And maybe if you do you'll have some leverage over the commercial vendors.
We'll see. I do wonder where you'll get your resources though.
I think a cheaper solution is for someone who cares to just make their
free implementation so compellingly better that no one wants to use the others.
It will take less "negotiation", and so will put more energy into the
actual work of making something good.
> 2. Do what it takes to make porting this implementation as simple as
> possible. This implementation should compile and run on any new
> posix-compliant platform with nothing more than a C compiler. It should
> also be trivial to make executables with this Lisp and it should function
> as an executable script interpreter with no extra gymnastics.
All you have to do is volunteer the work.
I don't see what's stopping you.
Unless you're trying to volunteer someone else's work... and that's the
problem with free efforts, IMO.
> 3. Take advantage of having a single reference implementation to define a
> clean and straightforward CPAN-like system and a well-documented and
> straightforward C FFI. I know there are efforts to define something like
> this now, but, to my eyes, they are all severely limited by the difficulty
> of supporting so many implementations.
What stops you in what exists now?
> 4. Build full-featured and reliable interfaces to common toolkits and
> protocols on this infrastructure. Lisp should talk to databases as
> fluently as Perl, parse XML as well as Python, speak SNMP, HTTP, LDAP,
> IMAP, and SOAP, and couple with C as easily as Ruby.
What stops you in what exists now?
> 5. Take advantage of having a single implementation to extend and change,
> when necessary, the Common Lisp standard.
Why? It's VERY expensive to change. It cost just shy of a half a
million dollars to produce the standard in costs that we can easily
measure, and maybe again as much in hidden costs. That's just cost of
publishing paper. I'd estimate any actual technical change to the standard
at all will cost tens or hundreds of thousands of dollars for the community
to deploy. To what end? What in the present standard keeps work from
being done? I don't think you can point to something, but you're welcome
to try.
> Obviously this should be done
> extremely carefully, but I think it's a mistake at this point to consider
> the CL spec to be either complete or sacrosanct.
Why? Is this a purist or a monetary argument? I don't consider it complete
or sacrosanct from any conceptual point of view. I have lists of things I'd
love to change. But I cannot justify monetarily to anyone who would have to
actually accept the change and update their system that anything could change
at all. I wouldn't want the guilt. Would you?
> Speaking as an open-source hacker that would love to do what he can to help
> promote and grow lisp, I think these kinds of changes are necessary. As it
> stands now, trying to get anything done in the current Babel is even more
> frustrating than watching Ruby, Perl, and Python slowly reinvent Lisp the
> hard way.
IMO, you're better off just making eganlisp as a new language,
dialect, or implementation, whatever you seeok to make, taking anying
from public sources that you like, and ignoring anything you don't
like, and working from there. Trying to convince a community to agree
in advance of an outcome to follow you is VERY hard. I'm happy to
watch. But I'm not putting my betting good money on success coming in
the form you suggest.
I think success, if it comes, will come from people accepting that not every
language is the same, and that we are a community which has many bases.
We should work to have a set of common libraries and stop fussing over
the base. CL is done. Leave it. More resources spent on that are a commitment
individuals can reasonably make, but not the community. Our community has
paid once for its base and cannot afford to pay again for it, hoping for
a magic bullet fix.
JMO. The mileage of others will doubtless vary.
Kent M Pitman <······@world.std.com> writes:
> While there's some truth to what you say here, I'm not convinced
> that most of these need to be available through the implementation.
> I think your conclusions here would come out at least a bit
> different if you would separate things that "must come from the
> implementor" from things that "could come from anywhere".
Agreed. I think that a whole LOT could be done if a FFI library could
be created that ran efficiently in all of the existing
implententations and still allowed for the particular features and
strengths of each implementation to be used. There is no need for
this to be developed BY vendors, but I think it would greatly benefit
from the support of vendors. It seems a BSD style license is
essential for such a library.
Another important step would be a standardization of packaging and
building tools, defsystem in particular. There are efforts underway
to improve an open defsystem implementation that is portable to all
implementations. Again, a BSD style license IMO is most appropriate.
For the record, I like the GPL, but like Stallman I can also recognize
when other licenses would be better.
> Again I'd point to the users for not having indicated where they
> insist that vendors agree. Users could demand standards for the
> foothold components, and then lots more portable things could arise.
I agree as well. It seems that getting users to agree on a common FFI
interface, or socket interface is much more difficult then getting
vendors to support them. In many cases the vendors are not needed at
all. Perhaps the only agreement needed from vendors is that they ship
the library with their implementations.
> I really reject the notion that open-source is the answer to
> anyone's problems, though I don't have any problem with someone
> doing it. (I'm assuming by open source you mean like gnu style, and
> not where everything is still under trade-secret-do-not-use source
> access.) Giving away code is fine, and might bolster certain ailing
> sections of the community, but an economic model must be created to
> sustain us. Saying "others are giving stuff away, we must too"
> isn't an economic model. Where are Lisp people expected to make
> their money?
From writing applications which build on the open libraries, just like
in the other langauges with open implementations. Is there really a
need for these basic services like FFI, sockets, XML tools and the
like to be proprietary? I don't think so. I think that their
presence would be an economic boon to users and vendors alike.
> IMO, you're not going to get a single implementation even with open source.
> Lispers LIKE having their own little playgrounds. They aren't going to agree
> to a single core that is no longer theirs. There are already several
> free implementations. There is no motion to merge them except probably
> by outsiders who wish they would merge. But the people who maintain them
> seem content with their separate little worlds.
I can't see a single implementation ever emerging as dominant either,
and thank god for that. I think it's important for the long-term
health of the language to have these different implementations, even
if it does present a standardization challenge.
> I think a cheaper solution is for someone who cares to just make
> their free implementation so compellingly better that no one wants
> to use the others. It will take less "negotiation", and so will put
> more energy into the actual work of making something good.
I think that this approach applied to the libraries we mentioned
previously is worth pursuing. However, I don't think it will ever
work for CL implementations. I can see getting people to agree on a
FFI library, or socket library, at least to the point of it becoming a
de-facto standard, but not CL implementations.
> Unless you're trying to volunteer someone else's work... and that's the
> problem with free efforts, IMO.
Oh come on Ken. Sure there are some people who clamor for people to
volounteer their work, but that can't be attributed to "free efforts".
In fact, "free efforts" are precisely those people who volounteer to
*DO* the work and release it to the public. What you're talking about
are not "free efforts" because there is no effort. Do you really mean
you confuse those people who are attempting to develop an open code
base and contributing their own labors with the hangers-on that have a
million ideas but no code?
Besides, what real harm is there in asking someone who has written
code to release it under an open license? Are you just lamenting the
lack of tact and grace that some people do it with?
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Revolution begins by giving things and social relationships
their real names". -- L. Trotsky
Craig Brozefsky <·····@red-bean.com> writes:
> Oh come on Ken.
[Kent]
> Sure there are some people who clamor for people to volounteer their
> work, but that can't be attributed to "free efforts".
Yes, it can. A huge number of people are coming out of college these
days thinking that free stuff is an entitlement and that there is
something wrong with a community that doesn't provide it to them.
> Do you really mean you confuse those people who are attempting to
> develop an open code base and contributing their own labors with the
> hangers-on that have a million ideas but no code?
Partly I'm saying I don't see you saying how you're in the set of people
doing something. The key to doing something is not writing posts on why
it's needed, it's simply doing. You don't have to convince anyone.
> Besides, what real harm is there in asking someone who has written
> code to release it under an open license?
I think there is considerable harm.
I have spent a good part of this year only partly employed, trying to
figure out how to make a business out of things I know for once, rather
than to automatically join up with some company and make money for it.
I needed exactly the tools you're talkign about, but I didn't expect anyone
to provide them to me, so I wrote my own and now I have a bunch of
intersting stuff. It bugs me is that there is not a market for people to
license such tools; I don't hear anone posting here saying they'd pay a dime
for any of this stuff. I'm sufficiently not-optimistic about the prospect of
making any money on licensing this stuff because everyone seems so fixated
on the idea that this stuff should be freebie-ized that I'm not even trying
and I'm just going straight through to developing commercial applications
based on my tools.
I think that if there were a culture here of people saying that code was
worth something, I would have (a) licensed the stuff I needed from someone
who had written what I needed and (b) been able to license the stuff I built
atop that stuff to someone else. Each of us, adding value, would be
reimbursed for what we created. But instead, we're all involved in a rat-race
where as we develop anything, someone else wants to develop the same thing
and give it away. That undercuts those of us who are trying to invest in
something of value.
I think teaching college kids that it's good to give stuff away causes
them not to understand teh value of what they do, and causes them to
economically injure those people who are producing that stuff for money.
I don't see that as doing any good for anyone except the people who are
sitting around waiting for the freebies to be produced, and the whole system
is fueled by people who apparently need no income. That's not a stable system.
I can't figure out where the fuel is.
It is this (non)cycle, not anything technical, that is most in danger of
driving me away from Lisp after this many years of having held to it.
Certainly it makes me not even the slightest bit interested in spending time
on the market of creating tools, which is what people claim they so desperately
need in order to write applications. Because the writing of tools is something
people say they don't "value" as a market. Value means "pay money" in my
book. And so that makes it less likely, not more likely, tools will get
created...
> Are you just lamenting the
> lack of tact and grace that some people do it with?
I am very much demoralized and upset at the fact that it is considered
so socially acceptable these days to get virtually militant about the
absence of free stuff and it is considered so socially unacceptable to ask
a fair wage for having created something of value.
On Mon, 21 May 2001, Kent M Pitman wrote:
> Craig Brozefsky <·····@red-bean.com> writes:
>
> > Oh come on Ken.
>
> [Kent]
>
> > Sure there are some people who clamor for people to volounteer their
> > work, but that can't be attributed to "free efforts".
>
> Yes, it can. A huge number of people are coming out of college these
> days thinking that free stuff is an entitlement and that there is
> something wrong with a community that doesn't provide it to them.
Are these the same people you abuse later for giving their own stuff away?
Seems like a pretty fair tit-for-tat to me.
> > Do you really mean you confuse those people who are attempting to
> > develop an open code base and contributing their own labors with the
> > hangers-on that have a million ideas but no code?
>
> Partly I'm saying I don't see you saying how you're in the set of people
> doing something. The key to doing something is not writing posts on why
> it's needed, it's simply doing. You don't have to convince anyone.
If you're talking about Craig instead of the OP, you need to open your
eyes. Craig and OnShore are doing some of the more interesting open
source Lisp application work.
> > Besides, what real harm is there in asking someone who has written
> > code to release it under an open license?
>
> I think there is considerable harm.
>
> I have spent a good part of this year only partly employed, trying to
> figure out how to make a business out of things I know for once, rather
> than to automatically join up with some company and make money for it.
> I needed exactly the tools you're talkign about, but I didn't expect anyone
> to provide them to me, so I wrote my own and now I have a bunch of
> intersting stuff. It bugs me is that there is not a market for people to
> license such tools; I don't hear anone posting here saying they'd pay a dime
> for any of this stuff. I'm sufficiently not-optimistic about the prospect of
> making any money on licensing this stuff because everyone seems so fixated
> on the idea that this stuff should be freebie-ized that I'm not even trying
> and I'm just going straight through to developing commercial applications
> based on my tools.
>
> I think that if there were a culture here of people saying that code was
> worth something, I would have (a) licensed the stuff I needed from someone
> who had written what I needed and (b) been able to license the stuff I built
> atop that stuff to someone else. Each of us, adding value, would be
> reimbursed for what we created. But instead, we're all involved in a rat-race
> where as we develop anything, someone else wants to develop the same thing
> and give it away. That undercuts those of us who are trying to invest in
> something of value.
Hey, you're right, I wouldn't pay a dime for your stuff. Why should I
trust it? It's had one pair of eyes looking at it, not thousands. As far
as I'm concerned, your putative tools have no value unless they've been
given a public airing.
If you're now motivated to make commercial apps using your tools I can't
understand why you wouldn't want to lighten your support burden and
improve the quality of your tools by opening them up.
Now, I'm being a bit facetious here -- I believe that you, KMP, probably
have produced something of value and I do understand why you'd never make
them open source -- but not very.
> I think teaching college kids that it's good to give stuff away causes
> them not to understand teh value of what they do, and causes them to
> economically injure those people who are producing that stuff for money.
> I don't see that as doing any good for anyone except the people who are
> sitting around waiting for the freebies to be produced, and the whole system
> is fueled by people who apparently need no income. That's not a stable system.
> I can't figure out where the fuel is.
Ask Eric Benson about it sometime.
> It is this (non)cycle, not anything technical, that is most in danger of
> driving me away from Lisp after this many years of having held to it.
>
> Certainly it makes me not even the slightest bit interested in spending time
> on the market of creating tools, which is what people claim they so desperately
> need in order to write applications. Because the writing of tools is something
> people say they don't "value" as a market. Value means "pay money" in my
> book. And so that makes it less likely, not more likely, tools will get
> created...
>
gcc must be a figment of my imagination.
> > Are you just lamenting the
> > lack of tact and grace that some people do it with?
>
> I am very much demoralized and upset at the fact that it is considered
> so socially acceptable these days to get virtually militant about the
> absence of free stuff and it is considered so socially unacceptable to ask
> a fair wage for having created something of value.
I don't believe "socially unacceptable" is quite accurate unless you like
to hang out in certain IRC channels, which I doubt. "Economically
unacceptable" maybe. The fact that you can't think of a business model to
do what you'd generally like, despite the examples of others who have, is
hardly a chilling indictment of open source.
Tim
Tim Moore <·····@herschel.bricoworks.com> writes:
> On Mon, 21 May 2001, Kent M Pitman wrote:
>
> > Craig Brozefsky <·····@red-bean.com> writes:
> >
> > > Oh come on Ken.
> >
> > [Kent]
> >
> > > Sure there are some people who clamor for people to volounteer their
> > > work, but that can't be attributed to "free efforts".
> >
> > Yes, it can. A huge number of people are coming out of college these
> > days thinking that free stuff is an entitlement and that there is
> > something wrong with a community that doesn't provide it to them.
>
> Are these the same people you abuse later for giving their own stuff away?
>
> Seems like a pretty fair tit-for-tat to me.
Yes obviously and from what do they live?
>
> Hey, you're right, I wouldn't pay a dime for your stuff. Why should I
> trust it? It's had one pair of eyes looking at it, not thousands. As far
> as I'm concerned, your putative tools have no value unless they've been
> given a public airing.
You probably would not pay a dime for anything. That's you choice. Why
should you trust it? Now why shouldn't you? Do you think any vendor
will want to do you a harm?
>
> If you're now motivated to make commercial apps using your tools I can't
> understand why you wouldn't want to lighten your support burden and
> improve the quality of your tools by opening them up.
Ever considered from what he's living now? Maybe some support for work
on his way could let him going? Ever thought that he like to earn his
living by programming tools? You are not willing to pay for it fine,
live with whatever you like.
>
> I don't believe "socially unacceptable" is quite accurate unless you like
> to hang out in certain IRC channels, which I doubt. "Economically
> unacceptable" maybe. The fact that you can't think of a business model to
> do what you'd generally like, despite the examples of others who have, is
> hardly a chilling indictment of open source.
This is a no argument, just it worked for some does not imply it work
for anyone. I give you some counterexamples. Corel, Eazel, Adobe and I
have my problem with the following picture. On what do Linux
distributors live, on work from others which the hardly pay anything,
oh well they have their distributions but how much effort do the
really put in the software they bundle?
If we play tit-for-tat, I propose that anyone pays a small fee to the
original authar and get's paid for his software. So a repository in
which I can participate support and get supported would be really
nice.
However it would be nice to on what you learn you living. Maybe you
are paid by governement and than it's of course natural to insist on
Open Source. You get paid by all the tax payers. If you are not please
let us know how you get along.
Regards
Friedrich
Friedrich Dominicus <·····@q-software-solutions.com> writes:
> > Seems like a pretty fair tit-for-tat to me.
>
> Yes obviously and from what do they live?
They are paid to program. The majority of software is produced and
used in-house, so the license it is released under is less important
then getting it done.
> However it would be nice to on what you learn you living. Maybe you
> are paid by governement and than it's of course natural to insist on
> Open Source. You get paid by all the tax payers. If you are not
> please let us know how you get along.
I'm paid to program in CL. A large chunk of what I have written over
the last few years is released as Free Software. The freed chunks
were tools used to build an application, which itself is proprietary.
Now, regardless of what my political motivations are, selling tools is
just not something my employer was interested in. It's low, low
margin, difficult to support, and anyone who studies their software
industry history knows the hardships tool vendors faced even before
Open Source become a viable target for their ire.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
"Craig Brozefsky" <·····@red-bean.com> wrote
> Now, regardless of what my political motivations are, selling tools is
> just not something my employer was interested in. It's low, low
> margin, difficult to support, and anyone who studies their software
> industry history knows the hardships tool vendors faced even before
> Open Source become a viable target for their ire.
I quite agree with you here. I also developed and sold software tools in the
past, now I still make tools but I use them to write applications that I
sell. It's much more easy to get money that way.
I will add that, when you sell tools/libraries, very often you trigger the
animosity of the software guys of the companies that buy your tools. The
"not invented here" syndrome.
Marc
On 21 May 2001, Friedrich Dominicus wrote:
> Tim Moore <·····@herschel.bricoworks.com> writes:
>
> > On Mon, 21 May 2001, Kent M Pitman wrote:
> >
> > > > Sure there are some people who clamor for people to volounteer their
> > > > work, but that can't be attributed to "free efforts".
> > >
> > > Yes, it can. A huge number of people are coming out of college these
> > > days thinking that free stuff is an entitlement and that there is
> > > something wrong with a community that doesn't provide it to them.
> >
> > Are these the same people you abuse later for giving their own stuff away?
> >
> > Seems like a pretty fair tit-for-tat to me.
> Yes obviously and from what do they live?
They get paid to write software that isn't sold as a commercial product,
like the vast majority of software engineers.
> >
> > Hey, you're right, I wouldn't pay a dime for your stuff. Why should I
> > trust it? It's had one pair of eyes looking at it, not thousands. As far
> > as I'm concerned, your putative tools have no value unless they've been
> > given a public airing.
> You probably would not pay a dime for anything. That's you choice. Why
> should you trust it? Now why shouldn't you? Do you think any vendor
> will want to do you a harm?
I've actually paid money for a Linux application, Mike Hamrick's excellent
VueScan film scanner app. The price was right ($40), the results are
fabulous, the app embodies a great deal of domain expertise that I'm
unlikely to acquire myself. I don't see those factors converging for a
Lisp tool or development environment anytime soon.
I don't think that any vendor wishes me harm (although some certainly act
that way). I just don't have a particularly high level of confidence that
a closed-source vendor can supply reliable products. If I'm going to get
paged at 3 in the morning, I'd like it to at least be my fault, and I'd
also like a fighting chance of fixing the problem.
> > If you're now motivated to make commercial apps using your tools I can't
> > understand why you wouldn't want to lighten your support burden and
> > improve the quality of your tools by opening them up.
> Ever considered from what he's living now? Maybe some support for work
> on his way could let him going? Ever thought that he like to earn his
> living by programming tools? You are not willing to pay for it fine,
> live with whatever you like.
With his expertise and reputation, Kent Pitman should not need me to buy
his tools in order to make a very good living.
> >
> > I don't believe "socially unacceptable" is quite accurate unless you like
> > to hang out in certain IRC channels, which I doubt. "Economically
> > unacceptable" maybe. The fact that you can't think of a business model to
> > do what you'd generally like, despite the examples of others who have, is
> > hardly a chilling indictment of open source.
>
> This is a no argument, just it worked for some does not imply it work
> for anyone. I give you some counterexamples. Corel, Eazel, Adobe and I
> have my problem with the following picture. On what do Linux
> distributors live, on work from others which the hardly pay anything,
> oh well they have their distributions but how much effort do the
> really put in the software they bundle?
You keep mentioning Adobe, but I don't see what kind an example that is.
They determined there wasn't a market for FrameMaker on Linux; so what?
They were never going to make it open source.
Some open source companies are making it, some aren't. To me that's
indistinguishable from the rest of the business world.
> However it would be nice to on what you learn you living. Maybe you
> are paid by governement and than it's of course natural to insist on
> Open Source. You get paid by all the tax payers. If you are not please
> let us know how you get along.
It's really none of your business. Suffice it to say that I'm paid to
write software that is not commercially sold (except possibly in an
ASP-type arrangement), like most readers of this newsgroup. I'm not paid
to write Lisp, I do that on my own time, perhaps to relive the glory days
of my youth (such as they were), or perhaps because it's satisfying and
fun. Maybe that makes me a dilettante. *shrug*
Tim
Tim Moore <·····@herschel.bricoworks.com> writes:
>
> Some open source companies are making it, some aren't. To me that's
> indistinguishable from the rest of the business world.
you're right
>
> > However it would be nice to on what you learn you living. Maybe you
> > are paid by governement and than it's of course natural to insist on
> > Open Source. You get paid by all the tax payers. If you are not please
> > let us know how you get along.
>
> It's really none of your business.
No it is definitly not. To make the point it's quite different if you
do some hacking for fun in you spare time or depending fully on the
software you write. In the first case it's easy to postpone
open-source-Development whereas in the later it'll be bit harder.
> Maybe that makes me a dilettante.
Why should a hobbyist programmer be dilettante? He/she is under not
pressure and can work on whatever he/she likes. And can build
extraordinary nice software as professional can build extraordinary
bad software.
The point was just there must be a middle ground. Open source Software
can work for some, but insisting that all had to be open is
ridicolous
Regards
Friedrich
On 21 May 2001 07:19:28 GMT, Tim Moore <·····@herschel.bricoworks.com>
wrote:
> Now, I'm being a bit facetious here -- I believe that you, KMP, probably
> have produced something of value and I do understand why you'd never make
> them open source -- but not very.
Given that even open-source leader Eric Raymond writes[*] that a
closed-source approach has its place, it's probably wise not to take it for
granted that each piece of software should be released under the terms of
an open-source license.
Paolo
[*] In his essay "The Magic Cauldron".
--
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://cvs2.cons.org:8000/cmucl/doc/EncyCMUCLopedia/
Paolo Amoroso wrote:
>
> On 21 May 2001 07:19:28 GMT, Tim Moore <·····@herschel.bricoworks.com>
> wrote:
>
> > Now, I'm being a bit facetious here -- I believe that you, KMP, probably
> > have produced something of value and I do understand why you'd never make
> > them open source -- but not very.
>
> Given that even open-source leader Eric Raymond writes[*] that a
> closed-source approach has its place, it's probably wise not to take it for
> granted that each piece of software should be released under the terms of
> an open-source license.
I think only a very tiny, but also very vocal faction actually thinks
everything should be released as open source. Open source certainly has
its place, but it sure isn't the be-all and end-all of software
development. It's equally clear that closed source isn't all that great
for some things, otherwise this debate wouldn't even be happening. I
strongly suspect that Neal Stephenson was correct in his assertion in
his essay "In the Beginning was the Command Line..." that it is (general
purpose) "operating systems' destiny to be free," as well as many tools.
I certainly don't _expect_ freebies, but I do appreciate them if it's
something I can use. Free tools don't destroy the market for proprietary
ones -- witness Java. If you're going to charge for something people can
get elsewhere for free, there's got to be a good reason for them to pay
for it. I think open source has had a very healthy effect on industry,
unlike software patents, for example.
Larry
Kent Pitman has given away a LOT of stuff to the Lisp community. A lot of it
happens not to be code, but do course material, contributions to CLtL(2),
articles, newsgroup contributions, you name it, count for less?
--
Biep
Reply via http://www.biep.org
Kent M Pitman <······@world.std.com> writes:
>
> I have spent a good part of this year only partly employed, trying to
> figure out how to make a business out of things I know for once, rather
> than to automatically join up with some company and make money for it.
> I needed exactly the tools you're talkign about, but I didn't expect anyone
> to provide them to me, so I wrote my own and now I have a bunch of
> intersting stuff. It bugs me is that there is not a market for people to
> license such tools; I don't hear anone posting here saying they'd
> pay a dime
> for any of this stuff. I'm sufficiently not-optimistic about the prospect of
> making any money on licensing this stuff because everyone seems so fixated
> on the idea that this stuff should be freebie-ized that I'm not even trying
> and I'm just going straight through to developing commercial applications
> based on my tools.
I fully understand that, anyway it would be nice to know what you have
done. So why don't you post a list?
>
> I think that if there were a culture here of people saying that code was
> worth something, I would have (a) licensed the stuff I needed from someone
> who had written what I needed and (b) been able to license the stuff I built
> atop that stuff to someone else. Each of us, adding value, would be
> reimbursed for what we created. But instead, we're all involved in
> a rat-race
> where as we develop anything, someone else wants to develop the same thing
> and give it away. That undercuts those of us who are trying to invest in
> something of value.
I agree, I can not see the economical sense behind it. I'm a
programmer and like to build things. I pay my share for things I'm
interested in, but it seems that especially on free Unices hardly
anyone else does. Why isn't there nearly no commercial application in
Linux?
>
> I think teaching college kids that it's good to give stuff away causes
> them not to understand teh value of what they do, and causes them to
> economically injure those people who are producing that stuff for
> money.
The same is true in Germany for ABM stuff. That are firms funded by
taxes, in which unemployed get aquired to improve the unemployment
statistics. The wages are paid by all the people and the can underbid
any other offer obviously.
> I don't see that as doing any good for anyone except the people who are
> sitting around waiting for the freebies to be produced, and the whole system
> is fueled by people who apparently need no income. That's not a stable system.
> I can't figure out where the fuel is.
This has worked in the past as I read in a book. The thing was in the
past peopel desparatly wanted some things and did not get them. This
has changed massivly. If you look of balancing software you can find
hundreds of offers, so if you wrote the next one, you have to compete
with that, and some governement funded people can obviously underbid
any "commercial offer". One area I find this extremly, is in the
market for Development tools. I think the balance is lost. There must
IMHO be a middle ground for development tools.
>
> It is this (non)cycle, not anything technical, that is most in danger of
> driving me away from Lisp after this many years of having held to it.
>
> Certainly it makes me not even the slightest bit interested in spending time
> on the market of creating tools, which is what people claim they so desperately
> need in order to write applications. Because the writing of tools is something
> people say they don't "value" as a market. Value means "pay money" in my
> book. And so that makes it less likely, not more likely, tools will get
> created...
I could not state it better. The question is quite easy IMHO.
Are you willing to spend money on the things you want to have?
Someone in this thread posted he want a unified FFI interface. Are you
willing to pay for it's development?
Others posted about a GUI-Toolkit for Common Lisp for him/her the
anser is easy
CLIM is standard in commercial Lisps. Considered buying one?
If you want another one working an "all the Lisps" are you willing to
pay for it?
If someone feel everything is better in Java becasue you can get this
stuff for free. Please feel free to use Java. I wonder what you'll say
if you implemented some nifty Java stuff, wanted to sell it and will
hear, well why don't you just give it to me?
>
> > Are you just lamenting the
> > lack of tact and grace that some people do it with?
>
> I am very much demoralized and upset at the fact that it is considered
> so socially acceptable these days to get virtually militant about the
> absence of free stuff and it is considered so socially unacceptable to ask
> a fair wage for having created something of value.
Things usually have two sides. We have come from a market where I
would think tools were overcharged. Now the swing is the total
opposite, tools are thrown away. This extremist positions will be left
sooner or later, and the market will calm down. Maybe the swing is
getting back at the moment. A lot of firms betting on OSS ideas have
gone bankcrupt (more or less) and a few firms decided (wisly not to
enter that ruinous market (e.g Adobe they do not offer Framemaker on Linux)
But even OSS can work if someone is willing to pay for OSS
development. It underbid others that is true, but I can't see that the
"needs" will end somewhere. What I see with one laughing and one
crying eye is that a lot of OSS stuff still falls behind, because
no-one wants that stuff or thos who wanted it are not willing to take
their share. OK you can take over but how often does that happen?
Regards
Friedrich
Kent M Pitman <······@world.std.com> writes:
> I am very much demoralized and upset at the fact that it is considered
> so socially acceptable these days to get virtually militant about the
> absence of free stuff and it is considered so socially unacceptable to ask
> a fair wage for having created something of value.
If something *really* has value, people will be willing to pay for it.
And lets try see it from the user's point of view. If I *pay* (money)
for something, it is *mine*. *That* is at the heart of the Free vs
Proprietary conflict. I can resell what I have bought. I can change
it, remove parts, add parts etc. But the proprietary software model
does not want to recognise this. The proprietary firms want to keep
what they sell. That *sucks*.
Why are programmers different to everyone else, Kent? If a ditch
digger digs a ditch for a sewage pipe he should get a fair wage. He
shouldn't get a percentage every time a turd floats down the pipe which
is laid in the ditch he dug.
If someone writes a useful program, he should get paid for the time he
spent on it. Why he should feel justified in expecting to get paid
again and again is beyond me. But that's what proprietary licensing
is all about.
The attitude behind it is greed, arrogance and laziness. Work once,
get paid forever.
If something has to be *secret*, then make it in-house, otherwise
release it so everyone, yourself included, can benefit from your
work. The Free software model allows value to easily be added, for
the benefit of everyone. The proprietary model strives to limit and
control the value.
And if you really can't bear the thought of someone benefiting from
your work for nothing, then don't release. But please don't bitch
about not being able to sell it. If there *really* was a need for it,
someone would buy it.
Regards,
Peter
Peter Wood <··········@worldonline.dk> wrote in message news:<··············@localhost.localdomain>...
> Why are programmers different to everyone else, Kent? If a ditch
> digger digs a ditch for a sewage pipe he should get a fair wage. He
> shouldn't get a percentage every time a turd floats down the pipe which
> is laid in the ditch he dug.
I suspect that Wood hasn't arranged for sewage service. Interestingly
enough, most sewage bills have a charge that is roughly "per turd". (The
typical charge is based on water consumption, which is roughly proportional
to the number of turds. Service where the sewage mix is different have
different rates.)
The ditch diggers may not get a cut of each payment, but the sewage
company does, and Kent was talking about becoming a company. Moreover,
the "value" of the ditch digger's work is a net present value of some
fraction of the "per turd" income stream, so it would be an error to
assume that the ditch digger's salary is independent of the "per turd"
payment stream.
With septic tank systems, the regular service folks do get paid "per turd".
(They fill up and are pumped out periodically.)
I do appreciate the example though, as it demonstrates why licensing
is a perfectly rational charging scheme.
> If someone writes a useful program, he should get paid for the time he
> spent on it. Why he should feel justified in expecting to get paid
> again and again is beyond me. But that's what proprietary licensing
> is all about.
Nope - it's about freedom in how one chooses to get paid, and about
ones rate. I can't afford to pay Kent what he's worth on a one-time
basis, but I can afford to rent his expertise on a per-usage basis.
(Then again, I know how astoundingly valuable Kent's time/work is.)
Why is that arrangement, which we both find acceptable, so unacceptable
to you? Don't answer - I don't care why it's unacceptable to you. Where
do you get off telling us what deals we can make? MYOB.
> If something has to be *secret*, then make it in-house, otherwise
> release it so everyone, yourself included, can benefit from your
> work.
If Kent doesn't get paid, why will he work? More specifically,
why will he work on things that I'd benefit from?
My payment gives Kent a reason to do things that I'd like done, that
I'd benefit from.
> Regards,
Ah yes, the sneer closing.
-andy
Peter Wood <··········@worldonline.dk> writes:
> If something *really* has value, people will be willing to pay for it.
Maybe. Maybe not. Men don't pay their wives for housekeeping and child
rearing. They've gotten used to not doing it.
> And lets try see it from the user's point of view. If I *pay* (money)
> for something, it is *mine*.
If you buy a car from me, you don't own the design specs. You may
still have to have an item serviced at your authorized dealer or risk
having some part of your warranty voided.
Certainly I don't believe in runtime royalties, if that's what you mean.
But if you pay $50 for a disk full of tools, you ought not own development
rights to something it took me three weeks of not working somewhere else
to make. In round numbers, for a $100K annual salary, that's $6K it should
cost to truly own that component, and that's only even if the person writing
it has no overhead for marketing and sales, nor has taken on any risk. If
I sit in a desk job and write code, I get my salary no matter what (barring
layoffs). But if I write code speculatively, some percent of it might not
sell. Let's say for argument that 50% of it won't be a money-maker. Ok,
so that's $12K I need to cover sitting at my desk coding, and we still haven't
discussed marketing. Let's assume that over a long time, there's 100% overhead
for that, since a lot of companies employ 5 non-technical people to every 1
technical person, but maybe the technical people have more loaded salaries
than some of the others. So let's say half the resource of the company goes
to technical and half to marketing, in salary. And we'll ignore the cost of
taking out ads, which we shouldn't, because someone here will argue that ads
can be taken out for free at www.somerandomplacenoonehasheardof.com. So let's
say we're now up to $24K in "reasonable costs" to bring you a library it
takes 3 weeks to write if you want to be a "real business". So the true cost
of this item, if you really account for the details, is somewhere around $25K.
And if you think about it, real companies are probably paying that amount
to have ou do three weeks of coding, since probaly only certain weeks of your
salaried employ actually produce something that goes into product, and only
certain products sell. I think that's a very conservative analysis.
When you buy a car, you're buying iron. When you buy software--well, there
are a lot of ways to market it. Giving it away, I don't see where the money
comes from. (You tell me.) What's your best suggestion for how I recover
the $25K that I think I'm owed for the time and risk I've taken by not sitting
at a desk and saying "yes sir, no sir" to someone who's going to build my
work into some other product, profit from it, and leave neither me nor my
community able to buy the componentware that it took to build that product.
And the more you encourage free software, the more risk there is that every
time I sit down to write something based on whatever level is available on
the market today, that someone else will have done the same thing and
contributed it free. So that drives up the price by increasing the risk I
have wasted my time. Where is the money coming from?
People push free software, but they do not say where the money is in that
model. Can someone explain that?
Btw, MY model of "open software" is you pay EXTRA for source licenses.
If your concern is that you can't verify the code without seeing it, then
you have an additional concern that others don't, and you should pay more.
Also, I think seeing source is an abstraction violation and encourages
the use of non-published interfaces, leading to later breakages that will
be blamed on the vendor but ought to be blamed on the user who violated
the documentation. And if the user is just going to use documentation,
they don't need source.
> *That* is at the heart of the Free vs
> Proprietary conflict. I can resell what I have bought. I can change
> it, remove parts, add parts etc. But the proprietary software model
> does not want to recognise this. The proprietary firms want to keep
> what they sell. That *sucks*.
Yes, the free software model provides writes for everyone but the person who
created the item. If it were rewarding him for the time/risk, then this would
be a fine model. But it is not. It just takes and does not give back.
Where does the money come from to reimburse the person who contributed the
idea and the labor and the sense of self?
> Why are programmers different to everyone else, Kent?
They aren't. They need food just like anyone else. Where have I argued
they are different?
> If a ditch
> digger digs a ditch for a sewage pipe he should get a fair wage.
This is a bad analogy. And, by the way, these guys are unionized.
Programmers generaly are not. Programmers have traditionally commanded
value because they offered something to the corporate world that management
thought it couldn't get free. I'm quite worried this is being destabilized,
and soon programmers will enjoy a serious fall in community status.
Throughout the ages, it's not the smart people but the people with money
and guns, that have ruled the world. THere's been a brief period where
being a nerd was fashionable, but I think that will go by the wayside as
soon as the people in charge can figure out how to tuck us back in our place.
Money is our best protection against that. Money represents compensation
for value. Smart people bring value and ought to be duly compensated.
Who is compensating the fre software person? Where is the money coming from?
The only thing I see is that people selling other are able to fund this
mass delusion of "free software makes money" as a "loss leader" to get you
to buy iron. But no one has pointed to a sustained, stable, money-making
business that is based on free software per se.
> He
> shouldn't get a percentage every time a turd floats down the pipe which
> is laid in the ditch he dug.
Personally, I think a teacher should get a fractional percent of every
student's future worth. That would reward the teachers who
conssitently put out good students.
I think an artist is entitled to a slice of people viewing their paintings
or listening to their music. To pay all artists and musicians a day's wage
and say it doesn't matter is socialist robbery.
Back to the ditchdigger, too, that's something that's been done for
eons, and society has a good idea of what ditches it needs dug.
Society has little idea of what software it needs written, and there
are not standard offices one goes to to make sure "the obviously right
next softare" gets written. There are many choices and the value is
as much in what to write and how it's written as in the doing of the
act. And that's still very individualized, at least for now. I don't
doubt you can beat the art out of it if you bang hard enough on it,
but I won't see you've done the world a favor by reducing the art of
programming to the work-a-day labor of ditch-digging.
> If someone writes a useful program, he should get paid for the time he
> spent on it.
This is the fallacy. He should be reimbursed for the "true cost", not just
the apparent cost. That's why I did the worked analogy up front.
> Why he should feel justified in expecting to get paid
> again and again is beyond me.
This is not a free-software vs paid software rant, this is a capitalist
vs socialist discussion. People should be rewarded for taking risks.
Writing something in hopes it will sell is a risk. There has to be a
benefit to that or people will just go for the safe jobs since it's easy
to show mathematically they will always lose by taking risks. TO get
people motivated to take risks, one MUST reward risk in a way that makes
it at leaset possible to win bigger by taking them than by not taking them.
And on average you have to show that a reasonable person can at least
cover the expected losses they'll endure for having taken the risk.
> But that's what proprietary licensing is all about.
Proprietary licensing is what you get when you have value and the only
way you can market is privately because all sense that there is value to the
public is undercut. Copyright laws exist to PROMOTE sharing without loss
of control. Any sense that sharing implies loss of control is a return to
the idea that makers of great works should do so behind closed walls without
sharing, so they can control the price.
> The attitude behind it is greed, arrogance and laziness. Work once,
> get paid forever.
My desire to recover at least $25K for 3 weeks work on a direction of
my choice is not greed, it is implied by math. It's implied by the fact
that I'll have to market it and that that costs. It's implied by the
fact that I'll have downtime between that project and the next while I figure
out which way to go. If at the end of the year, I don't have more money
in my bank account for having chosen to sit around and write custom libraries
than for having been a cog in a big organization, then I might as well resign
myself to cog-hood. Calling that greedy, lazy, arrogant is plain unfair
and I'll up the ante to ignorant bordering on disrespectful.
> If something has to be *secret*, then make it in-house, otherwise
> release it so everyone, yourself included, can benefit from your
> work.
THis presumes (a) that this is an exhaustive partition of the choices
and (b) that I do in fact benefit.
> The Free software model allows value to easily be added, for
> the benefit of everyone. The proprietary model strives to limit and
> control the value.
No, it seeks reasonable compensation before giving something away.
> And if you really can't bear the thought of someone benefiting from
> your work for nothing,
Where does the money come from?
> then don't release. But please don't bitch
> about not being able to sell it. If there *really* was a need for it,
> someone would buy it.
I've been round and round ragging on Franz for this, so let me play the other
side in fairness.
I claim their product is priced too high. They claim they need the money
to survive as a company. The problem is, we're both right. But the truth
is that the reason their product is priced too high is NOT that they didn't
put that much work into it. It's not their desire for excess profits. It's
that others are willing to give stuff away for such a low cost that it drives
the market cost down. At minimum, the walmart effect makes what you say
be untrue. Anyone equipped for mass sales can clobber anyone not equipped
for it, making it very hard to recover costs. And this problem will only
get worse. But beyond that, what makes things really painful is that Java
sells for as little as it does. Had it sold for $2000 or $3000, then Franz
could command that and I wouldn't say "geez, isn't that a little high?"
I'd say that was the right market place and I wouldn't begrudge them the
money. And I don't like Java. So what I'm saying is that the thing I don't
like is driving the price of the thing I do. That shouldn't be. Bad art
shouldn't drive the price of good art. Bad music shouldn't drive the price of
good music. Why should bad programming languages drive the price of good
ones? Mostly because the people we want to sell to are not equipped to see
the difference. So they go for Java unless Lisp is priced the same and offers
some argunent for better value. Bleah.
When I'm grumbling at Franz for pricing, I'm saying many things, but one
thing I'm not saying is "clearly you will make money if you drop your
prices". What I'm saying is that that I saw this before at Symbolics--
people wanted lower priced machines. Symbolics didn't drop its prices.
People bought other machines. Poof. No more Symbolics. (That's
oversimplified--there were other factors I'd rather not engage here.)
But the point is that when prices come down, you ride with them. That doesn't
assure you'll stay in business, but if you fail to, you'll definitely
not stay in business.
So when you say if you have value, people will buy it, I don't believe it.
It's all about "perceived" value. And perceptions are subject to enormous
manipulation. It is not true that if you have a better mousetrap people
will beat a path to your door. If you have no mousetrap at all, but a cute
informercial about the clever box, then they might. The world is a lot
more complex than you're making it out. Any argument about free software
that takes as a given that "sharing works out better for everyone", that
doesn't say where the money comes from, and that talks about the time it
takes to make something and spends no time discussing either "overhead" or
"risk" as real phenomena carries no weight with me.
Kent M Pitman <······@world.std.com> writes:
> So let's say we're now up to $24K in "reasonable costs" to bring you
> a library it takes 3 weeks to write if you want to be a "real
> business". So the true cost of this item, if you really account for
> the details, is somewhere around $25K.
I think you missed out things like office costs and infrastructure
costs. The usual figure for these is that it costs double salary to
emply someone technical, don't know the equivalent for a sales-type
person. (Probably about the same because they always seem to need a
titanium laptop with a 1600x1200 screen and a Gb/sec of net bandwidth
for all the important games they need to download erm I mean research
they need to do). So I think you should about double your total.
--tim
Kent M Pitman <······@world.std.com> writes:
> Peter Wood <··········@worldonline.dk> writes:
<snip>
> > And lets try see it from the user's point of view. If I *pay* (money)
> > for something, it is *mine*.
>
> If you buy a car from me, you don't own the design specs. You may
> still have to have an item serviced at your authorized dealer or risk
> having some part of your warranty voided.
Why do so many so-called "capitalists" want to have their cake and eat
it? I *own* the *car*. I can sell it or give it away. The warranty
has nothing to do with it. If software was like this then a
proprietary firm could say to me, "OK, if you change it the warranty
is void." I would think that reasonable. Instead they say, "You
can't change it, and the warranty is void anyway in 90 days, and if it
causes your computer to exude noxious fumes which poison you, we are
in no way resposible."
> When you buy a car, you're buying iron. When you buy software--well, there
> are a lot of ways to market it. Giving it away, I don't see where the money
> comes from. (You tell me.)
Have a look at successful Free software (Apache, Linux, Perl,
etc). The "money" comes because they are *useful*. *Lots* of people
*use* Apache and make money, perhaps from something unrelated to
Apache itself. If they improve Apache they give the improvement back
to the community, because they know that the others will do the same.
So everyone who uses Apache benefits, clearly and concretely. The
compensation may be paid in cash, but there is a clear benefit. That
is *one* of the ways people benefit from free software. Who is the
Apache vendor? There isn't _one_. Who wrote Apache? Lots of people
did and still do. And they all benefit if they use the software. Who
wrote Linux? Lots of people, all of whom benefit from having a solid
os.
So, if you have written something which you use and you think others
can use too, you can choose to "Free" it, in the hope that others will
make it even better. If that happens, *that* is a clear benefit to
*you*.
This model works very well in some areas, and poorly in others. It
works well when the distinction between user and programmer is
blurred. It works poorly when the programmer isn't actually going to
use the product, because then he is *not* going get recompensed. The
solution to that is simple and the free market will take care of it.
If there *really* is a need for that type of software, (eg flashy
games) there will be money set aside to pay for it.
People sometimes say Free software is not viable because "software
development is so expensive." They completely miss the point. Free
software is attractive and viable (for developers who are also users)
for precisely that reason.
So what if there are lots of people who get a free ride, ie just use
without contributing. I think it's *good* that people can get
something for nothing. I have a colleague who always breaks stuff
before throwing it onto the dump, so that the Gypsies who scrounge
stuff won't get something for nothing. I find that attitude revolting
and appalling. It actually sickens me. I think it is actually evil.
<snip>
>
> Yes, the free software model provides writes for everyone but the person who
> created the item.
Under the GPL, he has the same rights as everyone else. And if he
wrote the software for the right reason (because he needed to use it),
he will benefit, over time, from those rights, just as everyone else
who uses the software will. On the other hand, if he wrote it in
expectation that he could palm it off on a bunch of lusers and lean
back in his chair and watch his bank balance rise, he is in for a
shock.
<snip>
> To pay all artists and musicians a day's wage and say it doesn't
> matter is socialist robbery.
Then for a musician to get lucky with a few big hits, and live in ease
for the rest of his life is a capitalist con trick. Thousands of
other musicians, lots of them much better than the lucky hit maker,
are impoverished. So what. It's a choice they made, and if they are
unhappy then they can change occupations. Do you know the investment
required to become a really good musician? By comparison, software
development pales into insignificance. They invest their *lives*.
Hows that for a risk? Most of them are never fairly recompensed for
that (in money). But the record companies often make a killing.
<snip>
> > Why he should feel justified in expecting to get paid
> > again and again is beyond me.
>
> This is not a free-software vs paid software rant, this is a capitalist
> vs socialist discussion. People should be rewarded for taking
> risks.
No they should not. People should be rewarded for doing useful work.
If you take a risk on something, and luck out, well and good. That
implies the work you did was *needed* since people will pay. If you
take a risk and your enterprise bombs, well tough shit. That's why
it's called a "risk".
<snip>
> So when you say if you have value, people will buy it, I don't
> believe it. It's all about "perceived" value. And perceptions are
> subject to enormous manipulation.It is not true that if you have a
> better mousetrap people will beat a path to your door. If you have
> no mousetrap at all, but a cute informercial about the clever box,
> then they might. The world is a lot more complex than you're making
> it out. Any argument about free software that takes as a given that
> "sharing works out better for everyone", that doesn't say where the
> money comes from, and that talks about the time it takes to make
> something and spends no time discussing either "overhead" or "risk"
> as real phenomena carries no weight with me.
The market knows better than you. Maybe people's real need is to be
conned into buying the vapour-ware mouse trap. If they *really* need
to catch lots of mice, then they *will*, in time, use the better
design. Especially if they can download it free from the internet and
build it themselves :-) And everyone will benefit (including the
designer) because there will be fewer mice around to spread
disease. :-) On the other hand, if the device is too pricey, people
might just get a cat or make do with what they've got.
I can sum up my position very easily: If you can't sell something,
that means *you* are doing something wrong. *Not* that your buyers
are idiots, or brain washed, or that those funky Free software commies
have poisoned your well.
Regards,
Peter
>>>>> "Peter" == Peter Wood <··········@worldonline.dk> writes:
Peter> Why do so many so-called "capitalists" want to have their cake and eat
Peter> it? I *own* the *car*. I can sell it or give it away. The warranty
Peter> has nothing to do with it. If software was like this then a
Peter> proprietary firm could say to me, "OK, if you change it the warranty
Peter> is void." I would think that reasonable. Instead they say, "You
Peter> can't change it, and the warranty is void anyway in 90 days, and if it
Peter> causes your computer to exude noxious fumes which poison you, we are
Peter> in no way resposible."
Except that software doesn't even say that. Software usually only
warrants that the the media is ok. There's no warranty that the
software actually does anything. Can you imagine buying a car like
that?
Ray
In article <··············@rtp.ericsson.se>,
Raymond Toy <···@rtp.ericsson.se> wrote:
>Except that software doesn't even say that. Software usually only
>warrants that the the media is ok. There's no warranty that the
>software actually does anything. Can you imagine buying a car like
>that?
>
I think that this practice among commercial software houses throws
away a potential competitive advantage over open source, although
quite a few people don't seem to have realized it.
If I buy a thing, for almost any value of thing that isn't software,
and it proves defective and dangerous, I can sue the manufacturer.
It's not going to be worth it for me as an individual, but it may
well be worth it (if only for the threat value) for a business
or somebody interested in organizing a class action suit.
Commercial software usually comes with as effective a warranty as
you'll see on Gnu software, and actually less recourse if it
doesn't work as advertised. (At least with GPLed software, you
can hire somebody to hack the source.) For this reason, I consider
open source software to be better backed and supported than the
commercial software I've bought.
This is not the ideal situation. Ideally, commercial software
firms would provide some sort of legally actionable responsibility
to their users. In practice, they don't, and anybody buying
commercial software rather than downloading free in the belief that
the commercial stuff is backed by anybody's promises is a fool.
In "Good Omens", by Pratchett and Gaiman, the devil Crolley
takes a standard software license agreement, attaches a sticky
note saying "Learn, guys" and sends it to Hell's Soul Contract
Division.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
"James Hague" <···········@volition-inc.com> writes:
> Peter Wood wrote:
> >
> > Have a look at successful Free software (Apache, Linux, Perl,
> > etc). The "money" comes because they are *useful*.
>
> A bit off topic, but has anyone else wondered why the list of great success
> stories, especially for a movement that is based on an endless supply of
> free labor, is so *short*? The "etc" usually translates to Python, TCL, and
> some of the gnu packages. Now I certainly understand that there are many
> more open source programs than this, but it is interesting how the same
> small group of relatively old programs (Linux dates to 1991; Perl started
> life in the 1980s) roll off the tongue so quickly as the canonical examples.
Because they are the most widely recognizable. I could give you a
dpkg listing of my Debian box, but 90% of the software I find useful
every day and that I'm thankful for would be unidentifiable to the
public at large, gaim, gnapster, emacs, bbdb, gnus, mozilla, mpg123,
xmms, csound, freebirth, bash, gzip, tar, GNUStep, GTK, Gnome, xterm,
X11, WindowMaker, aewm, nawm. just to name the ones I use daily.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
"James Hague" <···········@volition-inc.com> writes:
> Peter Wood wrote:
> >
> > Have a look at successful Free software (Apache, Linux, Perl,
> > etc). The "money" comes because they are *useful*.
>
> A bit off topic, but has anyone else wondered why the list of great
> success stories, especially for a movement that is based on an
> endless supply of free labor, is so *short*? The "etc" usually
> translates to Python, TCL, and some of the gnu packages. Now I
> certainly understand that there are many more open source programs
> than this, but it is interesting how the same small group of
> relatively old programs (Linux dates to 1991; Perl started life in
> the 1980s) roll off the tongue so quickly as the canonical examples.
Apache, Perl, Python, TCL, (x)emacs and a few others differ
from *most* other free software in a few ways:
- they are well-written
- they are well-documented
- they are relatively easy to build and install
Note that well-written does not imply "has aestethically
pleasing syntax and semantics" :-)
--
Raymond Wiker
·············@fast.no
"James Hague" <···········@volition-inc.com> writes:
> A bit off topic, but has anyone else wondered why the list of great success
> stories, especially for a movement that is based on an endless supply of
> free labor, is so *short*?
This is pretty ignorant.
I'm working on Mac OS X as I write this, with its free underpinnings.
And I'm completely dependant on OpenSSH, OpenSSL and a lot of other
technologies that you didn't include in your "etc".
--
(espen)
"James Hague" <···········@volition-inc.com> writes:
> You missed the point. The point is not there aren't other examples, but
> that everyone blindly cites the same stock examples. When the topic is
> "Open Source is great!" there's always a line like "Look at Linux, Perl, and
> Apache!" right near by.
If _that_ was your point, I have to say I'm not as puzzled as you.
Linux, Perl and apache _are_ probably the most prominent examples.
Just look at http://www.netcraft.com/survey/ and think about all
the marketing $ MS puts into IIS...
--
(espen)
In article <··············@localhost.localdomain>,
Peter Wood <··········@worldonline.dk> wrote:
>
>Have a look at successful Free software (Apache, Linux, Perl,
>etc). The "money" comes because they are *useful*. *Lots* of people
>*use* Apache and make money, perhaps from something unrelated to
>Apache itself. If they improve Apache they give the improvement back
>to the community, because they know that the others will do the same.
>So everyone who uses Apache benefits, clearly and concretely. The
>compensation may be paid in cash, but there is a clear benefit. That
However, I don't think Kent is interested in "money", but rather in
money. Suppose you are not working for a company. How are you
going to support yourself writing free software? I've read Eric
Raymond's essays (at least what's available for free on his web
site), and I'm not convinced that there's a viable business model
there, aside from being somewhat parasitic on other businesses.
(For example, it makes sense for O'Reilley to spend money on open-
source software development, so that people will buy their books
on the software.)
If Kent spends his time writing free software, then he is not making
money. I personally would like him to continue to eat, program Lisp,
and have a net connection so that he continues his well-thought-out
contributions to comp.lang.lisp. (Not that I'm willing to pay much
for that.) This means that he has to find ways to sell software,
and in practice to do that he has to avoid giving it away to
everybody who wants it.
I think it a wonderful thing that so many people will contribute
their work in the form of excellent software for no financial
remuneration, but as a matter of practice these people aren't
quitting the day job. Perhaps, if they were paid to do so, they
would quit the day job and write much more and better free software,
and perhaps society would greatly benefit from this, but I don't
see it happening any time soon, nor do I know of any way to
establish this as a stable part of the economy.
So, I wish Kent luck on finding ways to sell software so he can
maintain himself in the style to which he would like to become
accustomed.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
········@visi.com (David Thornley) writes:
> In article <··············@localhost.localdomain>,
> How are you going to support yourself writing free software?
How are you going to support yourself trading in free markets?
Free Market is NOT a business model. It is a legal statement of justice.
It is the refusal of political action and government-protected monopolies.
Free Software is just the same. It doesn't dictacte how you are to
earn money. It does give a damn. It tells you how you won't "earn" money:
you won't make money by preventing people to exchange services.
When monopolies are gone, everyone is back to the bottom line:
to get money from other people, you must provide value to *them*.
So here's the question: do you provide value to anyone?
If you do, then you needn't fear Free Software;
it will be a great way to do business in a way that satisfies everyone.
If you don't, then fear and loathe Free Software,
for it will come and terminate your racket.
> If Kent spends his time writing free software,
> then he is not making money.
I spend time writing free software and making money.
Maybe Kent just isn't good at marketing his talent.
That's what firms are about: gathering proficiencies
to make them into a marketable service offer.
The advantage of LISP is all about metaprogramming.
Metaprogramming is much less interesting
when you just don't have the license to do it.
Actually, it's detrimental, when you want to hoard your own software.
http://fare.tunes.org/articles/ll99/index.en.html
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
Because people confuse information and information-related services
(which include searching, creating, processing, transforming, selecting,
teaching, making available, guaranteeing, supporting, etc), they are afraid
that Free (libre) Information mean free (gratis) information-related services,
which would indeed kill the industry of said services. On the contrary,
Free Information would create a Free Market in these services, instead of
current monopolies, which means they will be available at a fair price,
so the result would be a flourishment of that industry!
In article <··············@Samaris.augustin.thierry>,
Francois-Rene Rideau <·······@SPAM.tunes.org> wrote:
>So here's the question: do you provide value to anyone?
>If you do, then you needn't fear Free Software;
>it will be a great way to do business in a way that satisfies everyone.
The question is not whether you provide value to anyone, but
rather do you provide value you can get paid for. The Gnu
Project has written software of great value, but last I looked
at the web site it still asked for donations, and said that
X many business purchases would support a programmer for a year
(apparently at a salary I'm not willing to work for).
If the Gnu Project had some sort of compensation proportional to
the value they have created, say 10%, they wouldn't have to
worry about anything financial.
So, unless you have a suggestion for how people can get paid
well for writing free software, don't blame others for wanting
to do something they can be paid for.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
Kent M Pitman <······@world.std.com> writes:
> My desire to recover at least $25K for 3 weeks work on a direction of
> my choice is not greed, it is implied by math. It's implied by the fact
> that I'll have to market it and that that costs. It's implied by the
> fact that I'll have downtime between that project and the next while I figure
> out which way to go. If at the end of the year, I don't have more money
> in my bank account for having chosen to sit around and write custom libraries
> than for having been a cog in a big organization, then I might as well resign
> myself to cog-hood. Calling that greedy, lazy, arrogant is plain unfair
> and I'll up the ante to ignorant bordering on disrespectful.
Short question: what does your toolbox do? If it would be very useful
for me I could imagine paying, say $100, for making it free. So, if
250 readers of this list would join me, we might buy it from you and
make it free.
Taking this further, one could imagine the following alternative to
free software: people could found a club with annual fees (<= 100$)
with the only goal to buy software from vendors and authors to make it
available as open source for the club members. If sufficiently many
people would join, we could buy your tools, or perhaps even the
commercial version of Macsyma from Symbolics:-)
To make free software zealots also like the club, one could establish
the rule that bought software would be made free after some years.
N.B.: It is important that it is a club with every member having a
vote concerning the policy.
Yours, Nicolas.
Nicolas Neuss <·····@ortler.iwr.uni-heidelberg.de> writes:
> Short question: what does your toolbox do? If it would be very useful
> for me I could imagine paying, say $100, for making it free. So, if
> 250 readers of this list would join me, we might buy it from you and
> make it free.
I've thought about this scenario a great deal, and numerous close
variants. I won't say it's impossible, but it's quite a lot harder, I
think, than you seem to acknowledge here, though in fairness I assume
your intent wasn't to get too deep into it on first post.
The problem is that it has to not be free while the money piles up,
since the goal mihgt not be met. The money COULD be placed in escrow,
though that complicates things substantially. The money is then doing
no one good while it waits. Or people could just pledge, but then the
money might not really be there by time all the pledges are tallied.
Btw, as to my tools in particular, I'm not quite ready to release
information about what they do. I was speaking only generally right
now about this business scenario, and not meaning to turn this into
something about me and my wares particularly. Also, of course, the
particular numbers we've mentioned here are slightly contrived for
discussion. Give me another month or two before I've got it in a form
where it's rational for me to entertain public discussion--having to
defend it at this point would be more trouble than writing it in the
first place.
> Taking this further, one could imagine the following alternative to
> free software: people could found a club with annual fees (<= 100$)
> with the only goal to buy software from vendors and authors to make it
> available as open source for the club members. If sufficiently many
> people would join, we could buy your tools, or perhaps even the
> commercial version of Macsyma from Symbolics:-)
This is an expensive club. Even the ACM is cheaper at a baseline. I don't
know how popular it would be. And, moreover, those who really believe in
free software often cite their ability to trade code instead of money as
a way of contributing. That, of course, would not work. Of course, you
could make it so that everyone puts in, say, $1000 and then everyone writes
code worth $1000, and then everyone draws out $1000 for their code, reimbursing
their membership fee and making them feel good about having gotten a deal.
But if you looked closely at the numbers you might find it was a lot equivalent
to not having the club... ;-)
> To make free software zealots also like the club, one could establish
> the rule that bought software would be made free after some years.
Software should become free as soon as it's paid for. It's just a matter
of accounting which is and which isn't. In a pooled situation, you have
to resolve that or it gets slushy and weird.
> N.B.: It is important that it is a club with every member having a
> vote concerning the policy.
Ugh. Don't forget to budget a huge amount of the $$ to red tape. One
reason businesses run efficiently is that they are all dictatorships.
The power/recourse isn't a vote, it's the ability to leave the company
(as an employee or consumer) if you don't like the mode of autocracy
and find a better mode. But thankfully they don't waste time on
voting. They just either get it right or go out of business. If they
did stay and try to perfect things, things would be a lot pricier.
It's a worthwhile area of discussion you raised, I think. I just
happen to think it requires more depth to really get at the dynamics
that would drive it. You've got to get past the idealism and down to
the day-to-day mechanics and economics.
"Kent M Pitman" <······@world.std.com> wrote in message
····················@world.std.com...
> Nicolas Neuss <·····@ortler.iwr.uni-heidelberg.de> writes:
>
> > Short question: what does your toolbox do? If it would be very useful
> > for me I could imagine paying, say $100, for making it free. So, if
> > 250 readers of this list would join me, we might buy it from you and
> > make it free.
>
> I've thought about this scenario a great deal, and numerous close
> variants. I won't say it's impossible, but it's quite a lot harder, I
> think, than you seem to acknowledge here, though in fairness I assume
> your intent wasn't to get too deep into it on first post.
>
> The problem is that it has to not be free while the money piles up,
> since the goal mihgt not be met. The money COULD be placed in escrow,
> though that complicates things substantially. The money is then doing
> no one good while it waits. Or people could just pledge, but then the
> money might not really be there by time all the pledges are tallied.
There is a public goods problem with this model. If someone else puts up
enough money to get Kent to release the software for free, then I get the
software for no money -- so I have an incentive not to put up any money. If
I put up part of the money to get Kent to release the software but not
enough other people do, then I lose the use of the money while Kent waits,
and I get nothing -- so I have an incentive not to put up any money. Or Kent
doesn't require the money to be in escrow before releasing the software and
maybe I decide not to pay now that I have it -- so I have an incentive not
to put up any money.
So Kent isn't likely to get paid his full amount and therefore isn't likely
to want to release his software.
Maybe we should all think a minute about what and how mathematicians
get paid. Have you ever seen one getting paid for a mathematical
invention? No? An then think about any andvanced product on the
market and guess wether it would exist without more or less advanced
mathematics. (e.g. cimputer tomography, mail encryption, CAD software
etc.)
Maybe capitalism (ok, make that market) doesn't just work for the real
good things.
Just think about it.
Andreas
--
Wherever I lay my .emacs, there�s my $HOME.
Andreas Eder <············@t-online.de> writes:
> Maybe we should all think a minute about what and how mathematicians
> get paid. Have you ever seen one getting paid for a mathematical
> invention? No? An then think about any andvanced product on the
Yes I have seen mathematicians get paid for their discoveries. I have
even seen mathematicians starting their own companies based on their
own discoveries. But mathematical discoveries are sufficiently
dissimilar from software that I can see no reason to bring them into
this discussion.
> Maybe capitalism (ok, make that market) doesn't just work for the real
> good things.
Maybe capitalism works great for producing enough wealth that
governments can siphon off certain parts of that in order to employ
large numbers of mathematicians at public research institutes. Maybe
it also works by creating companies that are sufficiently well funded
in order to employ mathematicians as part of their R&D departments.
Maybe this is a real improvement over former times, where mathematics
(and all science) more or less rose and fell with the benevolence of
the rulers?
> Just think about it.
What is slowly getting on my nerves in all discussions involving free
or open-source software, is that a) no metaphore and no comparison is
too bad to be brought up, and b) that many participants seem to have
no clue at all about either the exact workings of the current social
systems, nor any clue at all about the historic development of such.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
"Pierre R. Mai" <····@acm.org> writes:
> What is slowly getting on my nerves in all discussions involving free
> or open-source software, is that a) no metaphore and no comparison is
> too bad to be brought up, and b) that many participants seem to have
> no clue at all about either the exact workings of the current social
> systems, nor any clue at all about the historic development of such.
You make me laugh. *Noone* has a clue about the "exact workings" of
the current social systems, or about their historical development!
Reality can support many different models.
Also, perhaps you need to understand that a metaphor is not literal.
Creative people can use metaphors and analogies, sometimes
experimentally, to understand *aspects* of complicated things which
are not immediately obvious. A metaphor might be very bad in some
respects, and apply quite nicely in some others. If the aspects you
are trying to get a better handle on are the ones where it applies,
then that is a good metaphor. Of course, to someone who thinks other
aspects are more important, it might appear to be a bad metaphor.
Regards,
Peter
[Andreas Eder]
| Maybe we should all think a minute about what and how mathematicians
| get paid. Have you ever seen one getting paid for a mathematical
| invention? No?
mathematicians do get paid for cranking out formulas. in fact, powerful
algorithms are kept secret.
--
Rolf Lindgren http://www.roffe.com/
·····@tag.uio.no
Peter Wood <··········@worldonline.dk> writes:
>
> If someone writes a useful program, he should get paid for the time he
> spent on it. Why he should feel justified in expecting to get paid
> again and again is beyond me. But that's what proprietary licensing
> is all about.
Oh no problem just tell me what you want (under the condition you'll
wan it in Lisp) you get a figure from me how
much to pay for the devlopment and of we go. Yes you can have the
sources too.
>
> The attitude behind it is greed, arrogance and laziness. Work once,
> get paid forever.
Is that all you can read out of Kents messages?
My situtation is very simple. I wan to develop software it does not
bother me if I work for OSS or not. I just want to get paid for the
job and that's it. Now assuming you wan to have a new super-cool
GUI-Toolkit for Common Lisp, let's assume a year of programmer time
will cost you $50 000. I guess for getting a system really good it
took between 5-7 man years. So you just get along between $250 000 to
$350 000. What you do with it is up to you.
Now alternativly I took the risk implementing that out of my own
pocket. And I hope to get the money in later. Now the costs are
simular high + some extra for the profit (who knows when the next
project comes along) I calculate with $250 a copy. Now that are just
1000-1400 copies. Ah yes I have to package the product and maintain it
well I need at least another $50000 each year so another 200
copies. And you probably find I get paid forever.
>
> If something has to be *secret*, then make it in-house, otherwise
> release it so everyone, yourself included, can benefit from your
> work. The Free software model allows value to easily be added, for
> the benefit of everyone. The proprietary model strives to limit and
> control the value.
I do think you are both wrong and right at the same time. It's neither
that all proprietary software is bad and open software is good. It's
in between. Anyway sometimes I even want to buy something but do not
get any offer. Pleas try to get accounting software for Linux. You
won't get neither a free nor a closed one. That is partly due to the
fact that a) the vendors don't get paid for it b) hackers don't spend
time on such undertacking.
>
> And if you really can't bear the thought of someone benefiting from
> your work for nothing, then don't release. But please don't bitch
> about not being able to sell it. If there *really* was a need for it,
> someone would buy it.
Again it would be nice to know what he has to sell would you agree?
Regards
Friedrich
Peter Wood <··········@worldonline.dk> writes:
> And lets try see it from the user's point of view. If I *pay* (money)
> for something, it is *mine*. *That* is at the heart of the Free vs
If you *buy* something it is yours. If you rent, lease, or license
something (all of which involves you paying for something), it is not
yours, but you obtain some rights to it. Just because you rented a
flat, does not mean that you may sub-let it at will, and when you
lease a car that doesn't mean you may lend or sell it to anyone you
want.
That is a distinction between buying something (which effectively
transfers most, but not always all, rights to you) and other forms of
obtaining certain rights, which aren't as pervasive (and hence not as
expensive either).
Since you think that sucks, I imagine that you bought your house, car,
and all other possessions.
> Proprietary conflict. I can resell what I have bought. I can change
Yet you can't resell what you have leased.
> it, remove parts, add parts etc. But the proprietary software model
Again all those rights are severely curtailed for e.g. a rented flat.
Yet no-one goes around saying that renting flats or houses should be
universally stopped (you might make that determination for yourself,
though).
> does not want to recognise this. The proprietary firms want to keep
> what they sell. That *sucks*.
If they wanted to do that, that wouldn't suck, it would be legally
untennable. Yet they don't sell you their software (at least not
normally), they grant you certain rights to it, in exchange for
money. If the rights you obtain aren't to your liking, you are of
course free to negotiate a different contract if you can afford to.
Most software producers that don't sell shrink-wrapped commodity
software will grant you further rights if you negotiate a
corresponding contract.
> Why are programmers different to everyone else, Kent? If a ditch
> digger digs a ditch for a sewage pipe he should get a fair wage. He
> shouldn't get a percentage every time a turd floats down the pipe which
> is laid in the ditch he dug.
>
> If someone writes a useful program, he should get paid for the time he
> spent on it. Why he should feel justified in expecting to get paid
> again and again is beyond me. But that's what proprietary licensing
> is all about.
Just like the producer of music should be paid just once, and then
everyone should be free to profit of his work, with him not getting
paid a dime. And I'm sure you agree that the greatest rip-off of all
are those pesky authors of books, who somehow are so greedy that they
really consider it their right to get a percentage of profits, instead
of leaving all of that to the poor publishing houses.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
"Pierre R. Mai" <····@pmsf.de> writes:
> Peter Wood <··········@worldonline.dk> writes:
>
> > And lets try see it from the user's point of view. If I *pay* (money)
> > for something, it is *mine*. *That* is at the heart of the Free vs
>
> If you *buy* something it is yours. If you rent, lease, or license
> something (all of which involves you paying for something), it is not
> yours, but you obtain some rights to it.
I think it was fairly obvious that I was referring to buying.
Software vendors should make it much clearer to the public exactly
what transaction is taking place. But they don't want to do that!
People might start *reading* the licence agreements!
> That is a distinction between buying something (which effectively
> transfers most, but not always all, rights to you) and other forms of
> obtaining certain rights, which aren't as pervasive (and hence not as
> expensive either).
>
> Since you think that sucks, I imagine that you bought your house, car,
> and all other possessions.
Yes, I did. It very seldom makes sense to rent. Sometimes its
convenient for a short while, but its almost always a money loser.
> > it, remove parts, add parts etc. But the proprietary software model
>
> Again all those rights are severely curtailed for e.g. a rented flat.
> Yet no-one goes around saying that renting flats or houses should be
> universally stopped (you might make that determination for yourself,
> though).
>
I don't think the leasing model is appropriate for software. I can
return a flat or a car. I can't return software in the same way,
because I *never took it away*. I see software, at least partly, as
ideas. If I have an idea and tell you, then we are both "have" the
idea. The situation is completely different if I rent you a flat.
Either you have it or I do. So, I was talking about buying. And
anyway, leasing physical things like a flat usually gives you more
rights than you get when "leasing" software. Amongst other things,
most civilised countries have legislation to protect renters from slum
lords. But the slum lords of the software world are given almost free
rein. Free software makes things harder for the slum lords but anyone
who applauds that fact is a commie free loader, right?
> > does not want to recognise this. The proprietary firms want to keep
> > what they sell. That *sucks*.
>
> If they wanted to do that, that wouldn't suck, it would be legally
> untennable. Yet they don't sell you their software (at least not
> normally), they grant you certain rights to it, in exchange for
> money. If the rights you obtain aren't to your liking, you are of
> course free to negotiate a different contract if you can afford to.
"If you can afford to"? In practice, the small guy is powerless
against the cut-throat megacorp. Shit! How much *time* do you think
the small business can afford to waste on legalese? Free software
helps to level the playing field.
>
> > Why are programmers different to everyone else, Kent? If a ditch
> > digger digs a ditch for a sewage pipe he should get a fair wage. He
> > shouldn't get a percentage every time a turd floats down the pipe which
> > is laid in the ditch he dug.
> >
> > If someone writes a useful program, he should get paid for the time he
> > spent on it. Why he should feel justified in expecting to get paid
> > again and again is beyond me. But that's what proprietary licensing
> > is all about.
>
> Just like the producer of music should be paid just once, and then
> everyone should be free to profit of his work, with him not getting
> paid a dime.
So you think that when a musician has played a concert, he should
*also* be payed next year for that concert? Artists (musicians,
authors, etc) should get payed fairly for the work they do. The
system today rewards the record companies more than the musicians.
Thats a fact. OTOH *Noone*, IMO, is justified in feeling that 3 weeks
of work (or 3 months or 3 years) entitles them to a lifetime of
income. If they can work it so they actually get a lifetime of
income, well congratulations to them! But I won't accept bitching and
whining when they can't work it that way.
> And I'm sure you agree that the greatest rip-off of all
> are those pesky authors of books, who somehow are so greedy that they
> really consider it their right to get a percentage of profits, instead
> of leaving all of that to the poor publishing houses.
The publishing house (or record company) controls the means of
realizing the artist's work, so the recording artist has traditionally
been forced into a partnership with them. Under those conditions, the
percentage system was necessary. That situation seems to be changing,
and artists need to find new ways of realizing their work, and new
ways of generating income. Anyway, software is a bit different.
Anyone with a compiler and network connection has the means of
realizing theirs (and others) works. [I expect someone will jump in
and beat me on the head now with the fact that "the compiler doesn't
come from nowhere, someone's work has gone into it." Yes - gcc has
been made by its users, who share it to their mutual benefit]
"Art" is not something static. The economic model for the production
of works of art in the Rennaisance was very different to what it has
been for the last 100 years or so. Its probably changing now.
BTW, Please don't twist my words to make it look like I think people
shouldn't get paid. I have stated several times, that I think
everyone should be paid fairly for whatever useful work they perform.
I just hate to hear bleating. The world doesn't owe *anyone* a
living. The bottom line is "if you can't make money doing what *you*
want, then get off your arse and do something else." Most people
learn that lesson as adolescents.
Free software is not going to totally replace Proprietary software in
all areas. It is going to (and does) give users/developers a real
choice, rather than the pseudo-choice involved in dealing with
megacorps (or even just big corps). It is a good deal for the guys
with knowledge but little capital. The best Free software is made (in
the fullest sense of the word) by its users. That is probably what
really bugs a lot of conservative programmers who still think of
themselves as members of an elite, exclusive group: Programmers vs
lusers.
Regards,
Peter
"Knowledge itself is Power" - Francis Bacon
Peter Wood <··········@worldonline.dk> writes:
> I don't think the leasing model is appropriate for software.
Buying is out. If I sell it to you, as a matter of law, *I* don't own
it any more. The only legal mechanism that can work where information
is copied is licensing. Licensing is done by dictating use but
retaining ownership. No software developer, other than a contractor
being paid a large wage for a one-shot deal, will ever give up his
rights to the first customer.
> I can return a flat or a car. I can't return software in the same
> way, because I *never took it away*.
It is always your option not to license the software. The beauty of the
legal system of contracts is that their conditions, whatever they are,
are just about always allowed because they are voluntarily entered.
It is your option not to enter the contract.
> > If they wanted to do that, that wouldn't suck, it would be legally
> > untennable. Yet they don't sell you their software (at least not
> > normally), they grant you certain rights to it, in exchange for
> > money. If the rights you obtain aren't to your liking, you are of
> > course free to negotiate a different contract if you can afford to.
>
> "If you can afford to"? In practice, the small guy is powerless
> against the cut-throat megacorp.
Nonsense. If enough people felt this way, there would be a market for
another mode of business. Small businesses can't afford all kinds of
things that big businesses can, and so other businesses spring up to
serve them.
> Shit! How much *time* do you think
> the small business can afford to waste on legalese? Free software
> helps to level the playing field.
Of all the problems I have in my emerging business, the availability
of affordable software with a usable license is not one of them. Just
the opposite. Software is SO affordable from some sources that I may
not be able to make any money selling certain kinds of software that I
produce.
> So you think that when a musician has played a concert, he should
> *also* be payed next year for that concert?
Yes, I do, actually. Because if you don't, then they will be playing
this same concert next year instead of new music. They MUST be payed
enough money not just for the show but to pay for travel, to pay for
advertising, to pay for vacation time, to pay for the time it takes to
write new material, the time it takes to rehearse that new material.
If they are not given that kind of wage, I'd better not expect them to
do anything but play someone else's music over and over. The creative
process is fundamentally different than sweeping floors. It costs money
to create things and the consumers of the creative ware must invest in
that or expect that the content creator will have inadequate funds.
> Artists (musicians,
> authors, etc) should get payed fairly for the work they do. The
> system today rewards the record companies more than the musicians.
This is a red herring. Live with it. You will not invent a system
that does not unfairly reward someone except a system that rewards no
one.
> Thats a fact. OTOH *Noone*, IMO, is justified in feeling that 3 weeks
> of work (or 3 months or 3 years) entitles them to a lifetime of
> income.
I'm going to post a response to this under separate cover. This statement
is based on a HUGE fallacy about how the capitalistic market process works.
It sounds very persuasive rhetorically but really is inappropriate.
(And, incidentally, I think it is false on its face. If Einstein or Salk
or Mozart did all their work in 3 weeks or 3 years, they are absolutely
entitled to a lifetime of pay because of the benefit others received.
Moreover, in some situations, the doing of the deed burns the person out.
Ballplayers get a lot of money in part because their career depletes them
physically by a young age and they must live on that money the rest of
their lives. The same might be said for war veterans, lamentable as war
is.)
> If they can work it so they actually get a lifetime of
> income, well congratulations to them! But I won't accept bitching and
> whining when they can't work it that way.
So your sense of appreciation for what someone does for you is not based on
the degree of value you have received? That seems callous.
> The publishing house (or record company) controls the means of
> realizing the artist's work,
You're really fixated on this, but it is irrelevant. The opposite might be
true in some cases. It depends on how contracts are negotiated. Yet your
opposition to the situation is not based on any test to determine on a case
by case basis whether the artist or the record house is getting the dominant
share.
> BTW, Please don't twist my words to make it look like I think people
> shouldn't get paid.
No one's saying that, that I have seen. What they are saying is that
you are not acknowledging the numerous reasonable things for which
they require pay. There's an old joke about a contractor showing up
at a site and surveying a situation and then pounding in one nail and
submitting a huge bill. Upon being questioned about the bill, the
person says "$1 for hitting the nail. The rest is for knowing where
to put the nail." If you can't acknowledge the reasonableness of this
simple concept, that's where you should focus in understanding
people's ire about some of your remarks, ire that I think is quite justified.
> I have stated several times, that I think
> everyone should be paid fairly for whatever useful work they perform.
> I just hate to hear bleating. The world doesn't owe *anyone* a
> living.
A priori, no. But those who have proven their worth? I'm not so sure.
My only concern is that people put back more than they take. If someone
builds 1-2-3 and each user pays a dollar for it, and they make a million
dollars, I'm not so sure they haven't put in more than they've taken...
I think what the world doesn't owe anyone is free software.
> The bottom line is "if you can't make money doing what *you*
> want, then get off your arse and do something else." Most people
> learn that lesson as adolescents.
This same remark could be applied to your earlier remark about how hard it
is for a small business to survive. Others run their small businesses just
fine on the existing set of rules.
> Free software is not going to totally replace Proprietary software in
> all areas. It is going to (and does) give users/developers a real
> choice, rather than the pseudo-choice involved in dealing with
> megacorps (or even just big corps).
I doubt this. I think it will just financially disempower programmers,
leaving all the money in the hands of businessfolk. Then I think the
businessfolk will take computer science in a different direction than
tech people would have because the tech people can't afford to chart their
own course.
> It is a good deal for the guys
> with knowledge but little capital.
I don't see it, since it supplies them with no capital. It does nothing
to free them from their day job. It just allows them to amuse themselves
at the end of their work day doing something someone else dictated for
them. Until you show how it translates into more control of one's would-be
work time, you have not made your case.
> The best Free software is made (in
> the fullest sense of the word) by its users. That is probably what
> really bugs a lot of conservative programmers who still think of
> themselves as members of an elite, exclusive group: Programmers vs
> lusers.
This is so far from the truth that I'm not even going to reply.
> * In message <···············@world.std.com>
> * On the subject of "Re: evolving lisp"
> * Sent on Tue, 22 May 2001 11:24:10 GMT
> * Honorable Kent M Pitman <······@world.std.com> writes:
>
> It is always your option not to license the software. The beauty of
> the legal system of contracts is that their conditions, whatever they
> are, are just about always allowed because they are voluntarily
> entered. It is your option not to enter the contract.
This is not always the case.
It is still quite hard to buy a PC without paying the "MicroSoft tax".
"market failure" is quite common in a monopolised market, and the
_commodity_[1] software market is very monopolistic, not just due to the
(now illegal) MS business practices, but because of its nature - high
fixed costs/negligible marginal costs.
This is why I do not think that it is very productive to compare the
software market to, say, car market.
I think it is more similar to the cable/power/phone markets.
Maybe mandating (_full_) source code release, eliminating the software
patents and anti-reverse-engineering laws, and limiting copyright term
would help?
1. full source release would prevent copyright violations (AKA
plagiarism) - you just compare the sources and see if anything was
stolen (the comparison would be more complicated than just diff(1) or
cmp(1)).
2. the license can be the same as the current "shrink-wrap" licenses - no
redistribution &c. The sources may be compiled - to check that they
are the actual sources of the program.
3. The evaluation copy can come without sources - so that one would not
be able to bypass the copy protection tricks.
I am sure there will be a lot of opposition to this.
Please be aware that I am not in a position to force this upon you (and
even if I were, I would not have done it anyway), so you are safe. :-)
[1] _commodity_, like OS, editor &c, as opposed to "one of a kind"
consulting contracts.
--
Sam Steingold (http://www.podval.org/~sds)
In the race between idiot-proof software and idiots, the idiots are winning.
Sam Steingold <···@gnu.org> writes:
>
> This is not always the case.
> It is still quite hard to buy a PC without paying the "MicroSoft tax".
> "market failure" is quite common in a monopolised market, and the
> _commodity_[1] software market is very monopolistic, not just due to the
> (now illegal) MS business practices, but because of its nature - high
> fixed costs/negligible marginal costs.
Well the solution to that, surely, is anti-monopoly legislation that
works, and is enforced (as far as I can see the US has neither of
these at the moment - the legislation is such that it takes forever to
enforce and the current president doesn't seem too keen on anything
that will hurt the large corporations who funded him). You don't
need any special magic software case, you just have to fix the system
you already have.
--tim
Tim Bradshaw wrote:
>
> Sam Steingold <···@gnu.org> writes:
>
> >
> > This is not always the case.
> > It is still quite hard to buy a PC without paying the "MicroSoft tax".
> > "market failure" is quite common in a monopolised market, and the
> > _commodity_[1] software market is very monopolistic, not just due to the
> > (now illegal) MS business practices, but because of its nature - high
> > fixed costs/negligible marginal costs.
>
> Well the solution to that, surely, is anti-monopoly legislation that
> works, and is enforced (as far as I can see the US has neither of
> these at the moment - the legislation is such that it takes forever to
> enforce and the current president doesn't seem too keen on anything
> that will hurt the large corporations who funded him). You don't
> need any special magic software case, you just have to fix the system
> you already have.
The problem with American anti-trust law is that it is quite literally
impossible for a business to comply with some regulations without
breaking others. There are a number of mutually exclusive requirements
in the tangle of laws we have, which seem to benefit only bureaucrats
and lawyers. Historically, there are no examples of a coercive monopoly
that was not enforced by government edict. And Microsoft is not a
coercive monopoly, though it's dominant, overbearing and repugnant in
some ways. It's generating it's own counterweights in the marketplace,
and I think it will lose or be losing it's dominant position in the
industry (much as IBM did) long before the anti-trust suit finally winds
its way through the appeals process and anything actually _happens_ on
that front.
Larry
Larry Elmore <········@home.com> writes:
> Microsoft is not a coercive monopoly
Of course it is! That's what "intellectual property" is all about:
coercive monopolies settled by governments who pretend
"To promote the progress of science and useful arts,
by securing for limited times to authors and inventors
the exclusive right to their respective writings and discoveries;"
(US constitution, Article I, Section 8, Clause 8).
Just because it's legal doesn't mean it's not coercive.
Quite on the contrary: Just because a piece of paper
signed by a mob ruler in a big city says it's legit
doesn't make it andy righter or wronger. It just makes it coercive.
So it's being legally mandated verily ENTAILS that it's coercive.
Funny how the events in the american colonies started as a protest
against government-enforced monopolies on tea trade, that were
allegedly granted to promote innovation and secure investment
in intercontinental trade.
Nowadays, governments are more clever: never give a nominal privilege;
grant innocent-looking rule-driven privileges
that never explicitly name anyone.
Then once in a while, isolate a one privilege holder, chastise him,
and boast that you are the saviour of the people against
the "problems of Free Markets". And most people buy it,
and ask for more and more government intervention.
Protectionism means a less adaptative, slower, smaller, market,
where everyone loses in the end: monopoly holders make an extra undeserved
buck, but two deserved bucks disappear from the world, and in the long run,
even the monopoly holder has a poorer world in which to spend his buck
and in which to earn more bucks. Free Markets shy away from it.
Just why do you think the market chose the technically crappy PC platform?
Because it was the one least protected by IP.
LISP would win with a Free Market in LISP services.
If Xanalys, Franz, etc, can't make a profitable buck without a monopoly
on their selling, servicing, supporting, etc, the software they write,
then they don't deserve to exist. If they can, then who needs protectionism?
Free Software will but expand the market. Don't fear freedom.
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
Too many people are thinking of security instead of opportunity. They seem
more afraid of life than death.
-- James F. Byrnes
Francois-Rene Rideau <·······@SPAM.tunes.org> writes:
> Larry Elmore <········@home.com> writes:
> > Microsoft is not a coercive monopoly
> Of course it is! That's what "intellectual property" is all about:
> coercive monopolies settled by governments who pretend
> "To promote the progress of science and useful arts,
> by securing for limited times to authors and inventors
> the exclusive right to their respective writings and discoveries;"
> (US constitution, Article I, Section 8, Clause 8).
I guess that you are against privacy protection laws as well, then?
Because the only thing that these laws install are monopolies on the
personal data of your self, allowing only you to control the sale,
processing and use of such data.
Don't fear freedom.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
This is way off-topic from comp.lang.lisp;
Followup-To: ············@tunes.org
or any other more appropriate forum than you want (and announce here).
>>>: Larry Elmore <········@home.com>
>>: Fare Rideau <···········@tunes.org>
>: "Pierre R. Mai" <····@acm.org>
>>> Microsoft is not a coercive monopoly
>> Of course it is! That's what "intellectual property" is all about:
>> coercive monopolies settled by governments [...]
> I guess that you are against privacy protection laws as well, then?
I reckon the natural right of anyone to take whatever measures to
protect one's own privacy, within the limits of one's property:
walls, roofs, curtains, opaque envelopes, strong encryption,
safes, faraday cages, non-disclosure agreements, barbed wire, etc.
I reckon the right of anyone to counter-attack or sue,
when an aggressor trespasses or destroys his property
while trying to invade privacy (or trying to do whatsoever, for the matter).
I deny anyone the right to counter-attack or sue
someone who would manage to gather information
without trespassing, destroying, or breaching anything.
I deny governments or anyone the right either to
invade the privacy of some people when it entails
trespassing or destroying private property,
or to enforce the privacy of other people after the fact,
when these people failed to take appropriate measures to protect themselves.
Actually, I deny governments the right to enact laws of any kind,
or to do anything that citizens couldn't do individually
in an voluntary organized way.
I reckon the validity of common law and case law,
in as much as it hasn't been tainted by forged law.
> Because the only thing that [IP] laws install are monopolies on the
> personal data of your self, allowing only you to control the sale,
> processing and use of such data.
No. They install monopolies on the use of their own resources
by non-contracting third parties.
They are no different in principle than the monopoly on trade with India.
You needn't any monopoly to withhold the your data and your services
until you agree to some conditions of sale with some customer.
Just like you needn't any monopoly to conduct trade with India.
Competition is not piracy. See abundant literature at
http://fare.tunes.org/patents.html#pointers
> Don't fear freedom.
I don't.
Next, know what freedom is.
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
Economics is the plural of Morality.
Francois-Rene Rideau <···········@tunes.org> writes:
> I deny anyone the right to counter-attack or sue
> someone who would manage to gather information
> without trespassing, destroying, or breaching anything.
So you are opposed to privacy protection laws, which at least makes
your argumentation consistent.
> > Because the only thing that [IP] laws install are monopolies on the
> > personal data of your self, allowing only you to control the sale,
> > processing and use of such data.
Please don't misquote me. I was only speaking about privacy
protection laws, not intellectual property laws in my last post.
Anyway, given that you are consistent in your reasoning, I'd be
interested how you defend the use of the GPL, which makes use of
strong IP laws in order to enforce the adherence of 3rd parties to the
original authors wishes.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
"Pierre R. Mai" <····@acm.org> writes:
>>: Pierre Mai
> Francois-Rene Rideau <···········@tunes.org> writes:
>> I deny anyone the right to counter-attack or sue
>> someone who would manage to gather information
>> without trespassing, destroying, or breaching anything.
> So you are opposed to privacy protection laws, which at least makes
> your argumentation consistent.
Thank you for acknowledging this consistency.
Apparently, not everyone is willing to look deep enough to see that.
>>> Because the only thing that [IP] laws install [...]
> Please don't misquote me. I was only speaking about privacy
> protection laws, not intellectual property laws in my last post.
Oops. My misunderstanding. I stand corrected.
> Anyway, given that you are consistent in your reasoning, I'd be
> interested how you defend the use of the GPL, which makes use of
> strong IP laws in order to enforce the adherence of 3rd parties to the
> original authors wishes.
Free Software is negation of Intellectual Property.
Free Software licenses are basically agreement, whereas authors say:
"look, I'm not going to enforce any iniquitous privilege against you!"
The softer licenses (MIT/X/amended BSD) add:
"if you too come and play the game with us, we'll have fun;
if you come and don't play fair, we will mostly just ignore you"
The harder licenses (GNU GPL), add "you can join and have fun,
but don't bother coming if you don't intend to play fair".
Indeed, the GNU GPL mostly wouldn't be enforceable in absence of IP
(in which case everything would basically be as if BSD licensed);
more precisely, the one relevant clause that would become unenforceable
(whereas all the others will be redundant with natural rights)
is the clause about the need to public sources alongside binaries.
However, in presence of such IP privileges, the GNU GPL is
an aggressive message clearly delimiting a community that fights
Information Protectionists back with their own weapon, whereas
BSD-style licenses are more like pacific messages
that either think or hope that IP is not a problem,
or that it will be solved just by letting things happen
"just you hit us with your privileges," "I don't care",
"it doesn't hurt", "it doesn't matter", or otherwise "you'll lose anyway".
I think it is sane that both attitudes exist;
people will choose depending on their temperament.
However, I think multiplying mutually incompatible licenses
(all this NPL, MPL, QPL, APSL, etc.), is just WRONG,
all the more so since it prevents any kind of advanced (manual or automatic)
metaprogramming that'd mix legally incompatible programs (see my ll99 paper).
Some years ago, I was leaning toward using the GNU GPL or LGPL.
These days, my favorite license is bugroff
http://www.geocities.com/SoHo/Cafe/5947/bugroff.html
Sorry you didn't choose to heed the followup to ············@tunes.org
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
With Information Protectionism, people make money by slowing down the flow
of information. With Free Information, people make money by accelerating it.
Francois-Rene Rideau <·······@SPAM.tunes.org> writes:
> Larry Elmore <········@home.com> writes:
> > Microsoft is not a coercive monopoly
> Of course it is! That's what "intellectual property" is all about:
> coercive monopolies settled by governments who pretend
> "To promote the progress of science and useful arts,
> by securing for limited times to authors and inventors
> the exclusive right to their respective writings and discoveries;"
> (US constitution, Article I, Section 8, Clause 8).
Coercive or not, the fact remains that copyright is the one tool that
allows me to stand up to Microsoft's obvious desire to acquire the
works of others and redistribute those works for its own gain.
Ignoring the issue of whether anything I personally might write is of
suitable quality to interest Microsoft (let's assume for argument's
sake that it is), I can register my own works of authorship (for
example, "fiction writings" or "graphical icons" or "java applets" or,
if I may be permitted to speak "on topic" just for a moment--"lisp
libraries") with the US government copyright office, publish them on
the web, and feel secure that if Microsoft tries to take those things
from me against my copyright notice, I can not only sue for actual
damages (probably minor) but for statutory damages (huge enough to get
the attention even of Microsoft) AND for attorney's fees (so it's
worth a lawyer going after on contingency). Although people often see
companies apply copyright, it is the right of everyone to apply, and
it is substantially more powerful when used by the little guy against
the big guy than vice versa. Copyright is a powerful tool of
protection of individual freedom.
Even Stallman&co had to acknowledge this implicitly (and perhaps
explicitly as well, for all I know) when they had to use the powerful
tool of copyright as an implementation vehicle to enforce the GPL and
other legal vehicles they have created. It was the ability to control
the exact use of a work which allowed them the power to do what they
needed to do without new legislation.
Coercive? Of course. If what you mean is that it's meaningless to
make a law against an infraction without having teeth that allows you
to enforce that infraction. Laws are by nature coercive. The additional
explicit use of the word seems to me just scare tactics. "Oooooh, it's
not just a law, it's a coercive law...." Shrug. I'm happy to discuss
the effects of legislation but not to blur them.
The whole business of Microsoft being a monopoly is an issue of
anti-trust law, not an issue of IP law. The issue of the OS-tax,
where you can't buy a machine without Windows on it, is an anti-trust
issue. If MicroSoft were MicroHard, and it sold hammers, they'd
pobably have made a deal with tool belt vendors or vendors of boxes of
nails that made it impossible to get those items without also getting
a MicroHard hammer, Sure, you could fix the problem by eliminating
their ability to own something, but that would have other side-effects
as well--like removing the small author/artist/programmer's ability to
control their work. You could fix my imaginary MicroHard scenario by
removing the ability to own real property, too, but that would mean a
lot of looting would happen elsewhere in the world. Anti-trust
problems should be fixed by anti-trust law. A monopoly, however
implemented, hurts the market. The correct fix is to enforce the
existing laws where applicable, or to make new legislation that is
targeted at that problem--not to throw the baby out with the
bathwater.
[Were it me, I would fix the problem in the following way that would
not lend itself to subjective quibbling about vertical markets or good
of the customer: (a) bound the maximum size of a company; when it has
net worth more than a certain amount, it must split; (b) require that
no one human being own majority stake in more than one company; they
may continue to invest, but only as minority partners. The real
problem, it seems to me, is not property, but the singular control of
too much property by one person. One person is just not wise enough
to control everything. If lots of people were in control of medium
size things, the market would take care of itself. Another problem
that makes the present situation unstable is that a few vendors have
gotten so large as to be "obvious" investment choices, causing people
to invest in them because they know the entity will be protected from
catastrophe (as with the FDIC) rather than because of a belief they're
doing good. As long as nothing gets to be so large as to need to be
protected from failure, people are forced to think about where to put
their money, or to diversify to reduce risk. The paradox of today's
market is that diversification increases risk; there's just going to
be one or two winners and throwing your money after much of anyone
else looks like a loss. Well, perhaps fortunately for all I know,
it's not up to me, so we continue with the system we have.]
On Wed, 23 May 2001 13:19:50 GMT,
"Kent" == Kent M Pitman <······@world.std.com> wrote:
Kent> issue. If MicroSoft were MicroHard, and it sold hammers, they'd
Kent> pobably have made a deal with tool belt vendors or vendors of boxes of
Kent> nails that made it impossible to get those items without also getting
Kent> a MicroHard hammer
For MicroHard, read Shimano (the bicycle parts producers, who have fairly
effectively established a MicroSoft-style monopoly using MicroSoft-style
coercive practices, apparently)
Philip
--
If you have a procedure with 10 parameters, you probably missed some. (Kraulis)
-----------------------------------------------------------------------------
Philip Lijnzaad, ········@ebi.ac.uk \ European Bioinformatics Institute,rm A2-08
+44 (0)1223 49 4639 / Wellcome Trust Genome Campus, Hinxton
+44 (0)1223 49 4468 (fax) \ Cambridgeshire CB10 1SD, GREAT BRITAIN
PGP fingerprint: E1 03 BF 80 94 61 B6 FC 50 3D 1F 64 40 75 FB 53
Kent M Pitman <······@world.std.com> wrote in message news:<···············@world.std.com>...
> [Were it me, I would fix the problem in the following way that would
> not lend itself to subjective quibbling about vertical markets or good
> of the customer: (a) bound the maximum size of a company; when it has
> net worth more than a certain amount, it must split; (b) require that
> no one human being own majority stake in more than one company; they
> may continue to invest, but only as minority partners.
I think this plan is excellent. Instead of a specific dollar-value limit
on the size of a company, cap of the company's net worth to some fraction
of the GDP might be more appropriate. Part b is a thought that had been
incubating in the back of my head, but I only just realized that when you
articulated my thought so well.
But I think the real problem is that American businesses, at the least,
seem to me to be focused on following the crowd (I suppose the same could
be said for individuals). It's horribly ironic when you also think about
the mistaken idea that the US is a country that values individualism. Right
now, I'm not sure at all as to what the US values, as a society... I
suppose it would be money, which would be supported well by the way that
big business is dominating the economy. People see lots of money and they
flock to it. Or maybe I'm just being a naive youngster. :)
--
-> -/- - Rahul Jain - -\- <-
-> -\- http://linux.rice.edu/~rahul -=- ·················@usa.net -/- <-
-> -/- "I never could get the hang of Thursdays." - HHGTTG by DNA -\- <-
|--|--------|--------------|----|-------------|------|---------|-----|-|
Version 11.423.999.220020101.23.50110101.042
(c)1996-2000, All rights reserved. Disclaimer available upon request.
·····@rice.edu (Rahul Jain) writes:
> Kent M Pitman <······@world.std.com> wrote in message news:<···············@world.std.com>...
> > [Were it me, I would fix the problem in the following way that would
> > not lend itself to subjective quibbling about vertical markets or good
> > of the customer: (a) bound the maximum size of a company; when it has
> > net worth more than a certain amount, it must split; (b) require that
> > no one human being own majority stake in more than one company; they
> > may continue to invest, but only as minority partners.
>
> I think this plan is excellent. Instead of a specific dollar-value limit
> on the size of a company, cap of the company's net worth to some fraction
> of the GDP might be more appropriate. Part b is a thought that had been
> incubating in the back of my head, but I only just realized that when you
> articulated my thought so well.
>
> But I think the real problem is that American businesses, at the least,
> seem to me to be focused on following the crowd (I suppose the same could
> be said for individuals). It's horribly ironic when you also think about
> the mistaken idea that the US is a country that values individualism. Right
> now, I'm not sure at all as to what the US values, as a society... I
> suppose it would be money, which would be supported well by the way that
> big business is dominating the economy. People see lots of money and they
> flock to it. Or maybe I'm just being a naive youngster. :)
Well, American liberty began with nothing and had no experience with
the far extreme. We want to say, as an ideal, that individual liberty
should be limitless. The risk is of the extinguishing of liberty, so
we are careful to put no bounds on it. (Interesting parallels with
the "be fruitful and multiply" remark in the bible, that has led to
overpopulation.) However, in a finite container, such as the Earth,
there comes a point where telling a person they can be wealthy and
controlling without bounds is the same as telling other people they
cannot be, and that's where it gets complicated.
I find myself taken back to virtue ethics, which I missed in my
education growing up and only ran across in web searches for related
topics and found fascinating for their insight. Aristotle's "mean"
between unreasonable extremes (of excess and defect)--the notion that
the the things we find virtuous are, invariably, not merely the
finding of a compass direction and the blind trodding as fast as one
can in that direction, but rather the balancing of perpetual tugs from
incompatible forces going in different directions. Courage, for
example, being a mean between rashness and cowardice; Generosity being
a mean between wastefulness and stinginess. This topic is copiously
studied on the net, but I suggest, somewhat at random after 10-15
minutes' browsing, http://www.philosophypages.com/hy/2s.htm
as a nicely summed up example for those not familiar and wanting to read
a little more.
(I find myself wondering if "free software", having no notion of
"moderation", "harmony", or "balnce" built into it will or does have
similar problems as those observed in unbounded capitalism.)
Even free speech, which for various reasons I prefer to frame as "the
right to hear", I now regard as a balance betwen the right to know
what may affect you and the right of others that you not know what may
affect them. Of course, that will never be a good quote, like
Aristotle's remarks, until someone picks some more pithy one-word ways
of expressing those various virtues...
In article <···············@world.std.com>,
Kent M Pitman <······@world.std.com> wrote:
>
>Well, American liberty began with nothing and had no experience with
>the far extreme.
Historically, I don't think this is accurate. There was liberty
in the colonies, and the Founding Fathers had good classical
educations, and theoretically some knowledge of some of the
classical excesses.
>overpopulation.) However, in a finite container, such as the Earth,
>there comes a point where telling a person they can be wealthy and
>controlling without bounds is the same as telling other people they
>cannot be, and that's where it gets complicated.
>
This is one very large difference between the early US and the
current situation. Early on, if you were free, you had a choice
between putting up with whatever government you were under or
moving West and setting up anew. Nowadays, there is no similar
way to opt out (as opposed to drop out). (Of course, this
freedom was gained at the expense of the natives.)
And, yes, as part of student government I have received complaints
that US women living in dorms were sunbathing in bikinis and offending
students from the Middle East. If you define somebody else's behavior
as impinging on yourself (such as I do with people who smoke around
me) then somebody's going to have liberties infringed.
>I find myself taken back to virtue ethics, which I missed in my
>education growing up and only ran across in web searches for related
>topics and found fascinating for their insight.
FWIW, there are similar things in some more Eastern religions and
philosophy. The I Ching is full of statements about what the
superior man would do, expressed in terms of virtues. You may
want to look into some of that.
>(I find myself wondering if "free software", having no notion of
>"moderation", "harmony", or "balnce" built into it will or does have
>similar problems as those observed in unbounded capitalism.)
>
Most people that I know that write free software do have notions
of moderation, harmony, and balance. The GPL is based on notions
of harmony and balance. Some of the people who advocate free
software are not in fact moderate, harmonious, or balanced,
and I fear you've been paying too much attention to them.
And, of course, the extreme free software viewpoint that is devoid
of moderation does have problems. It removes most of the
incentive to write mass-market software backed by usability studies,
to give one example.
>Even free speech, which for various reasons I prefer to frame as "the
>right to hear", I now regard as a balance betwen the right to know
>what may affect you and the right of others that you not know what may
>affect them.
Could you explicate? I don't understand the balance. I think I
understand part of what you mean as "the right to hear", but probably
not all.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
········@visi.com (David Thornley) writes:
> In article <···············@world.std.com>,
> Kent M Pitman <······@world.std.com> wrote:
> >
> >Well, American liberty began with nothing and had no experience with
> >the far extreme.
>
> Historically, I don't think this is accurate. There was liberty
> in the colonies, and the Founding Fathers had good classical
> educations, and theoretically some knowledge of some of the
> classical excesses.
Oh, I just knew that statement was going to get me in trouble and I thought
about working harder on it but didn't. What you say is right, of course.
I had really meant they had no experience with liberty running so out of
control as to trample other people. I think only sometime in the 20th
century, accompanying the so-called "death of shame" did the phenomenon
to which I'm referring even exist. Though I probably explained myself
badly. [Now you might still claim I was wrong in thinking this part,
but hopefully at least then I'll be corrected on what I did mean. :-]
> >[overpopulation]
> This is one very large difference between the early US and the
> current situation. Early on, if you were free, you had a choice
> between putting up with whatever government you were under or
> moving West and setting up anew. Nowadays, there is no similar
> way to opt out (as opposed to drop out). (Of course, this
> freedom was gained at the expense of the natives.)
I think this, not just scientific curiosity, underlies the reason many
of us hate seeing the NASA budget cut...
> >Even free speech, which for various reasons I prefer to frame as "the
> >right to hear", I now regard as a balance betwen the right to know
> >what may affect you and the right of others that you not know what may
> >affect them.
>
> Could you explicate? I don't understand the balance. I think I
> understand part of what you mean as "the right to hear", but probably
> not all.
http://world.std.com/~pitman/pfaq/freedom-to-hear.html deals with this
in slightly more detail, though not as much as I'd like to if I had more
time to write on it. I don't want to drag down this newsgroup overly on
philosophy, so I'll correspond with people privately if they find it
leaves open questions, and I'll update the pfaq entry as appropriate to
summarize such correspondence, so those who just want to read the result
can check back in a week or two to see if anything got added.
········@visi.com (David Thornley) writes:
> In article <···············@world.std.com>,
> Kent M Pitman <······@world.std.com> wrote:
...
> >overpopulation.) However, in a finite container, such as the Earth,
> >there comes a point where telling a person they can be wealthy and
> >controlling without bounds is the same as telling other people they
> >cannot be, and that's where it gets complicated.
> >
> This is one very large difference between the early US and the
> current situation. Early on, if you were free, you had a choice
> between putting up with whatever government you were under or
> moving West and setting up anew. Nowadays, there is no similar
> way to opt out (as opposed to drop out). (Of course, this
> freedom was gained at the expense of the natives.)
>
Curious that the fate of the Native Americans should pop up in this
thread. Here is what Stallman said about how he felt after Symbolics
pillaged the MIT AI lab: "I felt that I was the last survivor of an
extinct tribe, doomed to spend my life among uncomprehending
strangers."
http://www.lysator.liu.se/history/garb/txt/87-2-rms.txt
Free software is an attempt to create the conditions under which a
similar tribe can flourish.
Regards,
Peter
Ps. If I remember correctly, one source of great conflict between the
settlers and Native Americans was mutual misunderstanding about what
it was reasonable to claim *ownership* of.
Kent M Pitman <······@world.std.com> writes:
> (b) require that
> no one human being own majority stake in more than one company; they
> may continue to invest, but only as minority partners.
Does this rule out a mostly-owned company buying other companies? If
not it's meaningless (everyone who wants to own more than one company
will simply start a holding company which will own others).
(Your (a) sounds pretty good though, possibly with some
fraction-of-GDP modification)
--tim
Tim Bradshaw <···@tfeb.org> wrote in message news:<···············@tfeb.org>...
> Kent M Pitman <······@world.std.com> writes:
>
> > (b) require that
> > no one human being own majority stake in more than one company; they
> > may continue to invest, but only as minority partners.
>
> Does this rule out a mostly-owned company buying other companies? If
> not it's meaningless (everyone who wants to own more than one company
> will simply start a holding company which will own others).
I think treating a company with the same rules as you would an individual
(I beleive that's how US law works already) fixes that problem. Or better
yet, the law could apply to the indirect stakes one has in other companies,
as well.
--
-> -/- - Rahul Jain - -\- <-
-> -\- http://linux.rice.edu/~rahul -=- ·················@usa.net -/- <-
-> -/- "I never could get the hang of Thursdays." - HHGTTG by DNA -\- <-
|--|--------|--------------|----|-------------|------|---------|-----|-|
Version 11.423.999.220020101.23.50110101.042
(c)1996-2000, All rights reserved. Disclaimer available upon request.
Kent M Pitman <······@world.std.com> writes:
> Peter Wood <··········@worldonline.dk> writes:
>
> > I don't think the leasing model is appropriate for software.
>
> Buying is out. If I sell it to you, as a matter of law, *I* don't own
> it any more. The only legal mechanism that can work where information
> is copied is licensing. Licensing is done by dictating use but
> retaining ownership. No software developer, other than a contractor
> being paid a large wage for a one-shot deal, will ever give up his
> rights to the first customer.
It depends on what you are selling. The GNU people do exactly that.
Because their stuff is GPL'd they can sell something and keep it.
Then the buyer can also sell it *and keep it*. That's because
software is not really like anything else. It's mostly like ideas,
which are better shared than hoarded.
>
> > I can return a flat or a car. I can't return software in the same
> > way, because I *never took it away*.
>
> It is always your option not to license the software. The beauty of the
> legal system of contracts is that their conditions, whatever they are,
> are just about always allowed because they are voluntarily entered.
> It is your option not to enter the contract.
>
That's what we are supposed to believe, yes. In reality the balance
of power is such that the little guy often does *not* have an option.
Fortunately, Free software is changing that.
> > > If they wanted to do that, that wouldn't suck, it would be legally
> > > untennable. Yet they don't sell you their software (at least not
> > > normally), they grant you certain rights to it, in exchange for
> > > money. If the rights you obtain aren't to your liking, you are of
> > > course free to negotiate a different contract if you can afford to.
> >
> > "If you can afford to"? In practice, the small guy is powerless
> > against the cut-throat megacorp.
>
> Nonsense. If enough people felt this way, there would be a market for
> another mode of business.
Ah, but there is a market for another mode of business. It's called
Free Software.
> > So you think that when a musician has played a concert, he should
> > *also* be payed next year for that concert?
>
> Yes, I do, actually. Because if you don't, then they will be playing
> this same concert next year instead of new music.
No they won't. Even the Pop public gets tired eventually of hearing
the same crap again and again.
> > Artists (musicians,
> > authors, etc) should get payed fairly for the work they do. The
> > system today rewards the record companies more than the musicians.
>
> This is a red herring. Live with it. You will not invent a system
> that does not unfairly reward someone except a system that rewards no
> one.
>
> > Thats a fact. OTOH *Noone*, IMO, is justified in feeling that 3 weeks
> > of work (or 3 months or 3 years) entitles them to a lifetime of
> > income.
>
> I'm going to post a response to this under separate cover. This statement
> is based on a HUGE fallacy about how the capitalistic market process works.
> It sounds very persuasive rhetorically but really is inappropriate.
>
> (And, incidentally, I think it is false on its face. If Einstein or Salk
> or Mozart did all their work in 3 weeks or 3 years, they are absolutely
> entitled to a lifetime of pay because of the benefit others received.
> Moreover, in some situations, the doing of the deed burns the person out.
> Ballplayers get a lot of money in part because their career depletes them
> physically by a young age and they must live on that money the rest of
> their lives. The same might be said for war veterans, lamentable as war
> is.)
I think both of your examples are factually wrong. Only young(ish)
people have the physique and reflexes to handle elite sport. Your
ball players aren't burned out - just aging ordinarily. War vets who
can't cut it often are suffering from treatable psychological and/or
physical disorders. If they were properly treated by their country,
they would have as good a chance as anyone else to earn a living.
>
> > If they can work it so they actually get a lifetime of
> > income, well congratulations to them! But I won't accept bitching and
> > whining when they can't work it that way.
>
> So your sense of appreciation for what someone does for you is not based on
> the degree of value you have received? That seems callous.
You know nothing about my sense of appreciation for what people do for
me. However I do insist on my right to decide what is useful for me.
I reserve the right not to feel gratitude for something I can't use.
>
> > The publishing house (or record company) controls the means of
> > realizing the artist's work,
>
> You're really fixated on this, but it is irrelevant.
No I think its important. It's one of the things which appears to be
changing. If it weren't important, artists *certainly* wouldn't pay
publishers. The musicians I know think it's pretty important and
would *really* like to be independant. New technology makes very
small runs of books feasible (cheap and easy). If that becomes
widespread, authors will be liberated from publishers. Of course,
promotion of works will still be an issue which they will have to pay
for.
> > BTW, Please don't twist my words to make it look like I think people
> > shouldn't get paid.
>
> No one's saying that, that I have seen.
Please reread the post I was responding to.
> What they are saying is that you are not acknowledging the numerous
> reasonable things for which they require pay.
This is laughable. You have made a straw man. I certainly am *not*
obliged to acknowledge anyone's right to require pay for what they
have done. Wether or not a product is worth anything is entirely
determined by wether it's needed and useful. If you can't get money
for what you have made, then you have misjudged the market.
> If you can't acknowledge the reasonableness of this simple
> concept, that's where you should focus in understanding people's ire
> about some of your remarks, ire that I think is quite justified.
>
I have no problem with people getting hot under the collar about my
remarks.
> I think what the world doesn't owe anyone is free software.
Of course it doesn't. I don't *care* wether you release your stuff
for free or make it proprietary. I *DON'T CARE* . If there's
something I need and its not available I make it myself. That is not
what my posts have been about. I *object* to your claim that Free
software is poisoning your well, because behind that claim is the
belief that you have some right to sell your stuff. You don't. You
have the right to try to sell it (or do whatever else you godamn
please). Your righteous indignation at people who "give stuff away"
is unjustified. People give stuff away for lots of reasons. One of
which is that they might get it back improved.
Regards,
Peter
Peter Wood <··········@worldonline.dk> writes:
>
> I think both of your examples are factually wrong. Only young(ish)
> people have the physique and reflexes to handle elite sport. Your
> ball players aren't burned out - just aging ordinarily.
Rubbish. People who do serious sport get damaged by it. For God's
sake even *I* have knee damage from when I used to ride a lot.
> War vets who
> can't cut it often are suffering from treatable psychological and/or
> physical disorders. If they were properly treated by their country,
> they would have as good a chance as anyone else to earn a living.
As far as I'm aware having your legs blown off is not, yet, treatable.
--tim
Tim Bradshaw <···@tfeb.org> writes:
> Peter Wood <··········@worldonline.dk> writes:
>
> >
> > I think both of your examples are factually wrong. Only young(ish)
> > people have the physique and reflexes to handle elite sport. Your
> > ball players aren't burned out - just aging ordinarily.
>
> Rubbish. People who do serious sport get damaged by it. For God's
> sake even *I* have knee damage from when I used to ride a lot.
>
1) Not even close to everyone who does serious sport gets damaged by
it.
2) The risks you run are *your* risks, not mine.
3) The people who do get damaged often do so because they try to milk
the situation for all its worth. So they continue with their sport
long after their bodies are capable of dealing with the wear and
tear.
> > War vets who
> > can't cut it often are suffering from treatable psychological and/or
> > physical disorders. If they were properly treated by their country,
> > they would have as good a chance as anyone else to earn a living.
>
> As far as I'm aware having your legs blown off is not, yet, treatable.
>
Most war vets don't have their legs blown off. And, of course, those
who do should be compensated appropriately. Civilised practice would
be for their employer to be economically responsible for the damage
they (or their families) suffer. But don't expect civilised practice
in the US or UK any time soon.
(I did use the word "often" to cover cases like this, but I see you
are another Straw Man Arguer.)
Regards,
Peter
Peter Wood <··········@worldonline.dk> writes:
>
> 1) Not even close to everyone who does serious sport gets damaged by
> it.
>
> 2) The risks you run are *your* risks, not mine.
>
> 3) The people who do get damaged often do so because they try to milk
> the situation for all its worth. So they continue with their sport
> long after their bodies are capable of dealing with the wear and
> tear.
>
Actually for physical sport the porportion is pretty high (note I
count stuff like arthritis due to joint wear which is a very
significant problem).
But this is beside the point: if I take a risk of something which may
prevent me being able to work (or equivalently I concentrate on
winning at the the olympics rather than building a career which will
pay me well later) I absolutely expect to be compensated for that
risk. That's what I think KMP was saying. But I don;t really expect
you to understand, unfortunately.
--tim
Tim Bradshaw <···@tfeb.org> writes:
> But this is beside the point: if I take a risk of something which may
> prevent me being able to work (or equivalently I concentrate on
> winning at the the olympics rather than building a career which will
> pay me well later) I absolutely expect to be compensated for that
> risk. That's what I think KMP was saying.
You sure should. But if you know or ought to know you're taking the risk,
then you cannot sue anyone for compensation after the fact:
it is your responsibility to negociate it beforehand.
In other words, nobody owes you compensation but yourself.
If you can't find compensation, don't take the risk.
If there's enough demand for your taking the risk,
then get paid accordingly and take the compensation into account.
That's what responsible investment is all about:
taking risks that you believe are worth taking.
Spending unpaid time writing software is an investment.
If you think it isn't worth it, just don't do it;
no one owes you a profit off it.
> But I don't really expect you to understand, unfortunately.
I think the person you replied to understands perfectly.
I recommend this thread moves away if it goes more off-topic than it is
(the only remaining link to LISP being the reference to KMP).
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
There has grown up in the minds of certain groups in this country the notion
that because a man or a corporation has made a profit out of the public for a
number of years, the government and the courts are charged with the duty of
guaranteeing such profit in the future, even in the face of changing
circumstances and contrary public interest. This strange doctrine is not
supported by statute nor common law. Neither individuals nor corporations have
any right to come into court and ask that the clock of history be stopped, or
turned back, for their private benefit. That is all.
-- Robert A. Heinlein ("Life-Line")
Francois-Rene Rideau <···········@tunes.org> writes:
>
> I recommend this thread moves away if it goes more off-topic than it is
> (the only remaining link to LISP being the reference to KMP).
>
I certainly have no further interest in discussing it with you or
Peter Wood, so yes, let's let it die, please.
--tim
I've been asked by some why I'm engaging in this conversation, given
that I'm doubtless not going to change the minds of those I'm
debating.
I'll probably wind down shortly, but the reason I wanted to take on the
issue was because I feel that the open/free software position is often
advanced as unambiguously good. It cannot possibly harm, etc. etc.
I'm scared by dogma. I think people should adopt thoughtful positions
on things, not merely look to this or that position and take it because it
seems like what they should do, without thinking.
I'd be as scared if anyone said "oh, I'll adopt Kent's position just
because Kent says so" as I would if you fill in the names of some of
my debate opponents here.
My point has been and will continue to be this:
Open/free software might have their good effects, but also their ill
effects. Certainly the same is true of Copyright. On balance, I
prefer Copyright, and not by some small margin. I don't disagree it
has sometimes ill effects, and I even take notes to myself about those
ill effects. I'd like to see those things repaired. I think giving
creative works of fiction a long timeline is good, because book
authors can take a long time to recover their investment in a book; I
don't think software needs multi-decade protection, and I certainly
don't think patent law has a lot of application to software. But I'll
still defend the system. And I hope people hear both my support for
the things I like and my concern about tuning the parts that aren't
perfect. I don't see that from many on the other side of the debate,
and that's how I find myself in these debates. I see them tell me the
world will be obviously and unambiguously good with their system.
This discussion started because someone suggested (or, at least, I
perceived that they suggested) that there was no possible way that
free software could injure the makers of non-free software. I've made
a case that this is not always so.
If people have gotten anything out of what I've said, I hope it's that
there are no easy answers here. Free software offers a kind of power,
at the possible expense of loss of another kind of power. It ought not
be accepted unquestioningly any more than any other system.
I use free software sometimes because it provides me with things the
market system doesn't; I'm sure there are proponents of free software
who use market-supplied software because the free software system has
failed to meet their needs. It's not a perfect world.
I'm just tired of being portrayed as the one with the black hat,
and seeing the other guys portrayed as the white hat people (in which
group I include those wearing red hats). I'd feel a lot happier if
we were all seen to be wearing gray hats, and we just left it at that.
We're all in search of an ethical way to behave, I think. But ethics
is a funny thing. I don't think you ponder an ethical question for a
long time with the goal of arriving at an ethical answer. Rather, I
think you remain ethical so long as you endure the burden of knowing
there is no one general answer, and as long as you continue to worry
about whether your interim actions are "the best you can do". At the
point where you've dismissed the problem as either solved or
unsolvable, you have ceased to be ethical because the very essence of
ethical behavior is caring, and the very essence of thinking you've
reached an answer is the cessation of caring.
Kent M Pitman <······@world.std.com> writes:
> I've been asked by some why I'm engaging in this conversation, given
> that I'm doubtless not going to change the minds of those I'm
> debating.
The goal of a rational debate is not to convince the other party,
but to construct a solid net of argumentative theory and counter-theory.
Change of mind is something that might or not happen, after the fact,
as a side effect, if the argumentative net has extended far enough
to reach and change the premises of either party. But the net has
a utility of its own even in absence of such possible change of mind.
For instance, some time ago, I had instructive debates about gun control,
and it was a challenging experience exploring the ins and outs of the issue.
Only long after the email exchange, and with complementary information from
other points of view, did I change my mind.
> I'm scared by dogma.
I reject the naive notion that opinions shouldn't be strongly held,
lest they be "dogma". Dogmatism vs rationality is about the processes
by which one makes and changes opinions, not about the strength of these
opinions. Cf. KR Popper's theories about all knowledge being but
conjectures, anyway, and the important part being the conjecture
filtering process.
Merely having an open mind is nothing; the object of opening the mind,
as of opening the mouth, is to shut it again on something solid.
-- G.K. Chesterton
> I think people should adopt thoughtful positions
> on things, not merely look to this or that position and take it because it
> seems like what they should do, without thinking.
Don't underestimate people who do not share your positions.
We might disagree with you and think you're misled,
but we don't question your rationality, honesty, etc.,
and we still very much appreciate your opinions,
and the way you defend them.
> I'd be as scared if anyone said "oh, I'll adopt Kent's position just
> because Kent says so" as I would if you fill in the names of some of
> my debate opponents here.
I have no shame in stating that, like possibly most lurkers on c.l.l,
"oh, I'll ponder to Kent's position just because Kent says so".
You're on my (and likely most lurkers') livefile (anti-killfile?),
and will remain there as long as you post your usual knowledge and insight.
> My point has been and will continue to be this:
>
> Open/free software might have their good effects, but also their ill
> effects. Certainly the same is true of Copyright.
> On balance, I prefer Copyright, and not by some small margin. [...]
>
We know your position, and we appreciate when you argue it,
even when we think you're ultimately wrong.
> And I hope people hear both my support for
> the things I like and my concern about tuning the parts that aren't
> perfect.
Nobody ever claimed that freedom yields an immediate or differed
perfection in its outcome. Rather, depending on one's using
the idealist or the utilitarian point view (which are actually isomorphic),
we claim the perfection of freedom is about the respect of an absolute
moral principle, or the respect of a practical principle that cannot be
violated without creating at the same time injustices and inefficencies.
On the topic of imperfection and freedom, I particularly recommend
paragraphs Y.55 to Y.63 of this great book on freedom (now online):
http://www.econlib.org/library/Bastiat/basHar0.html#To%20the%20Youth%20of%20France
> I don't see that from many on the other side of the debate,
> and that's how I find myself in these debates.
In as much as we appreciate your moderation as a sign of personal greatness,
it is not in itself and cannot be an argument for or against any opinion.
> This discussion started because someone suggested (or, at least, I
> perceived that they suggested) that there was no possible way that
> free software could injure the makers of non-free software. I've made
> a case that this is not always so.
I agree that some people might be less well because they or other people
write free software than they would if they could benefit from an undisputed
copyright monopoly. That doesn't entitle them to claim any damage.
And that doesn't mean that they can't live a decent worker's life
with or despite free software.
> If people have gotten anything out of what I've said, I hope it's that
> there are no easy answers here.
Things are not always simple, but they are not always complex.
Dismissing arguments and debates because "things are complex" is the
ultimate negation of reason, and appeal to what remains: violence.
> [Free software] ought not
> be accepted unquestioningly any more than any other system.
Who says it is accepted unquestioningly or should?
Excuse me, but that's a straw man. There sure are dogmatic such tenants;
just like there are dogmatic tenants of copyright.
But I think you deeply misunderstand free software
if you can't see any of the rational argumentation behind it.
I don't mean for you to accept this argumentation, much less on face value;
but to understand it (or them -- not all defend free software
out of the same basic principles).
For instance, I have a friend who, despite not sharing my libertarian ideas,
can reasonably predict my argumentation on most political topics, because
he understands the coherence deep down in them, independently from their
validity. That's all that rational debates are or should be able:
communicating not just raw bits of data, but vast coherent structures.
> I use free software sometimes because it provides me with things the
> market system doesn't; I'm sure there are proponents of free software
> who use market-supplied software because the free software system has
> failed to meet their needs. It's not a perfect world.
I'm sorry, but that's a _complete_ straw man that shows
your ignorance or deep misunderstanding of your opponent's opinions.
There is no opposition between free software and the market system.
Free software is not against the market system;
not anymore than e.g. abolition of slavery was against the market system.
Just because some believe that ownership and hence trade of individuals
(all the more justified by conquest and heredity) shouldn't exist doesn't
mean that they believe ownership and trade in general shouldn't exist.
Free Software is perfectly compatible with a free market in a variety
of services, from development to packaging to retail to support, etc.
Free Software opposes the use of Intellectual Property to protect
monopolies on developing, copying, using, and distributing software.
The notion of selling services for (free) software development
is at least as old as the GNU Manifesto. Actually, if you accept that
software is but one branch of selling creative thinking, opponents of
Intellectual Property and Industrial Property back in the nineteenth century
have long argued how creators can make money without monopolies,
by constantly working to get an edge over competition, or
selling their services to people who will use them so.
> I'm just tired of being portrayed as the one with the black hat,
> and seeing the other guys portrayed as the white hat people (in which
> group I include those wearing red hats). I'd feel a lot happier if
> we were all seen to be wearing gray hats, and we just left it at that.
We're not putting hats on *people*, but on *opinions*.
A greatest man on earth can defend wrong opinions.
The worst bandit on earth can defend correct opinions.
If Saint Francis says earth is flat, and Hernan Cortes says it is round,
it doesn't make the earth any flatter, Saint Francis any less saint,
or Hernan Cortes any less of a bogeyman.
In other words, I may be a genius or an utter moron, that doesn't make LISP
any better or worse when I use it, defend it, breath it (or attack it, etc.).
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
To appreciate an intelligent man is to have some pleasure when one agrees
with him, and to have an instructive pleasure when one does not.
Francois-Rene Rideau <···········@tunes.org> writes:
> I agree that some people might be less well because they or other people
> write free software than they would if they could benefit from an undisputed
> copyright monopoly. That doesn't entitle them to claim any damage.
Well, that's a subjective position. Maybe not legal damage. But
surely there is an adverse impact on their planned way of life.
Let me risk trying to pull this back onto the original topic for a
moment by citing the specific issue that's of concern to me. I'll use
the first person pronoun, but please keep in mind we're using me only as
an example. The real question is what to do with people and situations
of this general kind...
We have a lot of tools we need written. Who's going to write them and what's
going to motivate them? Is this black magic? Do I just hope? Does crossing
my fingers, knocking on wood, or some such thing speed up the process?
I understand what it means when I say "get out your wallets and start
paying for what you value". It might have no effect, but hopefully it
at least improves the chances someone will pay. But if you say, "put
your wallets away and stop paying content-producers for value you
receive". What follows after that exactly?
>>: Francois-Rene Rideau <···········@tunes.org>
>: Kent M Pitman <······@world.std.com>
>> I agree that some people might be less well because they or other people
>> write free software than they would if they could benefit from an
>> undisputed
>> copyright monopoly. That doesn't entitle them to claim any damage.
> Well, that's a subjective position. Maybe not legal damage. But
> surely there is an adverse impact on their planned way of life.
Adverse impact as compared to just WHAT?
I'd be better off if people just worshipped me,
if bankers would rival to put gold in my accounts,
if women would fight for my bed, etc.
Your refusing to worship me, etc.,
has an adverse impact on my planned way of life.
Can I have any lawful claim to any of these?
No, because you don't owe me any of these in any way.
Such plans that I may have of your worshipping me are ludicrous,
and deserve but to be laughed at, and certainly not be backed by law.
Similarly, declaring citizens equals has an adverse impact on nobleme;
freeing slaves has an adverse impact on slave-owners' lives;
abandoning communism has an adverse impact on the nomenklatura, etc.;
but can they complain? No, because nobody owes them wealth and power.
In short, having a claim is not enough. I have no less a claim than you do.
Then question is to determine which claims are just, which are not. I
still await whatever arguments you may have to back claims that IP is good.
Funny how, with different tones, this reply is essentially the same as
my reply to our Ayn Rand following Eric Langjahr: rational discussion
of disagreements consist in exploring premises, until the disagreement
disappears, or the speakers agree to disagree. So if you want to usefully
say more than basically "I disagree", then you should explore the reasons
the basic divergent claims may or may not be justified. [And to Eric,
combining the premises with more and more complex extra premises
doesn't serve to advance the debate, but only to make things more obscure.]
As for me, I'll argue that freedom is justified by moral principles,
which can be translated into the isomorphic utilitarian assertion that
in the long run, and statistically, if you're a honest hardworker,
you will have a better expectation of gain with liberty
than with any privilege you can get today.
[In the long run, I'm dead, will object some protectionist,
so if I can seize any big privilege, I should take it.
Yes you will be dead, I'll reply to him, and society will eventually
stop heeding such ludicrous claims as those that were yours,
and stop supporting any privilege whatsover.]
> We have a lot of tools we need written.
> Who's going to write them and what's going to motivate them?
If YOU need them written, then just YOU pay. If you don't pay,
they don't get written, and your "need" is not fulfilled, end of story.
In the end, you get to choose where to put your marginal dollar
for what (expected) marginal effect. That's how the company where
I work earns its dollars (well, euros): people need solutions
to THEIR problems, we build the marginal increment required to achieve
these solutions, as compared to the pool of published free software.
Either these solutions remain secret (which is not an infringement
to the principle of free software), or they get published as free software.
> Is this black magic? Do I just hope? Does crossing
> my fingers, knocking on wood, or some such thing speed up the process?
No black magic. You pay someone to do the job.
Mind you, free software or not free software, over 90% of all
software written in the world is such specific customer-built solutions,
so abolishing IP privileges won't change much about the stability
of most of the software industry. As for the remaining 10%,
even if it disappears, the end of the wasteful course toward redundant
developments due to IP barriers, and the ensuing sharing of free code
will more than make up for the lost production. And it needn't disappear,
since people can imagine ways to organize differently than with
proprietary licensors (development consortiums, support companies, etc).
> I understand what it means when I say "get out your wallets and start
> paying for what you value". It might have no effect, but hopefully it
> at least improves the chances someone will pay.
> But if you say, "put your wallets away and stop paying content-producers
> for value you receive". What follows after that exactly?
Once again, that's a straw man argument, since I heartfully
agree with the former statement and disagree with the latter.
I find it sad that many old time LISPers (you, RPG, other ones)
seem to not even understand what free software is about. Mind how
I'm not speaking about adhering to its theses, but about understanding them.
You can't claim disagreement with a thesis when you misunderstand it.
You can claim you oppose it nonetheless; you can claim it is utter nonsense
and that there is nothing to understand. But you can't claim disagreement.
Free software is about freedom to use, copy, modify or redistribute.
Just because your gardner is free to tend your garden without permission
from the garden architect doesn't mean he'll do it for you free of charge.
If you don't pay him, he'll tend *his* garden, and if you like it the way
he does it, you'll have to copy his ways yourself (if you manage).
Likewise, just because any engineer can enhance FooLISP doesn't mean,
any particular engineer will fix *your* bugs free of charge.
If you need those bugs fixed tomorrow, or even better, yesterday
(through proactive code review bug-removal as in OpenBSD),
as opposed to "whenever someone else gets tired of the bug and fixes it",
then put your money where your ass needs protection, and pay someone.
If you want a warranty that someone will be here to fix bugs, pay him.
Fixing bugs tomorrow is the classical technical support service;
fixing bugs yesterday is achieved by funding an appropriate consortium.
Debugging the empty program (or whatever's publicly available)
into a software solution to your problem is
the classical software development business.
I could have invented it, except that it's in the GNU manifesto as regards
software, in nineteenth century papers as regards IP in general, and in
a wealth of free trade literature as regards other kinds of protectionism.
Again, it's not quite about your agreeing to it, but about honestly
acknowledging what your contradictors are or aren't arguing.
To people who ask what is the business model of Free Software,
I reply: "what is the business model of Free Trade?".
PS: some private email I got indicates that it hasn't been made clear enough
how each speaker's opinions and arguments only commit said speaker, and not
any other people who share any particular such opinion or argument. Different
free software siders do it for different reasons, with different arguments,
and similarly for opponents. I do not claim to have made a review of such
arguments in my messages. On the contrary, I only tried to explain
a particular point of view from someone who, unlike most free software siders,
has the positive opinion that intellectual property in general is wrong.
Again, for literature against IP (including reviews of arguments on both
sides, the most comprehensive one being in french), see
http://fare.tunes.org/articles/patents.html#pointers
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
As long as software is not free, we'll have hardware compatibility,
hence bad, expensive, hardware that has decades-long obsolete design
Francois-Rene Rideau <·······@SPAM.tunes.org> wrote in message news:<··············@Samaris.augustin.thierry>...
> I find it sad that many old time LISPers (you, RPG, other ones)
> seem to not even understand what free software is about. Mind how
> I'm not speaking about adhering to its theses, but about understanding them.
> You can't claim disagreement with a thesis when you misunderstand it.
It isn't clear that they do misunderstand it. Based on what they've
written, they seem to understand it quite well, and have come to a
different conclusion.
I realize that it's easier to argue against certain positions, but that
ease doesn't obligate other people to take those positions or justify
implying that they have. Heck - it's poor form to act as if they've
taken those positions.
Examples: No one has argued that someone else has an obligation to produce
software above a minimum price. No one has argued that people who choose
to produce free software are wrong to do so. No one has argued that taking
a risk entitles one to reward. (Note that rewarding someone for taking
a risk and being right/lucky/whatever is a very different thing.)
I'm particularly taken with the notion that it's wrong to use the copyright
mechanism to extract monetary payment but the GPL's use of that same mechanism
to extract "work" payment is acceptable. (The GPL doesn't "free" software;
it extracts significant consideration in return for permission to create
a derived work. Note that other copyrightable work paid for by US taxpayers
tends to be public domain, which is "free", as in "you can do whatever
you want with it.")
I'm amused by the thought that someone actually thinks "programmers shouldn't
be allowed to be anything other than wage slaves" is a good thing.
BTW - Was mentioning Cortez an attempt to side-step Godwin's law?
-andy
Francois-Rene Rideau <···········@tunes.org> wrote in message news:<··············@Kadath.augustin.thierry>...
> The goal of a rational debate is not to convince the other party,
> but to construct a solid net of argumentative theory and counter-theory.
There are lots of goals of rational debate.
> I agree that some people might be less well because they or other people
> write free software than they would if they could benefit from an undisputed
> copyright monopoly. That doesn't entitle them to claim any damage.
Strawman. Kent merely pointed out that the damage occured, and he's
right. You may feel that it's okay to break those eggs, but that
doesn't mean that they aren't broken.
> And that doesn't mean that they can't live a decent worker's life
> with or despite free software.
How depressing. Do you really want to argue that the "freedom"
of "free software" is the life of a wage slave?
If so, free software also pretty much guarantees that rational
geniuses will tend do something else if they have a choice.
There's nothing wrong with living to code, but that's the only way to
make a living at coding, you're not going to get code from many of the
interesting people. Why? Because they have other things to do, and
living to code, spending all of their time coding because that's the
only way to make a living by writing code, will keep them from doing
those other things.
-andy
Kent M Pitman <······@world.std.com> writes:
> Open/free software might have their good effects, but also their ill
> effects. Certainly the same is true of Copyright. On balance, I
> prefer Copyright, and not by some small margin. I don't disagree it
> has sometimes ill effects, and I even take notes to myself about those
> ill effects.
I don't think Free software (GNU sense) and Copyright are opposites
and that one must choose between them. In fact, Free software relies
critically on a strong copyright legislation to enforce the GPL. Some
interpretations of the legal concept of "derived work" would make it
trivial to take parts of a GPL program and embed it in a closed source
program.
--
Lieven Marchand <···@wyrd.be>
Making laws appears to win votes. Enforcing them doesn't.
See Rule One. Roger Burton West in the monastery.
Kent M Pitman <······@world.std.com> writes:
> I'm just tired of being portrayed as the one with the black hat,
> and seeing the other guys portrayed as the white hat people (in which
> group I include those wearing red hats). I'd feel a lot happier if
> we were all seen to be wearing gray hats, and we just left it at that.
The answer to that is not to turn around and call all of those with
white hats the bad guys either, which you have done multiple times in
this thread, insinuating that by releasing or advocating Free Software
they are depriving you of your ability to accumulate wealth. Should
we not do what we think is ethical because it could have detrimental
effects to a small portion of the population which we think is much
much smaller than the portion of the population which would benefit
from our works? I think it's an ethical decision much like you
describe below. I don't hold your opinion that by enabling a small
portion of society to accumulate wealth we bring greater benefit to
the whole of society, so I have to act accordingly when I make ethical
decisions about what to do with my code (and I don't always have the
same answer mind you).
In another post you lament the scapegoating effect, the attempts to
find an easy answer to the failure of some market. What
differentiates your accusations about Free Software destroying your
small infrastructure tool market (you talked specifically about a tool
that took three weeks to make) from the scapegoating you lament?
Could it be that the lisp community itself was too small, that it does
not have the funds to support a tool vendor, particularly
infrastructure like tools like XML parsers, DB interfaces and web
server interfaces? Could it be that hand-rolling a one-off solution
was in some cases more cost-effective, further reducing that market?
Could it be that supporting a tool on multiple vendors is not
cost-effective fora single entity, particular when it hits
non-standardized language features? Perhaps it was that a tool vendor
with such a small set of product is not one that many people are
willing to bet their product on (some people say this about companies
as big as Apple)? Or is it simply that anything that takes only three
weeks of work is much too much at risk of having that somewhat trivial
effort replicated in the near future, perhaps using weekends instead
of weekdays to do the work?
To me, it's obvious that at this time the lisp community can use both
types of software, and that it's toolkit does indeed consists of both
types of software.
> We're all in search of an ethical way to behave, I think. But
> ethics is a funny thing. I don't think you ponder an ethical
> question for a long time with the goal of arriving at an ethical
> answer. Rather, I think you remain ethical so long as you endure
> the burden of knowing there is no one general answer, and as long as
> you continue to worry about whether your interim actions are "the
> best you can do". At the point where you've dismissed the problem
> as either solved or unsolvable, you have ceased to be ethical
> because the very essence of ethical behavior is caring, and the very
> essence of thinking you've reached an answer is the cessation of
> caring.
Agreed.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
Craig Brozefsky <·····@red-bean.com> writes:
> Kent M Pitman <······@world.std.com> writes:
>
> > I'm just tired of being portrayed as the one with the black hat,
> > and seeing the other guys portrayed as the white hat people (in which
> > group I include those wearing red hats). I'd feel a lot happier if
> > we were all seen to be wearing gray hats, and we just left it at that.
>
> The answer to that is not to turn around and call all of those with
> white hats the bad guys either, which you have done multiple times in
> this thread, insinuating that by releasing or advocating Free Software
> they are depriving you of your ability to accumulate wealth.
It's simply a matter of balance. Right now I hear a lot more suggestion
that free software is without problems, and I feel that balance is correctly
established by a strong position in the other direction. If/when the pendulum
swings the other way, I will consider a change to my rhetoric.
> Should we not do what we think is ethical because it could have
> detrimental effects to a small portion of the population which we
> think is much much smaller than the portion of the population which
> would benefit from our works?
That's up to your conscience. My concern is to make it clear what the
cost is. I think the cost is grossly underestimated.
> I think it's an ethical decision much
> like you describe below. I don't hold your opinion that by enabling
> a small portion of society to accumulate wealth we bring greater
> benefit to the whole of society, so I have to act accordingly when I
> make ethical decisions about what to do with my code (and I don't
> always have the same answer mind you).
You're welcome to take this position, but I have watched this position
neutrally my whole career. I am coming to the opinion that the effect
is far less benign than what you believe or suggest, and so I feel it
important to speak.
> In another post you lament the scapegoating effect, the attempts to
> find an easy answer to the failure of some market.
The difference is that I am not trying to scapegoat anything. I have
no one to whom I must defend my present situation. Rather, I am
trying to understand a set of market dynamics that seem not to behave
as I wish they would by analyzing the forces at work. I'm happy to
let that lead wherever it might lead--I don't have a vested interest
in it leading to free software. I myself have contributed as much
"free" stuff as many others. I contribute tons of free advice here,
and I may yet reach the opinion that such is isomorphic to what I'm
railing against now--that perhaps it undercuts the ability of some to
write books. Perhaps it undercuts my own ability. I don't know. But
my point is to answer the substance of your question on what is the
difference between my remarks and scapegoating, and the answer is
motive. I have no special motive to suspect free software. I use
free software. I have written free software. However, in analyzing
the market dynamics, I cannot see the difference between its effect
and the already-disallowed effect of commercial "dumping".
> What differentiates your accusations about Free Software destroying
> your small infrastructure tool market (you talked specifically about
> a tool that took three weeks to make) from the scapegoating you
> lament?
I haven't actually tried to sell the tool in question. I don't know
that it won't sell. However, I know that there are not many sold, and
I find that alarming. Moreover, I have been at companies where I've
had to tell vendors their prices were too high because variuos tools
are offered free. And I have to wonder whether that's fair. I'm
telling people who are just trying to earn an honest buck that the
free contributions of others has injured there ability to demand pay
for work done. Sorry, but I think that's bad. That's not me
scapegoating. I've been as much a contributor to the process as a
victim of it. I am trying to do an honest analysis of the process,
not point fingers at the players.
> Could it be that the lisp community itself was too small,
> that it does not have the funds to support a tool vendor,
> particularly infrastructure like tools like XML parsers, DB
> interfaces and web server interfaces?
It could be indeed. That itself would be a problem.
I actually think, as I've mentioned before, that the world would be
well-served if there were a maximum corporate size. In case it's not
obvious, a consequence of this theory is that if everyone on earth wanted
a copy of microsoft word, they would not be able to buy it. At least
not without radically restructuring flows of profit-taking. Because
if everyone bought a copy, the company would get so much money it would
probably have to splinter. The hope would be to have more people offering
more kinds of things, rather than one single "commodity", which I think
is unhealthy to the market.
People talk about cloning as fragile in biology because it's not virus
resistant, for example. I have to wonder if the same isn't true or shouldn't
be true about software. Sure, it's efficient to have everyone have the
same http server, but it also means the entire world is always open to
identical security holes, etc.
I have to believe at least some what makes Lisp interesting design-wise is
that it encourages non-commodity thinking. (I like the way it's done, but
I don't have a lock on wisdom. It might be as bad for Lisp to corner the
market as for C or Java.)
> Could it be that hand-rolling
> a one-off solution was in some cases more cost-effective, further
> reducing that market? Could it be that supporting a tool on
> multiple vendors is not cost-effective fora single entity,
> particular when it hits non-standardized language features? Perhaps
> it was that a tool vendor with such a small set of product is not
> one that many people are willing to bet their product on (some
> people say this about companies as big as Apple)? Or is it simply
> that anything that takes only three weeks of work is much too much
> at risk of having that somewhat trivial effort replicated in the
> near future, perhaps using weekends instead of weekdays to do the
> work?
Possible, but unfortunate. It suggests I can't purchase mix-and-match
components. I think if the clothing industry can make it efficient to
make a whole product line out of hair twist-ties, we should be able to
do the same with software.
> To me, it's obvious that at this time the lisp community can use both
> types of software, and that it's toolkit does indeed consists of both
> types of software.
I've heard no argument in any of this discussion to convince me that there
is a valid business model at which free software is at the core. All I've
heard are rationalizations to how I can get free software produced and still
have money coming in, as long as what I'm selling is something else. I
happen not to want to do the something else, so that leaves me with nothing.
To me, it's the same as if you were selling one of these late night
infomercials where I make money from selling swamp water, so having a lot
of swamp water to sell, I buy your book and it says "just make sure to
include one Porsche automobile with every gallon of swamp water, and people
will pay you tons of money". Well, of course they will. If I wanted to
sell support, I could do that without developing software! I don't need
to do both. But I want to develop free software. And I want money.
That is my problem. You are creating an industry in which people expect
software for free, or at least some amount of the time, and that devalues
the cost of what I do. It's pretty easy to see that this drives down the
price. I don't see any factual arguments on your side. All I see are
arguments of why I could or should live with it. Ultimately, I have to
live with it unless I lobby to make it illegal, the same as "dumping",
which believe me, I haven't ruled out. But, regardless, I don't have to
like it.
At minimum, I think every person who contributes free software should
be ethically required to write a letter to whoever pays their
mortgage, tuition, or food and explain that they are giving away stuff
that might instead be possible to sell. In the case that a person is
supporting themselves, this might be a no-op. But a lot of the people
contributing this are doing it on someone else's nickel, and I don't
see it as much different than having your mom give you thouands for
"school expenses" only to find you gave away hundreds or thousands of
it to charities, since that's possibly what you might have made. To
some extent, I doubt that all employers allowing their coding efforts
to be released as open source are being offered a proper disclosure
either. Sometimes they may really be. Sometimes I bet they don't
understand the issue well enough to have an informed judgment.
Certainly I see nothing in the literature of free software suggesting
any kind of disclosure of this kind, which makes me leary of anyone
making a claim that the situation is being adequately dealt with.
I think people are just asserting without discussion that it's the
obviously right thing, or that it requires no discussion. I simply
don't agree.
On Wed, 30 May 2001, Kent M Pitman wrote:
> Craig Brozefsky <·····@red-bean.com> writes:
> > What differentiates your accusations about Free Software destroying
> > your small infrastructure tool market (you talked specifically about
> > a tool that took three weeks to make) from the scapegoating you
> > lament?
>
> I haven't actually tried to sell the tool in question. I don't know
> that it won't sell. However, I know that there are not many sold, and
> I find that alarming. Moreover, I have been at companies where I've
> had to tell vendors their prices were too high because variuos tools
> are offered free. And I have to wonder whether that's fair. I'm
> telling people who are just trying to earn an honest buck that the
> free contributions of others has injured there ability to demand pay
> for work done. Sorry, but I think that's bad. That's not me
What's bad about it? Your company is not a charity. It's hardly an
honest buck they're trying to make if the same thing can be had for free.
> > To me, it's obvious that at this time the lisp community can use both
> > types of software, and that it's toolkit does indeed consists of both
> > types of software.
>
> I've heard no argument in any of this discussion to convince me that there
> is a valid business model at which free software is at the core. All I've
Note that "Lisp community" does not equal "valid business model" in its
entirety.
Some people seem to want to use Lisp to solve their problems. They're not
in the business of selling software; it's not their core competency and
would be a distraction. Why shouldn't they share their clever tools and
get other willing hands to help maintain it?
> At minimum, I think every person who contributes free software should
> be ethically required to write a letter to whoever pays their
> mortgage, tuition, or food and explain that they are giving away stuff
> that might instead be possible to sell. In the case that a person is
What a total load of crap.
I recall that at one time you were learning Portuguese. Did you let
Symbolics/Harlequin in on that, that you were wasting valuable time
chasing after the Portuguese Lisp market instead of working away on
additional products that might sell?
I find it really hard to believe that someone with an obvious intellectual
gift and a sensitivity demonstrated in his writing has to denominate the
worth of everything in dollars.
Tim
Kent M Pitman <······@world.std.com> writes:
< BTW, I've cut out the parts that I agree with, or where our
differences are very minor. I thought I should make that customary
exercise explicit in this case. >
> > I think it's an ethical decision much like you describe below. I
> > don't hold your opinion that by enabling a small portion of
> > society to accumulate wealth we bring greater benefit to the whole
> > of society, so I have to act accordingly when I make ethical
> > decisions about what to do with my code (and I don't always have
> > the same answer mind you).
>
> You're welcome to take this position, but I have watched this position
> neutrally my whole career. I am coming to the opinion that the effect
> is far less benign than what you believe or suggest, and so I feel it
> important to speak.
There are two values being balanced here, the detriment to producers
and the benefit to consumers. In our situation there is also the
issue of technological advancement, so we have to balance those two
values over time, great consumer benefit now can have disasterous
consumer effects later, and vice versa.
It does neither of us any good to try and counter-balance one
another's rhetoric, since we'll never be able to see the terms of our
analysis in a way that would allow us to assess the effects of Free
Software on our community.
> motive. I have no special motive to suspect free software. I use
> free software. I have written free software. However, in analyzing
> the market dynamics, I cannot see the difference between its effect
> and the already-disallowed effect of commercial "dumping".
We have to look at what dumping is first. Most anti-dumping law has
to do with foreign trade. Selling in one market at a price
substantially lower than another. Since Free Software is the same
price anywhere, theoretically free, it's difficult to conceive of it
as dumping in that sense.
However, if we look at predatory dumping, restricted to a single
market, we have a useful target. Predatory dumping requires that the
the dumper have sufficient resources to survive underselling longer
than their competitors, and that there is a significant entry barrier
so that after competition is driven out and the dumper is attempting
to collect monopoly prices it is too costly for others to enter the
market and provide competition anew.
First, Free software authors usually do not have more money than their
competitors. In fact, in most cases their competitors are
exponentially larger in terms of cash available. Second, with Free
Software there is not a significant entry barrier once competition has
been eliminated. The previous competitors software still exists and
can be picked up again, and also anyone can fork the Free Software.
Lastly, I have yet to see a Free Software project try and re-introduce
monopoly pricing after dominating it's market (tho I have seen
database providers do this, CDDB in particular, but they were not
really Free Software to start since their license was not protective
of contributors rights). There seems to be no relation between Free
Software and predatory dumping.
There are other types of dumping that might be more akin to what Free
Software is. Persistent dumping is when the producer consistently
sells lower in one market than another. An example often given is the
U.S. Film industry which can recoup all it's costs in the U.S. and can
then sell abroad for very little. [1] Persistent dumping is legal
domestically, and most economists would argue that it is beneficial
since consumers benefit more than producers are harmed. Free Software
could be seen in a similiar light. The costs of production are
recouped or found in one market, support, consulting, custom software
development, or hobbyist time, and then it can be dumped into another
market, the commodity software tool market in this case.
Drastic price reductions can have disasterous effects on producers in
a market, but that does not mean they are, or should be illegal. As
the article I referenced points out, most economists would say that
consumers benefit more from price reductions than producers are
harmed. I'm not sure if I would use this as an argument in support of
Free Software having a greater benefit for a larger number of people
than the negative effect it has on producers, but it does point in
that direction. The main reason I wouldn't is that I don't think
their analysis of commodities in the abstract accounts for the costs
technological advancement of software thru research.
> I haven't actually tried to sell the tool in question. I don't know
> that it won't sell. However, I know that there are not many sold,
> and I find that alarming. Moreover, I have been at companies where
> I've had to tell vendors their prices were too high because variuos
> tools are offered free. And I have to wonder whether that's fair.
> I'm telling people who are just trying to earn an honest buck that
> the free contributions of others has injured there ability to demand
> pay for work done.
What about all those benefits of proprietary software that those
companies have over the Free offerings you have been telling us about?
The better support, the greater control over the evolution of the
tool? I mean if you really want to do this right then you should
suggest they file for patents in their arena, or work to differentiate
their product sufficiently, just as they would from any competitor who
entered the market. If their tool is not sufficiently differentiated,
or their business not run well enough, then I see no reason they
should be protected as a business.
You have yet to make a convincing argument as to why the failure of a
vendor in this general situation is overall detrimental to society.
It's quite obvious to both of us that it's detrimental to the company
and it's employees, and perhaps even it's current customers. But how
does that measure against the availability of Free Software that
performs the same job? No licensing fee needs be paid by users, users
can see the source, modify it if they like, or someone else can start
a business based on that software (if it was BSD licensed). And what
solutions to the problems we face from losing that vendor are
precluded by the existence of Free Software? Not the need for
commercial support, not the need for continued development, not the
need for distribution. All of those can exist with Free Software,
even if presently there are few companies in those markets.
The place I think is worth focusing on is the continued advancement of
the problem domain said vendor was addressing. It's hard for Free
Software projects and vendors to generate the revenue needed for
sustained research and development. What a coincidence that this is
the issue copyright and patents were designed to address in the first
place. Then again, it's no coincidence that this is the issue that
most political-economic analysis of intellectual property law focuses
on. This is also what leads me to believe that without significant
changes in the way our society funds advanced research and
development, Free Software will mostly be used in established fields,
infrastructure, and commodity markets, while proprietary licenses and
patents will be used extensively in new fields, at the edge of our
knowledge.
Within our community there are things like continued compiler
development whose costs may be best met thru the sale of proprietary
licenses. Of course Free implementations that recieve their funding
from elsehwere are quite possible. There are also things, like the
XML tools, web servers, and socket toolkits, which require little
research and are not great revenue generators for various reasons, and
therefor are more likely to have Free solutions than proprietary ones.
This doesn't preclude you or anyone from selling proprietary versions
of them, but your doing so doesn't warrant any special price
protection.
> People talk about cloning as fragile in biology because it's not virus
> resistant, for example. I have to wonder if the same isn't true or shouldn't
> be true about software. Sure, it's efficient to have everyone have the
> same http server, but it also means the entire world is always open to
> identical security holes, etc.
I don't see anything in Free Software that requires a single solution
in any market. OS's are a good example, we have Linux and the various
FreeBSDs, we even have multiple Linux distributions. Almost every
market segment with Free Software offerings has more than one FS
offering, ranging from sound drivers, word processors, http servers,
and mail transport agents, to lisp implementations.
> > Could it be that hand-rolling a one-off solution was in some cases
> > more cost-effective, further reducing that market? Could it be
> > that supporting a tool on multiple vendors is not cost-effective
> > for a single entity, particular when it hits non-standardized
> > language features? Perhaps it was that a tool vendor with such a
> > small set of product is not one that many people are willing to
> > bet their product on (some people say this about companies as big
> > as Apple)? Or is it simply that anything that takes only three
> > weeks of work is much too much at risk of having that somewhat
> > trivial effort replicated in the near future, perhaps using
> > weekends instead of weekdays to do the work?
>
> Possible, but unfortunate. It suggests I can't purchase mix-and-match
> components. I think if the clothing industry can make it efficient to
> make a whole product line out of hair twist-ties, we should be able to
> do the same with software.
I think wether you can mix and match components is largely unconnected
to wether code is Free Software or not, and has more to do with the
establishment of standard interfaces. Those interfaces can have free
and/or non-free implementations. In my experience, I've found that
putting togther solutions from various components is easy with Free
components, since I have source code and can solve the ever-present
interface issues as they arise. Proprietary software with source
available and easily patchable would have the same benefit here
perhaps.
> I've heard no argument in any of this discussion to convince me that
> there is a valid business model at which free software is at the
> core. All I've heard are rationalizations to how I can get free
> software produced and still have money coming in, as long as what
> I'm selling is something else.
I suggested you check out the FSB list, which has a continuous
discussion of this very topic on it and people who are more
knowledgable about running businesses based on Free Software than I.
I can also present some examples of companies that to varying degree
have the development of Free Software as their core business, Cygnus,
RH, Mandrake, Alladin, Sleepycat, Sendmail, Progeny, Collab.net. I
think it's fair to say that the support/devel mixes of these companies
is as varied as those companies that write proprietary software. I
think it's meaningless to discount any company which sells support, or
distribution, or consulting based around Free Software, particularly
when they are the primary drivers of Free Software projects, just
because they do some of the same things to generate revenue that
proprietary companies do.
Beyond those suggestions there is no way I can provide you with a
risk-free business plan for selling your software without knowing more
specifics about it. Even then, I think it's a severe handicap to
discount those revenue streams that other software companies use such
as support, consulting and distribution. If you want to talk about
specifics which will let you get the code you've written out into the
public in our current market, and at least pay for itself, then you
would have to look at all of those revenue streams, even if you did
use a proprietary license.
> it. Ultimately, I have to live with it unless I lobby to make it
> illegal, the same as "dumping", which believe me, I haven't ruled
> out. But, regardless, I don't have to like it.
Since you seem to have some misconception about how anti-dumping law
works, I would suggest that you not pursue that avenue. You're prolly
more likely to succeed by placing requirements on people distributing
software which make it unrealistic for individuals, or
non-incorporated entities of any kind, without sufficient
capitalization to distribute software. See UCITA for some attempts at
this.
> At minimum, I think every person who contributes free software should
> be ethically required to write a letter to whoever pays their
> mortgage, tuition, or food and explain that they are giving away stuff
> that might instead be possible to sell.
Should my mom write a letter justifying her cooking for me, or
bringing me my dinner because she could have spent that skill and time
making a meal for a customer at a eatery, or waiting on them? Should
my GF write a letter when she sings me to sleep because she could have
recorded it and sold it? Hell, if I lived in a state where
prostitution is legal should me and my GF write a letter when we have
sex to justify why we didn't charge one another, or charge someone
else, or record it and put it on a porno site? Perhaps you should
write a letter justifying why you didn't sell your kidneys, or your
kids if you have them (especially kids as they are obviously a cost
sink since child labor laws were put in place).
Do we really want our activities regulated, or even known for that
matter, by our creditors based upon any existent, or non-existent
possibilities that some exchange value could be eeked out of them?
Me, I don't think so.
> To some extent, I doubt that all employers allowing their coding
> efforts to be released as open source are being offered a proper
> disclosure either. Sometimes they may really be. Sometimes I bet
> they don't understand the issue well enough to have an informed
> judgment. Certainly I see nothing in the literature of free
> software suggesting any kind of disclosure of this kind, which makes
> me leary of anyone making a claim that the situation is being
> adequately dealt with. I think people are just asserting without
> discussion that it's the obviously right thing, or that it requires
> no discussion. I simply don't agree.
Then you haven't checked the FSB list or the books on licensing
issues, or any of the other forums discussing licensing issues in the
context of running a software business. They are readily available.
There are conferences on the topic, there are mailing lists dedicated
to it, and there are journals featuring articles on it regularly.
Free Software can't be held responsible for your, or any business'
lack of research into the issues at hand.
[1] Anti-Dumping Law
http://students.washington.edu/~achamber/subs/dumping.html
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
Kent M Pitman <······@world.std.com> writes:
> Moreover, I have been at companies where I've
> had to tell vendors their prices were too high because variuos tools
> are offered free. And I have to wonder whether that's fair. I'm
> telling people who are just trying to earn an honest buck that the
> free contributions of others has injured there ability to demand pay
> for work done. Sorry, but I think that's bad.
Scene 1:
Cowboy Kent: Sheeit! Mr Land Baron, sir, but I've just gotta tell ya,
ya can't demand payment from the homesteaders for using
your water, anymore.
Land Baron: Why the fuck not?
Cowboy Kent: Well, they've found a spring up in them thar hills. (pointing)
Land Baron: Fuck that! I'll buy a new law stopping them. Where the
fuck is my congressman?
... to be continued.
Regards,
Peter
Once a businessman gets a golden egg, he kills the goose to make
sure he has a monopoly. -Richard Stallman.
Peter Wood <··········@worldonline.dk> writes:
> Kent M Pitman <······@world.std.com> writes:
>
> > Moreover, I have been at companies where I've
> > had to tell vendors their prices were too high because variuos tools
> > are offered free. And I have to wonder whether that's fair. I'm
> > telling people who are just trying to earn an honest buck that the
> > free contributions of others has injured there ability to demand pay
> > for work done. Sorry, but I think that's bad.
>
> Scene 1:
>
> Cowboy Kent: Sheeit! Mr Land Baron, sir, but I've just gotta tell ya,
> ya can't demand payment from the homesteaders for using
> your water, anymore.
>
> Land Baron: Why the fuck not?
>
> Cowboy Kent: Well, they've found a spring up in them thar hills. (pointing)
>
> Land Baron: Fuck that! I'll buy a new law stopping them. Where the
> fuck is my congressman?
>
> ... to be continued.
>
> Regards,
> Peter
>
> Once a businessman gets a golden egg, he kills the goose to make
> sure he has a monopoly. -Richard Stallman.
I have exhausted my quantum for this thread for this week.
Probably best to end on a note of humor.
In my probably-welcomed silence, perhaps someone else will more
usefully take up the challenge of how to move the community ahead.
Just please don't characterize this as if it were MY problem
exclusively or as if *I*'d found some golden egg I wanted to have a
monopoly on. I'm as worried about having people to BUY from as having
people to SELL to. I just want a working system that both incentives
people to be involved in ways they haven't been up until now, and that
feeds them in ways they haven't been getting fed of late. The
proposed "free software" approach seems to me to feed people if taken
exactly as specified, but only because you have to do something else
to make the money, and that does not incentivize at least me. But
maybe I'm unique in that. Maybe there are tons of Lispers lined up
wanting to just code freely because money is flowing to them from
other sources. Maybe it's just me that doesn't get it. Maybe they
will all step forward and say "I volunteer" now that I'm being quiet
enough to listen. I don't know. Anyway, I'll be reading rather than
writing for a while on this thread, and maybe I'll learn something.
Have fun.
On Thu, 31 May 2001 10:34:00 -0700, Eric Langjahr <········@itsdata.com>
wrote:
>Not sure what you are so upset about! I believe in the marketplace.
>Eventually it will all sort itself out.
Arguing about free markets is like arguing about abortion. Such arguments
are a waste of time, because people don't agree on the fundamentals.
Copyrights are enforced by governments. Governments get their power from
their military forces.
Everything will of course sort itself out in the long run, and what will be
will be, but such sorting-out could include a big nuclear war making all
people equal and making copyrights obsolete.
Let us therefore focus on technology and not argue about free markets and
other such imponderables. The best we can hope for is that advancing
technology will lead to more freedom and safety over the long run.
The software market is an immediate practical concern for all of us, but we
should probably focus on the way it is rather than on the way it should be.
We can possibly influence Lisp vendors to change their policies, prices,
etc., but we have no hope of influencing millions of programmers to change
their attitudes about free markets or free software.
·······@wherever.com writes:
> On Thu, 31 May 2001 10:34:00 -0700, Eric Langjahr <········@itsdata.com>
> wrote:
>
> >Not sure what you are so upset about! I believe in the marketplace.
> >Eventually it will all sort itself out.
>
> Arguing about free markets is like arguing about abortion. Such arguments
> are a waste of time, because people don't agree on the fundamentals.
> Copyrights are enforced by governments. Governments get their power from
> their military forces.
>
> Everything will of course sort itself out in the long run, and what will be
> will be, but such sorting-out could include a big nuclear war making all
> people equal and making copyrights obsolete.
It's even more drastic than that. Remember what Keynes said:
in the long run we are all dead
to which we mus add...
Don't Panic!
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
"Marco Antoniotti" <·······@cs.nyu.edu> wrote in message
····················@octagon.mrl.nyu.edu...
>
> ·······@wherever.com writes:
> Remember what Keynes said:
>
> in the long run we are all dead
>
> to which we mus add...
>
> Don't Panic!
As long as your on the subject of being dead :-)...
Kent M Pitman <······@world.std.com> writes:
> in analyzing the market dynamics, I cannot see the difference between its
> effect [free software] and the already-disallowed effect of commercial
> "dumping".
Perhaps you can explain what you think is wrong with dumping even in the
standard case. I think it's a bit subtle why that needs to be regulated.
As I understand it, the problem isn't with the dumping itself. Dumping puts
goods on the market below their manufacturing cost, in order to build market
share. The loss is paid for by profits from other, more secure parts of the
large conglomerate. The hope is to drive single-product or single-industry
competitors out of business.
That part is all fine. The problem comes once the competitors have been
driven out of business. Then the dumping winner suddenly stops dumping,
and raises prices to monopoly rates.
You've made an analogy to free software, but I'm not sure that the same outcome
applies. Free software may drive some competitive commercial software out of
business. But the free software will never become not-free, so the evils of
above-market monopoly prices never appear.
You've also claimed concern that the existance of lots of free software will
kill off commercial software as an economic endeavor. I'm not sure why you
can't see them coexisting. If free software is available, then you're
correct that it's difficult to make money in that niche. But there are always
other niches. Can't you build on the foundation of ever-growing free software,
in order to deliver proprietary value-added software on top? Note that this
isn't the "give away the software, sell the service" model. It's just
peaceful cooperation between the topic areas where the open-source model is
effective, and those where the commercial model is effective.
-- Don
_______________________________________________________________________________
Don Geddis www.goto.com ······@goto.com
VP of Technology, GoTo Shopping Phone 650-403-2220
1820 Gateway Drive, Suite 360, San Mateo, CA 94404 Fax 650-403-2201
New York...when civilization falls apart, remember, we were way ahead of you.
-- David Letterman
In article <···············@world.std.com>,
Kent M Pitman <······@world.std.com> wrote:
>Craig Brozefsky <·····@red-bean.com> writes:
>
>> Kent M Pitman <······@world.std.com> writes:
>>
>It's simply a matter of balance. Right now I hear a lot more suggestion
>that free software is without problems, and I feel that balance is correctly
>established by a strong position in the other direction. If/when the pendulum
>swings the other way, I will consider a change to my rhetoric.
>
Balance is a matter of perception. I surprised my wife recently
by speculating on possible problems evidenced by the open MCL
kernel; she said she hadn't realized I was not a free software
fanatic. I was thinking that every conversation I was in about
the topic seemed to me anti-free software in ways that I disagreed
with. (Nor should being anti-Microsoft be construed as being
anti-commercial software.)
>That's up to your conscience. My concern is to make it clear what the
>cost is. I think the cost is grossly underestimated.
>
I disagree.
It's a market thing. I have great faith in the free market within
certain bounds. There are two possibilities in any given market
niche. If commercial software writers can produce software enough
better than free alternatives to justify its purchase, then
they can thrive and all is good. If they can't, then they fold
and all is good.
One case I know about is Macintosh C and C++ development systems.
For quite some time now, Apple has provided its system free of
charge, so you can download and go. This hasn't stopped Metrowerks
from dominating the market, because their product is sufficiently
better to dominate in the professional market and much of the
hobby market. If I were using it, at roughly my current salary, it'd
have to make me about 1% more productive than the free alternative
to be worth it.
>the market dynamics, I cannot see the difference between its effect
>and the already-disallowed effect of commercial "dumping".
>
As I understand it, dumping is illegal because it has a tendency to
destroy competition, allowing the dumper to charge more in the
future. With free or open-source software, this isn't possible,
since the free stuff is still available and can still be developed
on.
>I find that alarming. Moreover, I have been at companies where I've
>had to tell vendors their prices were too high because variuos tools
>are offered free.
More precisely, the prices were too high for the benefit. That's a
market thing again. You can use tool X for free, and wonder about
tool Y. If tool X didn't exist, you'd just have to bite the bullet
and buy tool Y, because the functionality is worth more than $Y
to you. Since tool X exists, the Y vendor has to make their tool
worth the money. Assuming a free market, you win either way. If
the added functionality isn't worth $Y, you get something that'll
work without spending any money. If it is worth $Y, you buy it
and benefit.
And I have to wonder whether that's fair. I'm
>telling people who are just trying to earn an honest buck that the
>free contributions of others has injured there ability to demand pay
>for work done.
An honest buck is one earned in a transaction in which both sides
profit. If you can't make a software product that's sufficiently
better than what's free, then selling your software isn't a way
to make an honest buck. It may make it more difficult to make
money in the job of your choosing, but that's not unique to you.
Plenty of people would make other job choices if earning money
was not a consideration.
This also completely ignores the effects on the consumers, who
have a choice. They don't have to accept any shoddy commercial
product and pay whatever the vendor charges. They can use the
free product, which may or may not be shabby.
Sorry, but I think that's bad. That's not me
>scapegoating. I've been as much a contributor to the process as a
>victim of it. I am trying to do an honest analysis of the process,
>not point fingers at the players.
>
Me too. There's several things to consider.
One is that software is reproducible for almost free, and is
valuable. This means that the wealth of the world increases
with each copy of some software somebody wants. Hence, free
software, by allowing such copying, makes the world wealthier
than commercial software of similar value. This is the basic
fact that makes me keep thinking that there has to be something
better than the current commercial model. In no other field
is it possible to make something so intrinsically valuable with
such a trivial copying cost, and there really should be some
way to take advantage of that as a society.
Another is that people need to be able to produce software
full-time without feeling deprived. I do it by working for
a company that sells the stuff, myself. I do a limited
amount of free stuff, but the fact that it's outside of
my job and competes with family and house time means I do
only a limited amount. This means that, to take advantage
of the effects in the preceding paragraph, programmers have
got to be paid for free software, and that's the problem.
There are business models in which free software makes
sense, but only peripherally. A business needing a tool might
write one and GPL it. This can contribute value, but makes
no money.
The nice thing about the commercial model is that it works
with well-understood methods of being paid for work done.
The nice thing about the free model is mostly potential.
>People talk about cloning as fragile in biology because it's not virus
>resistant, for example. I have to wonder if the same isn't true or shouldn't
>be true about software. Sure, it's efficient to have everyone have the
>same http server, but it also means the entire world is always open to
>identical security holes, etc.
>
Agreed. On the other hand, there's an advantage in standardizing on
base software.
>I've heard no argument in any of this discussion to convince me that there
>is a valid business model at which free software is at the core.
Nor have I. I'd love to see one.
>to do both. But I want to develop free software. And I want money.
>That is my problem. You are creating an industry in which people expect
>software for free, or at least some amount of the time, and that devalues
>the cost of what I do. It's pretty easy to see that this drives down the
>price.
Exactly the same thing is true of competition: it drives down
the price. As a believer in market forces, I don't really see
a difference between free or commercial competition. Both force
you to do better work in order to be paid.
To digress slightly, my father-in-law worked for a company that made
its money selling commodity items. It's a rough business, and it's
purely commerical. Nobody gives away trainloads of free corn
syrup, and it's still not possible to get paid what you want for
what you want to do.
I don't see any factual arguments on your side. All I see are
>arguments of why I could or should live with it. Ultimately, I have to
>live with it unless I lobby to make it illegal, the same as "dumping",
>which believe me, I haven't ruled out. But, regardless, I don't have to
>like it.
>
Um, dumping is a commercial maneuver. Free software isn't. I don't
see how to make free software illegal without distorting the market
to extremes and throwing away the ability of the market to set prices.
This means that software pricing would be almost completely political,
based on whoever lobbied the most wanted to set the prices at.
Suppose you wanted to address the desktop operating system market.
What would be the minimum cost for an OS, and why would you pick
that price? Should it be set at the price Microsoft wants to charge,
so as not to inconvenience Microsoft?
>At minimum, I think every person who contributes free software should
>be ethically required to write a letter to whoever pays their
>mortgage, tuition, or food and explain that they are giving away stuff
>that might instead be possible to sell.
Only if they use their employer's resources. My employer has no
right to anything I make on my own time with my own stuff. Are
you proposing a world in which an employer has a 100% claim on
employee time and resources?
>supporting themselves, this might be a no-op. But a lot of the people
>contributing this are doing it on someone else's nickel, and I don't
>see it as much different than having your mom give you thouands for
>"school expenses" only to find you gave away hundreds or thousands of
>it to charities,
If people are working on free software on their employer's time,
with their employer's stuff, without permission, that is wrong.
since that's possibly what you might have made. To
>some extent, I doubt that all employers allowing their coding efforts
>to be released as open source are being offered a proper disclosure
>either. Sometimes they may really be. Sometimes I bet they don't
>understand the issue well enough to have an informed judgment.
Um, from my experience in the business world, no uninformed employer
would dream about letting anything their employees do, that he or
she has a claim on, being released for free. I rather assume that
any employee doing free software on the job is doing so without
their employer's consent (shame!) or doing it with their employer's
informed consent.
If you have worked for a company that would allow its stuff to
be given away for free without darn good reason, do tell.
>Certainly I see nothing in the literature of free software suggesting
>any kind of disclosure of this kind, which makes me leary of anyone
>making a claim that the situation is being adequately dealt with.
I believe the Gnu people require legal documents stating that the
donor has the copyright and is legally able to give away the software,
and that covers that situation. I think that in general, it's
assumed as a matter of honesty.
>I think people are just asserting without discussion that it's the
>obviously right thing, or that it requires no discussion.
That free software is the obviously right thing? In some senses, it
is. In others, it isn't.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
On Thu, 31 May 2001 21:01:30 GMT, ········@visi.com (David Thornley) wrote:
>Agreed. On the other hand, there's an advantage in standardizing on
>base software.
That sounds obvious. But how big is that advantage really? The kind of
future I want to see, and which I try to work towards, has most of the
software being written by software. In that future it seems much more
practical to have everyone request their own version of their base software
than to all try to agree on what base software should be standard.
>>: Peter Wood
>: Kent M Pitman
First, let me say that I share the opinions expressed by Peter Wood,
though I admit the way they are stated will not win agreement lest
you see the big picture behind them.
> It is always your option not to license the software. The beauty of the
> legal system of contracts is that their conditions, whatever they are,
> are just about always allowed because they are voluntarily entered.
> It is your option not to enter the contract.
That's the basic fallacy of intellectual property.
Intellectual property is a government-granted monopoly
that binds even non-contracting parties.
If Intellectual Property was born in licensing agreements in a free market,
then yes, you'd be right. But it isn't. It's a government-granted monopoly.
So the "voluntary contracts" are biased toward the licensor.
Cf. the wealth of documents at
http://fare.tunes.org/articles.html#pointers
> The creative
> process is fundamentally different than sweeping floors. It costs money
> to create things and the consumers of the creative ware must invest in
> that or expect that the content creator will have inadequate funds.
This is no different from any activity that requires investment.
Mind you, sweeping floors efficiently, 24/4/7, no matter weather,
strikes, or government policies, does require quite some bit of
capitals, proficiency and dedication.
No need for special privileges here.
> You will not invent a system that does not unfairly reward someone
> except a system that rewards no one.
There is a system in which people are responsible for there deeds,
for how much they pay and how much they get paid, so that when they fail
to be fair, it serves them as a lesson to be fairer next time.
That's called Liberty. Let's abolish government privileges.
http://www.econlib.org/library/Bastiat/basHar10.html
> (And, incidentally, I think it is false on its face. If Einstein or Salk
> or Mozart did all their work in 3 weeks or 3 years, they are absolutely
> entitled to a lifetime of pay because of the benefit others received.
No. They are entitled to convince other people to voluntarily pay them
for life. Other people are entitled to do so if so they please.
No one, government or not, is entitled to decide who shall be privileged
in getting, by the threat of public force, more of a share than one
gets from the free will of others.
Who is to decide who is the Mozart and who is the Salieri?
A government commission? Anyone can proclaim himself Mozart?
Or maybe make an automatic legal device (such as intellectual property)
for the most wicked trickster to abuse? And finally one that is rigged
toward the corporate interests of political campaign subsiders?
No thanks, ma'am. If you want to give extra credit to some category
of people, do it with your own money, and more power to them!
Don't you pay judges and lawyers and policemen with MY money to
protect YOUR corporate interests.
> The same might be said for war veterans, lamentable as war is.
I don't owe war veterans nothing. I didn't draft them into war.
The crooked politicians who sent them to butchery do. Let THEM pay.
> There's an old joke about a contractor showing up
> at a site and surveying a situation and then pounding in one nail and
> submitting a huge bill. Upon being questioned about the bill, the
> person says "$1 for hitting the nail. The rest is for knowing where
> to put the nail." If you can't acknowledge the reasonableness of this
> simple concept, that's where you should focus in understanding
> people's ire about some of your remarks, ire that I think is quite justified.
Of course we acknowledge the concept that thinking yields value.
What we don't acknowledge is the concept that
the bill price could be decided unilaterally and/or afterwards.
If the price is decided in advance, possibly after the contractor has a look,
then it's ok. Otherwise, the case is open to litigation.
And more to the point, what we fully reject is that the contractor could have
the monopoly on hitting nails, just because he originally built the house.
Especially so when the place where to hit the nail in the second house
is the same, and anyone could have done the $1 hitting part without
needing the $99,999 expertise before.
> I think what the world doesn't owe anyone is free software.
If you don't want to publish it, keep it secret.
But if you publish it, I deny you the right to sue people
who use, copy, modify, and redistribute it.
Sue your licensees if you dare; but don't you sue third parties.
Free software is about recognizing these rights to other people.
GNU GPL is trying to enforcing them for software that you participate into.
>> The bottom line is "if you can't make money doing what *you*
>> want, then get off your arse and do something else." Most people
>> learn that lesson as adolescents.
I heartfully agree with this statement.
> This same remark could be applied to your earlier remark about how hard it
> is for a small business to survive. Others run their small businesses just
> fine on the existing set of rules.
We don't complain about the hardness and ask an entitlement to a result.
We complain about the intrinsic injustice of IP, and demand its abolition,
either by the withdrawal of iniquitous laws, or the voluntary cancellation
of their effect through the voluntary agreements that are free software
licenses.
> I think [free software] will just financially disempower programmers,
> leaving all the money in the hands of businessfolk.
In my article against patents (URL above), I argue (in the case of
a different form of IP, but with equally applicable arguments) that
this is not the case, and that protection is what empowers businessfolk
(and the worst among them: monopolyfolk and lawfolk).
> Then I think the
> businessfolk will take computer science in a different direction than
> tech people would have because the tech people can't afford to chart their
> own course.
I dispute the right of tech people to chart their own course.
"There could hardly be a more unbearable -- and more irrational --
world than one in which the most eminent specialists in each field
were allowed to proceed unchecked with the realization of their ideals"
-- Friedrich A. Hayek
If anyone has the right to chart any course, it is those who pay.
In the end, that means: the consumer.
Privileges empower the producer. Not the consumer, not the tech,
not even the entrepreneur. Just the monopoly holder and dealer.
I'm convinced that proprietary software is a big limitation to
an overwhelming success of LISP on the marketplace.
http://fare.tunes.org/articles/ll99/index.en.html
PS: Peter and I do not in any way pretend to represent everyone in
the Free Software movement and expose a universal rationale behind it.
He seems to be and I am libertarian. There sure are left-leaning backers
of Free Software, too, starting with RMS.
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
Because people confuse information and information-related services
(which include searching, creating, processing, transforming, selecting,
teaching, making available, guaranteeing, supporting, etc), they are afraid
that Free (libre) Information mean free (gratis) information-related services,
which would indeed kill the industry of said services. On the contrary,
Free Information would create a Free Market in these services, instead of
current monopolies, which means they will be available at a fair price,
so the result would be a flourishment of that industry!
On 22 May 2001 18:56:05 +0200, Francois-Rene Rideau
<·······@SPAM.tunes.org> wrote:
>Sue your licensees if you dare; but don't you sue third parties.
Would you suggest, that if I were to leave my car in your garage, and
it was stolen, that I would have NO right to pursue the third party
who stole it? Only YOU for letting it be stolen?
If I leave numerous valuable items on the street, in an unprotected
state, your taking them is still THEFT. You knew they were not yours
to take or use. The fact that I didn't have them under lock and key
doesn't change the nature of your act.
If you steal from me, I am quite prepared in the absence of other
recourse (civil or criminal charges utilizing government being one but
not the only possibility), to hold a gun to your head and take that
which is mine back.
>That's the basic fallacy of intellectual property.
>Intellectual property is a government-granted monopoly
>that binds even non-contracting parties.
I don't view the software I create as 'intellectual property',
whatever that is. Software is a tool which I have created. It is MY
tool. It is MY property. If you want to use it, than it must be
because it provides some value to you. You don't have the right to
steal my tools. Not my tractor trailer, not my forklift, and not
software which I have created and possibly licensed specifically to
some third party.
Just because software is easily copied, does not make it any less
tangible.
Feel free to duplicate my efforts as best you can. If you can build a
better mousetrap by all means do so. If you can sell your clone
cheaper, or choose to give it away, do so. I make no copyright or
patent claim.
But don't steal it from me. Don't claim that any third party has the
right to make a binary copy of my work and utilize or sell it without
my permission.
>But if you publish it, I deny you the right to sue people
>who use, copy, modify, and redistribute it.
Before you can deny me a right, you would need to tell me who would
enforce YOUR denial. Government?
I have not entered into a contract with you giving you permission to
deny me anything. Denying me a right, is asserting a right for
yourself that would need to be enforced.
Obviously to deny me that right, even you need some rights enforcing
mechanism<g>! Do you have some 'third party' rights enforcing
mechanism which I lack?
The fact is that without property rights you and I have nothing. Your
life is probably your most valuble property. If you have no property
rights, than ultimately you don't even have the right to your life.
Any third party would have the right to impress you into slavery and
force you to labor for them.
In my world, no third party has rights to my productive efforts
without mutual agreement.
It is MY choice to decide who can utilize the fruits of my productive
efforts and I can set the value for which I am willing to exchange
them.
If in your universe, you refuse to recognize my right to my life and
my property than I, as well as many other like minded citizens, would
be forced to take matters into our own hands. Even in an anarchistic
society, there can be respect for property rights, and enforcement of
simple morality, whether you call that government or not. In the
'wild west' they may not have had much government, or even a sheriff,
but they still knew what to do with a horse thief.
Eric Langjahr <········@itsdata.com> wrote in message news:<··································@4ax.com>...
> If I leave numerous valuable items on the street, in an unprotected
> state, your taking them is still THEFT. You knew they were not yours
> to take or use. The fact that I didn't have them under lock and key
> doesn't change the nature of your act.
> Just because software is easily copied, does not make it any less
> tangible.
Software could not possibly be any less tangible than it already is.
You talk about software as if it were a physical thing, but it isn't.
You can't leave a piece of software in the street. You can leave a
disk in the street, but the disk is not software. Taking the disk is
fundamentally different from making a copy of the information on that
disk. For one thing, if I take the disk then you no longer have it;
I have deprived you of the use of that disk. But if I copy the software
you still have it. This is not to say that copying is morally justified,
just that you can't glibly dismiss the complexities of the morality of
copying by drawing an analogy with physical objects. The physical
laws of software are different from the physical laws of things, and so
are the ethics.
> Feel free to duplicate my efforts as best you can.
What if the best way to duplicate your efforts is to copy your code?
What if I make that copy not in the usual way, but by studying the
code, memorizing it, and painstakingly typing in a brand new version?
What if I don't memorize the code verbatim, but just its general
structure and duplicate that? What if I don't duplicate its general
structure but just its functionality (cf. the Apple look-and-feel
lawsuits)? At what point do I cross the line from moral to immoral
conduct?
Erann Gat
···@flownet.com
On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
>At what point do I cross the line from moral to immoral
>conduct?
It's a shame that you don't know that
Eric Langjahr <········@itsdata.com> writes:
> On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
>
> >At what point do I cross the line from moral to immoral
> >conduct?
> It's a shame that you don't know that
But you still have not answered the question. I'm quite interested in
the answer (and don't feel a slightest shame about that).
--
Janis Dzerins
If million people say a stupid thing it's still a stupid thing.
On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
>I have deprived you of the use of that disk. But if I copy the software
>you still have it.
Theft or stealing, according to my dictionary, has nothing to do with
whether I have been deprived of the use of my property. It has only
to do with the fact that my property has been taken.
So it comes down to a few simple questions?
1. Do you respect property rights?
2. Do you recognize that software which I have created is my property?
If software which I have created and licensed to one specific entity
for a specific usage is not MY property, whose property is it?
Surely you don't think the answer is YOURS? How did it become yours?
By virtue of the fact that you could easily steal it?
Surely you don't think that I would bother to create something that
became your property without mutually agreed upon compensation.
Trust me, I wouldn't. I'd rather dig ditches or flip burgers than be
forced to give my software, to the likes of you.
Obviously only by a twisted leap of faith could you claim it to be
yours. As such I can only understand the position that you, and
others like you take, as one based on religous fanaticism and not
rational thought.
That's the last I am going to say on this. We are clearly off topic
and obviously, like all usenet discussions, this one is pointless. I
am not going to change your mind and you are certainly not going to
change mine.
Eric Langjahr <········@itsdata.com> wrote in message news:<··································@4ax.com>...
> On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
>
> >I have deprived you of the use of that disk. But if I copy the software
> >you still have it.
>
> Theft or stealing, according to my dictionary, has nothing to do with
> whether I have been deprived of the use of my property. It has only
> to do with the fact that my property has been taken.
>
> So it comes down to a few simple questions?
>
> 1. Do you respect property rights?
> 2. Do you recognize that software which I have created is my property?
>
> If software which I have created and licensed to one specific entity
> for a specific usage is not MY property, whose property is it?
>
> Surely you don't think the answer is YOURS? How did it become yours?
> By virtue of the fact that you could easily steal it?
You have clearly misunderstood me. I did not say, nor did I mean to
imply, that it was morally justified to make an unauthorized copy of your
code. We actually *agree* on that (I think). The only point I was trying
to make is that taking software is fundamentally different in important
ways from taking physical objects, so you can't use the theft of a physical
object as an analogy to provide moral guidance in the case of software.
You have to make your argument some other way -- if you want to be effective
that is. If all you want to do is rant and rave then by all means carry on
as you have.
Erann Gat
···@flownet.com
Eric Langjahr <········@itsdata.com> writes:
> On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
>
> >I have deprived you of the use of that disk. But if I copy the software
> >you still have it.
>
> Theft or stealing, according to my dictionary, has nothing to do with
> whether I have been deprived of the use of my property. It has only
> to do with the fact that my property has been taken.
But if one copies your software you still have it. So it has not been
taken away from you, not even talking about stealing.
> So it comes down to a few simple questions?
>
> 1. Do you respect property rights?
> 2. Do you recognize that software which I have created is my property?
I don't know much about property rights but if we're talking about
software then intellectual property rights is what should be called
upon.
> If software which I have created and licensed to one specific entity
> for a specific usage is not MY property, whose property is it?
So we are talking about licenses not software.
> Surely you don't think the answer is YOURS? How did it become yours?
> By virtue of the fact that you could easily steal it?
Copying is not stealing -- it's making another copy.
> Obviously only by a twisted leap of faith could you claim it to be
> yours. As such I can only understand the position that you, and
> others like you take, as one based on religous fanaticism and not
> rational thought.
But the point was -- if I don't take (steal, copy) your software, but
just reproduce it -- is it yours or mine? And to what degree? (So
where's the line?)
> I am not going to change your mind and you are certainly not going
> to change mine.
I call this stupid.
--
Janis Dzerins
If million people say a stupid thing it's still a stupid thing.
From: Kent M Pitman
Subject: Re: LISP and Free Software (was: evolving lisp)
Date:
Message-ID: <sfwu22cm0v2.fsf@world.std.com>
Janis Dzerins <·····@latnet.lv> writes:
> Eric Langjahr <········@itsdata.com> writes:
>
> > On 23 May 2001 08:12:26 -0700, ···@flownet.com (Erann Gat) wrote:
> >
> > >I have deprived you of the use of that disk. But if I copy the software
> > >you still have it.
> >
> > Theft or stealing, according to my dictionary, has nothing to do with
> > whether I have been deprived of the use of my property. It has only
> > to do with the fact that my property has been taken.
>
> But if one copies your software you still have it. So it has not been
> taken away from you, not even talking about stealing.
It isn't the action but the consequence that's relevant. You can play all
the word games you want, but if you "copy" the password to my computer or
the combination to my safe, there are still material consequences to me.
If you take a piece of software I wrote but did not authorize you to
use, then either you are just doing this for spite or you are gaining
value. If you are gaining value, then you have an unfair market
advantage over people who are using that same value but paying for it.
If you insist that they should compensate for that unfair advantage by
stealing, sorry--copying, my software, too, then pretty soon I will
have no paying customers because the terms of your arrangement, if
allowed to happen, are better than what I was offering. But,
paradoxically, if there was going to be no one to pay, I would never
have made the software. So there would be nothing for you to steal.
In effect, it is the fact that it is inappropriate for you to steal
which enables the creation.
Sure, some people might just contribute for fun. But eventually they
will get tired of not eating, and their would-be contribution to
computer science will be compromised by a need to spend their days
doing other things.
Personally, I find there are few enough good books that I'd very much
like to make it possible for the good authors to spend full-time
writing--books they choose to write, not books someone contracts them
to write. And I feel the same about programmers and programs. Some
of these people will succeed, and some will fail. The market doesn't
guarantee them success--just a chance. Without at least the hope of
success, they're not going to try.
In article <···············@world.std.com>,
Kent M Pitman <······@world.std.com> wrote:
>Janis Dzerins <·····@latnet.lv> writes:
>
>> But if one copies your software you still have it. So it has not been
>> taken away from you, not even talking about stealing.
>
>It isn't the action but the consequence that's relevant. You can play all
>the word games you want, but if you "copy" the password to my computer or
>the combination to my safe, there are still material consequences to me.
>
>If you take a piece of software I wrote but did not authorize you to
>use, then either you are just doing this for spite or you are gaining
>value. If you are gaining value, then you have an unfair market
>advantage over people who are using that same value but paying for it.
>If you insist that they should compensate for that unfair advantage by
>stealing, sorry--copying, my software, too, then pretty soon I will
>have no paying customers because the terms of your arrangement, if
>allowed to happen, are better than what I was offering. But,
>paradoxically, if there was going to be no one to pay, I would never
>have made the software. So there would be nothing for you to steal.
>In effect, it is the fact that it is inappropriate for you to steal
>which enables the creation.
>
In other words, you're invoking the Kantian categorical imperative,
which I find a very strong argument.
>Sure, some people might just contribute for fun. But eventually they
>will get tired of not eating, and their would-be contribution to
>computer science will be compromised by a need to spend their days
>doing other things.
>
As a matter of empirical fact, there is a whole lot of very useful
free software (for varying definitions of free) available. This includes
operating systems, language implementations, editors, networking
software, games, whatever. Since this has been continuing for
decades, it isn't just a passing fad.
As I've been arguing in another group, if it turns out that
free software takes over so that it is effectively impossible
to create commercial software, then that's a market decision and
commercial software lost fair and square. I don't expect that
to happen.
Here's my take on it:
The current commercial software market is based on artificial
scarcity. There is a product, software, that can be extremely
valuable and can be mass produced incredibly cheaply. If more
copies are made, the world is richer. If all software that
was produced was freely distributed, the world would be richer.
The trick is to make sure that people who make things of great
value get a cut.
Remember, here, that something that can be very widely
distributed is more valuable because of it. A baseball player
earns his salary by his ability to entertain up to tens
of millions of people at once, not because such entertainment
is itself more valuable than dentistry or teaching. The
Gnu Project has contributed an immense amount of value to
the world.
Let's trace this in commercial software. Digitool sells copies
of MCL and assorted licenses to Joe's Bait and Software for
a few thousand dollars. Joe writes software that he sells for
hundreds of thousands of dollars. Joe has obviously created
something of considerable value, with a considerable assist from
Digitool. Joe presumably got a large cut, but Digitool gets
the few thousand he paid. (I understand that Franz works with
a different model.) This is inequitable as it stands, but
at least Digitool is getting some money. (In real life, they've
been hard hit by converting to MacOSX, and I suspect could really
use more money.)
Switch to the free software model. In this case, many people
other than Joe use MCL, and write even more valuable things,
so Digitool is contributing great value. On the other hand,
Digitool is suddenly not getting paid for this value even as
much as it was paid for the lesser value.
What I think it comes down to is practicality. Free software
would be the ideal solution, if people could get paid in proportion
to the value they create. If Kent writes some tools, charges $100
for them, sells them to 200 people, he's created over $20K of
value and received revenue (note the difference between revenue
and profit) of $20K. If he put them out on a web site for free use,
more people might use them, but Kent wouldn't get paid.
Find some way of giving Kent a cut of the value he creates with
free software and we've got a technique that will revolutionalize
the economy. Otherwise, I'm going to continue to support his right
to sell software one copy at a time.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
Kent M Pitman <······@world.std.com> writes:
> Sure, some people might just contribute for fun. But eventually
> they will get tired of not eating, and their would-be contribution
> to computer science will be compromised by a need to spend their
> days doing other things.
I don't know if you intend it, but this implies that the only
sustainable way to develop software is to get paid for the licensing
of it. My own situation is a counter-example, a mixture if you will.
I used to get paid hourly to produce things which then were released
as Free Software. I now am producing Free Software and am salaried,
part of the monies for my salary come from licensing of a proprietary
product. I know of people who are in siiliar situations, and I also
know of people who are salaried and no appreciable amount of their
monies comes from proprietary licensing.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
From: Kent M Pitman
Subject: Re: LISP and Free Software (was: evolving lisp)
Date:
Message-ID: <sfw1ypf6b9f.fsf@world.std.com>
Craig Brozefsky <·····@red-bean.com> writes:
> I don't know if you intend it, but this implies that the only
> sustainable way to develop software is to get paid for the licensing
> of it. My own situation is a counter-example, a mixture if you will.
> I used to get paid hourly to produce things which then were released
> as Free Software.
No, I know there are situations like this. I mean to suggest that having
to work on projects that have some other purpose and take the side-effects
of that is not what I mean. In such a structure, if I want to make a couple
of libraries, I have to find someone who wants them. I'm not free to just
pursue them on my own unless I want to do it for no compensation. Doesn't
sound like free to me...
In article <···············@world.std.com>, Kent M Pitman wrote:
> ...
>of that is not what I mean. In such a structure, if I want to make a couple
>of libraries, I have to find someone who wants them. I'm not free to just
>pursue them on my own unless I want to do it for no compensation. Doesn't
>sound like free to me...
let's assume you have written a couple of libraries which are useful and want
to sell. for some reason somebody else has written similar libraries with
comparable functionality in support for a specific project. s/he is not
particularly interested in marketing, but figures if he releases it under
gpl like terms, he does the world a favor *and* can pick up all the
improvements other people make to it while concentrating on his real
project. probably a good business decision for him, but it has the
potential of killing your business. but would you consider his decision
legitimate or would you feel wronged?
--
hs
----------------------------------------------------------------
"The cheapest pride is national pride. I demonstrates the lack of
characteristics and achievements you can be proud of. The worst loser
can have national pride" - Schopenhauer
Followup-To: ············@tunes.org
Once again, this is getting off-topic from comp.lang.lisp,
and I recommend moving the discussion to another forum, such as
http://lists.tunes.org/mailman/listinfo/cybernethics
>>: Fare Rideau <···········@tunes.org>
>: Eric Langjahr <········@itsdata.com>
>> Sue your licensees if you dare; but don't you sue third parties.
> Would you suggest, that if I were to leave my car in your garage, and
> it was stolen, that I would have NO right to pursue the third party
> who stole it? Only YOU for letting it be stolen?
Indeed, if I accept the responsibility for safely guarding
your car in my garage, then *I* am the one you should sue,
should the car be stolen. And should the thief be caught,
*I* am the one to sue the thief. And if either of us is insured
against such events, then the insurance in turn gets the right to sue,
and we can no more sue but for the part not covered by the insurance.
Again, in as much as IP is born in contracts,
you shouldn't be able to sue non-contractors.
In as much as it is a governmental privilege,
it should be abolished as the protectionism it is.
If it were a natural property, it would have to be protected
not just for 17 or 90 years, but forever.
> If I leave numerous valuable items on the street, in an unprotected
> state, your taking them is still THEFT.
Taking them sure is theft, but copying them is not.
And you sure don't get the right to search and seize anyone's property
because you were stolen while negligent, nor to seize anyone else's property
(through taxes) to have police protect you.
> If you steal from me, I am quite prepared in the absence of other
> recourse (civil or criminal charges utilizing government being one but
> not the only possibility), to hold a gun to your head and take that
> which is mine back.
And rightly so. Now, mind the "IF":
There remains to establish whether copying is theft.
>> That's the basic fallacy of intellectual property.
>> Intellectual property is a government-granted monopoly
>> that binds even non-contracting parties.
>
> I don't view the software I create as 'intellectual property',
> whatever that is. Software is a tool which I have created. It is MY
> tool. It is MY property.
Yes, and MY copy is MY tool and MY property, too;
you shouldn't deprive me from MY right to use it.
> If you want to use it, than it must be
> because it provides some value to you.
Don't confuse VALUE and UTILITY.
UTILITY is about ordering states of the world according to preferences;
VALUE is about exchanging services to provide each one with utility,
at a rate established by the balance of market.
See around paragraph 15 of the following chapter of "Economic Harmonies":
http://www.econlib.org/library/Bastiat/basHar11.html
> You don't have the right to steal my tools.
How am I stealing them?
Property in ideas is an insoluble contradiction. [He who complains of
"theft" of his idea] complains that something has been stolen which he
still possesses, and he wants back something which, if given to him a
thousand times, would add nothing to his possession.
-- H. Rentzsch, Geistiges Eigenthum, 1866.
> Just because software is easily copied, does not make it any less tangible.
Either your sentence is a self contradiction, or it is begging the question.
What is tangibility, and how is software tangible, to begin with?
> Don't claim that any third party has the
> right to make a binary copy of my work and utilize or sell it without
> my permission.
Of course they can. They can also sing, speak english, grow crops,
and do everything they want without your permission. As long as they
do it using stuff they own, like for example, a nth-generation copy
of your original software. Once you sold a copy, it's not yours anymore.
Just like your selling farming tools doesn't earn you a perpetual right
on the way the farmer uses them, your selling software doesn't earn you
zilch on the way people use the copy you hand them.
Oh, you may negociate a contract with them, but in absence of contract,
they shouldn't be bound, just like their shouldn't be a law protecting
tool-makers from non-permitted use of their tools by farmers lest
they condescend to grant the farmers a license.
Of course, in presence of tool-protectionism, after a few generations
of indentured farmers, the new ones will know better and only buy
from tool-makers who grant Free Tool licenses, so the society will
outgrow the iniquitous law. Yet the law will have made victims,
grown monopolies, encouraged people to lobby for more protection, etc.
>> But if you publish it, I deny you the right to sue people
>> who use, copy, modify, and redistribute it.
>
> Before you can deny me a right, you would need to tell me who would
> enforce YOUR denial. Government?
>
Self-defence. If you try to enter MY home, seize MY computer,
and MY copy of some software whatsoever, etc.,
I'll greet you with bits of high-speed metal.
Government is but a way for people to associate in self-defence.
I'm as much of a property-respectful libertarian as you can ever be,
so please no straw man argument here.
So it all comes down to: what is property, and what is theft?
And please, just because a piece of paper makes it "legal"
doesn't make it any righter or wronger.
What is property, out of natural right?
And what is but government protectionism,
stolen from the public at large to be granted to a privileged few?
You're begging the question.
You say intellectual property is natural property, I argue it isn't.
No need to make a straw man argument about independent premises
like my respect of property in general.
> I have not entered into a contract with you giving you permission to
> deny me anything.
The right for me to copy your work is no natural property of yours
to give away to me by contract, to begin with.
> Denying me a right, is asserting a right for
> yourself that would need to be enforced.
I can use the very same argument, with as much strenght or maybe more so,
to defend my right to use, copy and redistribute software:
are you claiming the right to seize anyone's computer
and digital media so as to enforce your alleged right to prevent copy?
Whence does any police get the right to enter my property
for search and seizure?
> Obviously to deny me that right,
> even you need some rights enforcing mechanism<g>!
> Do you have some 'third party' rights enforcing mechanism which I lack?
Self-defence, organized as "government" or not, is the only mechanism needed.
Just you come claim your "right" to control my use of my computer.
The question is not whether there is an enforcement mechanism. The question
is what are the rights to enforce, and whence these rights stem from.
Can government originate new rights, like "intellectual property rights"?
I contend that it cannot, and than when it does, it is detrimental to
everyone in the long run, and that society will find ways around it,
like mutual agreements not to use these rights (e.g. free software licenses).
> The fact is that without property rights you and I have nothing.
> Your life is probably your most valuble property. If you have no property
> rights, than ultimately you don't even have the right to your life.
> Any third party would have the right to impress you into slavery and
> force you to labor for them.
I fully agree with this stance. This is not the premise I question.
> In my world, no third party has rights to my productive efforts
> without mutual agreement.
Copying your software, modifying it, and redistributing it was MY effort,
done on MY computer, with MY digital media, MY electricity, MY time,
MY creative thinking, without any additional effort required on your part.
You're denying me the right to my own productive efforts,
without mutual agreement.
> It is MY choice to decide who can utilize the fruits of my productive
> efforts and I can set the value for which I am willing to exchange
> them.
No it isn't. It is your choice to decide whether you are to make
productive efforts or not, but once you agreed to do these efforts,
you have no right to decide how other people will take advantage
of their result (least of course they agreed to it in advance).
Similarly, you may keep your recipees secret,
but if someone manages to bluntly copy them or to reverse engineer them,
or to enhance them, you have no right to prevent him from using them,
unless he specifically yielded his rights to you as part of a contract.
> If in your universe, you refuse to recognize my right to my life and
> my property than I, as well as many other like minded citizens, would
> be forced to take matters into our own hands. Even in an anarchistic
> society, there can be respect for property rights, and enforcement of
> simple morality, whether you call that government or not. In the
> 'wild west' they may not have had much government, or even a sheriff,
> but they still knew what to do with a horse thief.
Again, I fully agree with this stance.
My computer, my mind, etc, are my property.
If you're trying to rob me from the right to use them as I see fit,
I'll have to fight you back.
So, can you have a serious discussion on the heart of the problem,
instead of raising straw man arguments about a side question?
It all boils down to: is intellectual property natural property,
or is it a government-granted monopoly? How do you identify property? etc.
Among the many pointers in
http://fare.tunes.org/articles/patents.html#pointers
I particularly recommend you Roderick T. Long's
"The Libertarian Case against Intellectual Property".
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
I do agree that I have to pay for the *opportunity* to read a book or to use
a program; I do not agree that I have to pay for the *right* to do so.
From: George Neuner
Subject: Re: LISP and Free Software (was: evolving lisp)
Date:
Message-ID: <3b0c3c4a.391906732@helice>
On Wed, 23 May 2001 00:38:56 -0700, Eric Langjahr
<········@itsdata.com> wrote:
>If I leave numerous valuable items on the street, in an unprotected
>state, your taking them is still THEFT. You knew they were not yours
>to take or use. The fact that I didn't have them under lock and key
>doesn't change the nature of your act.
You are conflating ethics, morality and the law.
If you leave items on the street which cannot readily[1] be identified
as belonging to you, legally they are "salvage" and anyone is entitled
to claim them.
Morally, there may be an issue of "theft" in the sense of "covet thy
neighbor's ...". Ethically there may be responsibility to return
found items to the rightful owner. Legally there is neither and the
salvager may lawfully claim compensation for effort (reward) if the
owner is eventually found and the items returned.
[1] "readily" is defined in this context as involving reasonable
effort or querying a known authority. Your car may be identified by
by its registration or VIN numbers. Your diamond ring, unless it is
engraved or registered, is unidentifiable to anyone but (maybe) you.
>If you steal from me, I am quite prepared in the absence of other
>recourse (civil or criminal charges utilizing government being one but
>not the only possibility), to hold a gun to your head and take that
>which is mine back.
Regardless of circumstances, you can expect to find yourself quickly
dead if you point a weapon at me.
>I don't view the software I create as 'intellectual property',
Then what is it?
>It is MY property.
Explain your notion of "property" and why you alone are conferred
unalienable rights to it.
>Before you can deny me a right, you would need to tell me who would
>enforce YOUR denial. Government?
Before anyone can deny you a right, you have to prove you have it in
the first place. What authority grants you the right to have a right?
Who enforces that?
>I have not entered into a contract with you giving you permission to
>deny me anything. Denying me a right, is asserting a right for
>yourself that would need to be enforced.
Conversely, by that same lack of contract, I have not given you
permission to deny me anything. Funny how that works ...
>In my world, no third party has rights to my productive efforts
>without mutual agreement.
So you don't pay taxes? Where is it that you live?
George
--------------------------------------------------------------------
Disclaimer: I am not an Intellectual Property attorney. My father,
my sister and a dozen or so friends of family are.
--------------------------------------------------------------------
[I missed replying to this first time around, sorry. And then you made
another of the same remark in a later post, so I guess this warrants reply
after all.]
Peter Wood <··········@worldonline.dk> writes:
> Work once, get paid forever.
There is no chance this could happen. Suppose I make a piece of
software and sell it for $100. Market forces ALREADY assure that I
won't get $100 forever. One of two things is true: either it has
limited utility and the market will naturally fall off for lack of
demand, or someone who can multiply will eventually decide it's worth
it to do the same fixed-cost work in exchange for $90 or $80 or some
lesser figure that will beat out the person getting $100 and that will
get them a continuous stream at the new rate. This process will iterate
until it decays to a point where the cost charged is something no one
feels it worth competing with, which is the true market value.
Essentially, every year, a vendor has to put more and more into the
same product to get the same dollar on the market that they originally
got. The market won't allow a product to continue to be sold at an
unfairly high price. Photoshop 6.0 doesn't command orders of
magnitude more money than Photoshop 1.0, even though it has probably
orders of magnitude more functionality. This is because of exactly
the effect I said. Once someone has made the thing, and someone else
sees a profit being made, the market naturally invites others who can
deal with slightly less profit to move in. That's natural market
forces. The original guy only stays in business if he keeps investing
because either the price comes down or the functionality goes up on
the original product.
And we have laws against "dumping" because it can cause legitimate
businesses to go out of business, but for some reason they are not
applied to people giving away entirely volunteer labor, even though
such "gifting" has the same economic power to ruin that "dumping" does.
Also, your argument makes it sound as if what these people want to do
is get rich and sip Pina Coladas on the beach. (Not that I think
there's anything immoral about that. Makers of Pina Coladas and
yachts have to make a living, too.) But many who aspire to riches
don't do it for purely personal reasons. They do it because the
riches will enable them to pursue a vision in a way that they simply
cannot do if they are dependent on others for their income. Pursuing
a workaday income makes it a good bet most people will not to have the
energy left over to do great things. And for those who can do great
things in their free time, imagine what they could do if all of their
time was their own...
Personally, I've worked my whole life for others and have been
saddened by watching things not go technically in directions I want
them to go. I don't see how free software will help me fix that
because you can give me all the free software you want, but if I don't
get *time*, I can do nothing with it. To me, power is not
software--because given time, I can WRITE software. Power is control
of my own time and resources. Both are controlled by money. Either
my time/resources are my own, or they are not. An argument that all
software should be free is, to a creator of software, an argument that
I should be in another business. Because no amount of anything I do
will ever get me more than a day's wage. If I were a writer, I could
aspire to write a best-seller. If I were a musician, a platinum
record. As a carpenter, I could make finely crafted bookcases that
people might pay a lot for. But as a programmer, I must make only
commodities. No sir, this is not incentive.
I would and do work on what I do regardless of whether I'm paid or
not. But I can only do that with time I control. When I'm enough
not-paid, like now, I'm eventually forced to go work for someone who
has a different theory than I have about where the world should go.
What supports the possibillty of my vision of the future ever happening
instead of someone eles's is my ability to tell the people who would
employ me -- that is, the people who would control my time -- "sorry,
I don't need you". But if I cannot rely on the economic forces of the
market to keep me free in that way, then that's the end of my freedom.
And I *must* make more than my month's rent to do that, because that doesn't
empower me to invest ahead. That just gets me to the end of the month.
Going out on my own--doing my thing and not someone else's--REQUIRES
that I have enough money buffered to keep me taken care of for a long
time while I investigate my own need. So any system that just rewards me
for a day's work does not ever enable the flexibility for me to explore
my own path.
Money is freedom. The push for all software to be free seems a
counter-culture attempt to topple the folks in power. But what
results? A few companies who've built a commercial name on other
platforms see big bucks selling their wares ALSO on Linux. But how
many NEW products have emerged on Linux, charging dollars and getting
away with it? We keep hearing the same three success stories for open
software--apache, gnu, redhat--even fewer success stories than you
here touted for Lisp, btw; at least THAT'S heartening. I don't see
Linux as a hotbed of much new commercial activity, for all it has that
software that's so free and able to support unbounded growth. And I
think part of it is that, incrementally, no one wants to charge, I
think because they're embarrassed to. And so everyone contributes
free stuff. And the free stuff grows, but empowers no one
economically, except maybe the hobbyist...
Incidentally, I'm all for very short timelines on copyrights, and I
don't believe in software patents. Don't take my remarks as being an
endorsement of the present long-term entanglement of IP that is going
on. But I still beleive the concept of copyright is important and I'd
rather fix it structurally by adjusting its balance knobs in a way
that causes it to do the right thing socially than throw it out, which
in my opinion is extreme and damaging. I don't believe intellectual
property rights should keep IP locked in prison forever, but I do
think it has to protect someone's investment long enough to get time
to market and a reasonable return, enough to make it worthwhile to
have bothered.
It's nice when people provide free things, but that
doesn't mean people should expect free things. I'd rather see a
culture in which people who felt they got something felt morally
obligated to pass money toward the source of value than one in which
people were taught that it was their moral right to get things free.
About the only moral right I think people have is to stand in the
middle of the jungle, cold, alone and unclothed and be eaten by a
tiger. Everything else that people do in the world is an evolved
notion of cooperation among people to construct something better. Now
you can evolve a toy economy where people share software back and
forth among each other ad infinitum and call it cooperation, but if I
can't trade software for food, clothing, shelter, or other things that
money buys, then your "software economy" is just based on Monopoly(R)
money and has no real value. Oh, you can do that if you want--create
a trade in programs that is outside the market of trade in money, but
in so doing you exempt any progress in programming from ever affecting
how people earn food. And in doing that, all you do is assure that
programming will never be a profession, only an avocation. Programmers
becomes the new peasants of the modern farm, never able to build anything
that will help them climb out of their little plot. I don't see that
as a contribution to the democratization of society.
Kent M Pitman <······@world.std.com> writes:
> We keep hearing the same three success stories for open
> software--apache, gnu, redhat--even fewer success stories than you
> here touted for Lisp, btw; at least THAT'S heartening.
Actually, who is making money out of redhat? I know commercial
organisations (like redhat) exist that support it but last time I
looked they were bleeding money. Perhaps they are doing better now,
since people have realised that you can't live by share-price alone.
--tim
>>>>> "Tim" == Tim Bradshaw <···@tfeb.org> writes:
Tim> Kent M Pitman <······@world.std.com> writes:
>> We keep hearing the same three success stories for open
>> software--apache, gnu, redhat--even fewer success stories than you
>> here touted for Lisp, btw; at least THAT'S heartening.
Tim> Actually, who is making money out of redhat? I know commercial
Tim> organisations (like redhat) exist that support it but last time I
Tim> looked they were bleeding money. Perhaps they are doing better now,
Tim> since people have realised that you can't live by share-price alone.
I think they're doing better. However, I'm pretty sure Cygnus was
making money before Redhat bought them.
Ray
Tim Bradshaw <···@tfeb.org> writes:
> Kent M Pitman <······@world.std.com> writes:
>
> > We keep hearing the same three success stories for open
> > software--apache, gnu, redhat--even fewer success stories than you
> > here touted for Lisp, btw; at least THAT'S heartening.
>
> Actually, who is making money out of redhat? I know commercial
> organisations (like redhat) exist that support it but last time I
> looked they were bleeding money. Perhaps they are doing better now,
> since people have realised that you can't live by share-price alone.
Well, it seems that Red Hat has reached a slightly red 0 in the last
quarter, but I'd speculate that most of the income that contributed
to this result was coming from the Cygnus side of the business. I
think Cygnus is as successful as it is, because gcc and the whole GNU
toolchain as such has little value to its customers, it is the
continued porting and support that drives income. In other words
Cygnus works as an open-source business because there are more
platforms than customers in the embedded market, and hence gcc can't
usefully be shared between customers and non-customers. All that work
that affects the mainstream platforms of gcc is just a side-benefit.
It remains to be seen how well the Red Hat side of the business can
adopt a services-oriented business model that is successful in the
long-term.
In any case with free-software, since you can only live on support
contracts and work for hire, it remains doubtful how truly innovative
software (no I'm not a Microsoftie), i.e. software for which no market
exists yet, can be built. If I have to invest $10-20 million in order
to build this new piece of software, how do I recover those costs if I
have to release the software for free, and can only earn money for the
support work and enhancement work on the software? Especially if the
market for the software is limited?
Probably truly new software will only come out of the hobbyist and
university areas, and or maybe if I can convince enough existing
customers that they need the software in question, without it existing
yet. Strangely it is fairly difficult to imagine how e.g. Visicalc
might have been created in such a setting.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
"Pierre R. Mai" <····@acm.org> writes:
>
> In any case with free-software, since you can only live on support
> contracts and work for hire, it remains doubtful how truly innovative
> software (no I'm not a Microsoftie), i.e. software for which no market
> exists yet, can be built. If I have to invest $10-20 million in order
> to build this new piece of software, how do I recover those costs if I
> have to release the software for free, and can only earn money for the
> support work and enhancement work on the software? Especially if the
> market for the software is limited?
I think this is the point. The free software economy, from the
programmer's point of view, is a service economy: you sell your time,
once. No-one gets really rich by selling their time, so programmers
in the free software economy don't get really rich, or equivalently
they don't ever free up time to do other stuff.
And I think this really explains a lot. For a lot of people -
especially free-software people - there is almost nothing they'd
rather do than hack 14 hours a day, so this is just fine - they get
paid to do what they'd do anyway. But if you want to free up time to
do something else, you need to not be charging for time but building
up value. And One of Kent's points is, I think, that if you are
charging for time but someone else is building up value, that other
person is going to end up telling you what to do.
For me, I'm acutely aware of this kind of thing. I sell my time, and
I make a pretty good living off it, but I've spent enough of my life
staring at a screen for almost all my waking hours and not doing all
the other stuff I want to do that I'm pretty interested in anything
that gets me out of this (:-).
--tim
Kent M Pitman <······@world.std.com> wrote in message news:<···············@world.std.com>...
> I think teaching college kids that it's good to give stuff away causes
> them not to understand teh value of what they do, and causes them to
> economically injure those people who are producing that stuff for money.
> I don't see that as doing any good for anyone except the people who are
> sitting around waiting for the freebies to be produced, and the whole system
> is fueled by people who apparently need no income. That's not a stable
> system. I can't figure out where the fuel is.
This has been a really interesting thread, it is enlightening to read
so many well thought out points of view on this subject.
I have read one account that said that an early goal of the Free
Software movement was specifically to cause economic injury to
Symbolics as a protest against them not sharing their research and
code. However, I think most college students that do participate in
Free projects have other goals:
- To offer something to the community in compensation for having
access to the many tools that they haved learned, used, and profited
from. They realize that their education was enhanced by having access
to tools like Linux, Emacs, GCC, CLISP, etc.
- To show off their code, impress people, and generally take a step up
the ladder from student to recognized guru. I think most programmers
enjoy showing their work and getting slaps on the back from their
colleagues.
- To benefit from the talents of others. Sometimes you can reach a
point on a pet project where you are not sure what the best next step
would be, or you having something interesting to share but you know it
is not ready for 'prime time.' Sharing your code with other people
allows you to get the benefit of their comments and incites.
I think most of us need income, but sometimes employers are willing to
support Free software development or do not mind if a specific piece
of code is shared with the rest of the world. Many folks like writing
software on their own time as well and want the opportunity to share
it with others. Personally, I miss the old days of Apple ][ user
groups where we used to sit and share programs that we had written in
AppleSoft BASIC. I don't think we considered it Free software (and I
definitely don't think we ever injured anyone else economically), but
all three of these motives were important.
Thank you for your time and patience. I hope this makes sense.
Kent M Pitman <······@world.std.com> writes:
> Craig Brozefsky <·····@red-bean.com> writes:
>
> > Oh come on Ken.
>
> [Kent]
Typo. No really, I haven't been asleep over the last four years I've
been reading c.l.l 8)
> > Sure there are some people who clamor for people to volounteer their
> > work, but that can't be attributed to "free efforts".
>
> Yes, it can. A huge number of people are coming out of college these
> days thinking that free stuff is an entitlement and that there is
> something wrong with a community that doesn't provide it to them.
That doesn't address my initial point, that this confuses those who
are doing open work and advocating the use of open licenses from those
who are just asking for people for "free" stuff. We both know that
nothing is free, that open code requires alot of labor.
> > Do you really mean you confuse those people who are attempting to
> > develop an open code base and contributing their own labors with the
> > hangers-on that have a million ideas but no code?
>
> Partly I'm saying I don't see you saying how you're in the set of people
> doing something. The key to doing something is not writing posts on why
> it's needed, it's simply doing. You don't have to convince anyone.
Agreed. Pontificating about how the world would be a better place if
everything was open without considering the mechanisms by which the
work will be done and how those people will eat annoys me as much as
it annoys you. The only thing I'm disagreeing on is the attribution
of this annoying behavior to people who are actually writing open
code, or are discussing ways in which to better coordinate their
effort.
People doing the work have to figure out where the labor will come
from. For me on the USQL and IMHO projects it was the time my
employer paid me for to produce toolkits for which we saw no market,
to support a proprietary product. We did not have the money to
market, advertise or support the toolkits. The code was written and
it could either languish in our organization, or it could get outside
users, whom BTW have contributed alot of patches that we found useful
for our main commercial venture.
> > Besides, what real harm is there in asking someone who has written
> > code to release it under an open license?
>
> I think there is considerable harm.
So the asking someone to release their code, regardless of wether they
actually do it or not, is responsible for the lack of a market for the
class of software you are interested in selling? I think that's
giving a bit too much credit to the whiners. I would have to agree
with Miles that social and market conditions, along with new
technologies for distributed software development (that Internet
thing) have much more to do with it. Competing languages like Java,
Python and Perl are using open development as a motive force. They
are succesful and developers are coming to expect the same model in
other languages. CL is no exception.
> I think teaching college kids that it's good to give stuff away
> causes them not to understand teh value of what they do, and causes
> them to economically injure those people who are producing that
> stuff for money. I don't see that as doing any good for anyone
> except the people who are sitting around waiting for the freebies to
> be produced, and the whole system is fueled by people who apparently
> need no income. That's not a stable system. I can't figure out
> where the fuel is.
It doesn't do the people sitting around waiting for freebies any good
either. As for the "people not needing income", that's disingenuous.
I need an income, all my co-workers who helped with the Free tools
we've released need an income, people working on clocc need an income.
We've just found ways to get their income and distribute some of their
code as Free Software at the same time.
> It is this (non)cycle, not anything technical, that is most in
> danger of driving me away from Lisp after this many years of having
> held to it.
Noone wants to alienate you, you are afterall Kent Pitman (I on the
other hand am quite expendable). But you are alienating yourself if
you choose to chastise those doing work to produce open tools, a
sizeable part of the CL community now, because you can't find a market
for things you might have been able to sell a decade ago. The means
of social production of software are changing, and that is having it's
effects on the CL community.
Might I humbly, and I do mean *humbly*, suggest that you look for
another approach for supporting yourself as a CL programmer. For one,
even without the Open Source runup over the last few years, the market
for CL tools is woefully small, especially for baseline tools which
have partially working, or 90% solutions already (XML, FFI, sockets
etc...). Writing an application which is sold to people outside the
CL community may be a better bet. You could use an open model or not.
If you do want to examine open models, I suggest the Free Software
Business mailing list as a starting point for examing business models
using open licenses.
> Certainly it makes me not even the slightest bit interested in
> spending time on the market of creating tools, which is what people
> claim they so desperately need in order to write applications.
> Because the writing of tools is something people say they don't
> "value" as a market. Value means "pay money" in my book. And so
> that makes it less likely, not more likely, tools will get
> created...
If those tools can't be created thru a proprietary model then we need
to find other ways to get them produced. I don't think the market for
CL tools has been healthy enough in the last few years to support
proprietary tool vendors (save the implementation vendors), and we're
not going to change that without expanding the size of the community
considerably. That's not going to happen until we have these tools,
so it's a bootstrapping problem.
So, recognizing this condition people have been working to build those
tools. USQL has had people from all over the globe contributing
patches to work with new DBs, and new CL implementations. People are
working on portable socket libraries, they are working on ways to
organize and distribute packages as well.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
"Kent M Pitman" <······@world.std.com> wrote in message
····················@world.std.com...
> Craig Brozefsky <·····@red-bean.com> writes:
> > [...] what real harm is there in asking someone who has written
> > code to release it under an open license?
>
> I think there is considerable harm.
>
> I have spent a good part of this year only partly employed, trying to
> figure out how to make a business out of things I know for once, rather
> than to automatically join up with some company and make money for it.
> I needed exactly the tools you're talkign about, but I didn't expect
anyone
> to provide them to me, so I wrote my own and now I have a bunch of
> intersting stuff. It bugs me is that there is not a market for people to
> license such tools; I don't hear anone posting here saying they'd pay a
dime
> for any of this stuff.
Do you see dual licensing as a reasonable alternative?
Eg. The Norwegian company Trolltech makes an excellent C++ library called
Qt. For non-commercial porposes on X11, you can use it free of charge and it
comes with full source code. If you want to use it for money-making
ventures, you pay a couple of thousand dollars for a commercial license -
which is, IMO, well worth it.
Seems to me this approach has advantages for everyone. The company gets
publicity and advertising by widespread usage. The community gets a toolkit
that's good enough to build KDE. Students can make their IRC clients and MP3
players (and improve Qt itself, if they wish) without forking out money they
can't afford. And the company doesn't have to make its living by selling
T-shirts, stuffed monkeys and good will.
Unless there are hidden traps I'm not aware of, it seems like a good deal
all round ...
In article <···············@world.std.com>, Kent M Pitman
<······@world.std.com> wrote:
>I think that if there were a culture here of people saying that code was
>worth something, I would have (a) licensed the stuff I needed from someone
>who had written what I needed and (b) been able to license the stuff I
>built
>atop that stuff to someone else. Each of us, adding value, would be
>reimbursed for what we created. But instead, we're all involved in a
>rat-race
>where as we develop anything, someone else wants to develop the same thing
>and give it away. That undercuts those of us who are trying to invest in
>something of value.
Although I'm not such a conspiracy theorist that I think this was
planned, I am enough of a social evolutionist to believe that the Open
Source movement can only be an economic good for two groups (apart from
consumers who want freebies):
1. Those who have enough resources to support a coding effort that loses
money, and has been planned from the start as a money loser or loss
leader. The best example here is Sun and Java (Sun's real revenue comes
from selling servers. Java is a money loser for them).
2. Those who are selling service and support for Open Source software.
Examples here would include comanies like Redhat and IBM.
It remains to be seen whether Open Source will really be a long term
money maker for those in group 2. Of course, it doesn't really matter if
it ever becomes a money maker for those in group 1.
In the meantime, Anyone who wants to code for a living, and doesn't work
for a Sun, and doesn't have a solid plan for making money from support
contracts, had better seriously consider whether going Open Source is
really a good idea economically.
I think the OP lives in a world where the money comes from the project
contract, and the implementation language is just a technical detail. If
we all lived in such a world, then we could all collaborate on an Open
Source Lisp, and use it in our contract work. However, some people's
real expertise and experience (e.g., Kent?) lies in producing the
language tools and implementations that the rest of us use to do our
work. In the Open Source world, how do these people get paid?
In other words, Open Source is about making the common tool we all use
free, collaborating on it, all for mutual benefit. But if you are the
person who only (or mostly) does high quality tools, how do you get paid
in an Open Source world. There is no O'Reilly of Common Lisp to keep
Kent on retainer while he cranks out a great Open Source Common Lisp,
unlike Perl and Larry Wall.
Lisp is a difficult language to implement well (i.e, fast native
compiler, threading, etc.). The experts at it can't make money giving it
away in an open source world. The result is what we have - several good
Open Source implementations, none of which are as good as the better,
partially closed source, and farily expensive, commercial
implementations.
Good Common Lisp implementations aren't free because they're not trivial
to implement, and the implementors, understandably, want to get paid for
spending thousands of person-hours doing so. Until we find a Common Lisp
sugar daddy, the situation is likely to remain this way.
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
In article <·····························@news.ne.mediaone.net>,
Raffael Cavallaro wrote:
> ...
>Although I'm not such a conspiracy theorist that I think this was
>planned, I am enough of a social evolutionist to believe that the Open
>Source movement can only be an economic good for two groups (apart from
>consumers who want freebies):
>
>1. Those who have enough resources to support a coding effort that loses
>money, and has been planned from the start as a money loser or loss
>leader. The best example here is Sun and Java (Sun's real revenue comes
>from selling servers. Java is a money loser for them).
>
>2. Those who are selling service and support for Open Source software.
>Examples here would include comanies like Redhat and IBM.
3. those who write software for specific tasks, use open source software
to make it easier, and don't mind if others use the byproduct of their
work to make their tasks easier.
you could look at it as infrastructure on which application of computer
technology is built
--
hs
----------------------------------------------------------------
"The cheapest pride is national pride. I demonstrates the lack of
characteristics and achievements you can be proud of. The worst loser
can have national pride" - Schopenhauer
On Mon, 21 May 2001 21:57:07 GMT,
Raffael Cavallaro <·······@mediaone.net> wrote:
>I think the OP lives in a world where the money comes from the project
>contract, and the implementation language is just a technical detail. If
>we all lived in such a world, then we could all collaborate on an Open
>Source Lisp, and use it in our contract work. However, some people's
>real expertise and experience (e.g., Kent?) lies in producing the
>language tools and implementations that the rest of us use to do our
>work. In the Open Source world, how do these people get paid?
>
>In other words, Open Source is about making the common tool we all use
>free, collaborating on it, all for mutual benefit. But if you are the
>person who only (or mostly) does high quality tools, how do you get paid
>in an Open Source world. There is no O'Reilly of Common Lisp to keep
>Kent on retainer while he cranks out a great Open Source Common Lisp,
>unlike Perl and Larry Wall.
I think the real question is not whether or not the people that develop
commercial Lisp tools *deserve* to making a living doing it. The real
question is how likely they are to be able to continue to do it. If it
were up to me all the people like Kent that have made CL what it is today
would be the millionaires and the geek folk heroes. This doesn't seem to
me to be the way the wind is blowing these days and that's why I think we
need to find a new model. I don't think it's within the power of the
commercial vendors to sustain Lisp.
I don't have a problem with proprietary tools or applications and I would
be very happy to be sitting in a room full of hackers with Xanalys or Franz
Lisp on their desktops. I just think that larger forces beyond our control
have invalidated this business model. We may not like it but I think we
have to come to terms with it.
--
miles egan
Raffael Cavallaro <·······@mediaone.net> writes:
> Although I'm not such a conspiracy theorist that I think this was
> planned, I am enough of a social evolutionist to believe that the Open
> Source movement can only be an economic good for two groups (apart from
> consumers who want freebies):
>
> 1. Those who have enough resources to support a coding effort that loses
> money, and has been planned from the start as a money loser or loss
> leader. The best example here is Sun and Java (Sun's real revenue comes
> from selling servers. Java is a money loser for them).
>
> 2. Those who are selling service and support for Open Source software.
> Examples here would include comanies like Redhat and IBM.
3. Those who develop libraries/modules that are not directly
relevant to their business model, but used in some of their
products. By making these components open-source, they (hopefully) get
others to share in the burden of further development and maintenance.
--
Raymond Wiker
·············@fast.no
In article <·····························@news.ne.mediaone.net>,
Raffael Cavallaro <·······@mediaone.net> wrote:
>In other words, Open Source is about making the common tool we all use
>free, collaborating on it, all for mutual benefit. But if you are the
>person who only (or mostly) does high quality tools, how do you get paid
>in an Open Source world. There is no O'Reilly of Common Lisp to keep
>Kent on retainer while he cranks out a great Open Source Common Lisp,
>unlike Perl and Larry Wall.
Larry Wall wrote Perl long before he was employed by O'Reilly.
--
nosig
·····@comlab.ox.ac.uk (Philip Armstrong) wrote in message news:<············@news.ox.ac.uk>...
> In article <·····························@news.ne.mediaone.net>,
> Raffael Cavallaro <·······@mediaone.net> wrote:
> >In other words, Open Source is about making the common tool we all use
> >free, collaborating on it, all for mutual benefit. But if you are the
> >person who only (or mostly) does high quality tools, how do you get paid
> >in an Open Source world. There is no O'Reilly of Common Lisp to keep
> >Kent on retainer while he cranks out a great Open Source Common Lisp,
> >unlike Perl and Larry Wall.
>
> Larry Wall wrote Perl long before he was employed by O'Reilly.
And there's no good reason why O'Reilly (or ActiveState, or Micro$oft) can't pick up Lisp too.
g
In article <···························@posting.google.com>,
··········@my-deja.com (glauber) wrote:
>And there's no good reason why O'Reilly (or ActiveState, or Micro$oft)
>can't pick up Lisp too.
O'Reilly is not a charitable institution. They support Perl because they
know they can sell Perl books, which is how they make money.
I'd bet that O'Reilly have decided that there isn't sufficient demand
for lisp books that it would defray the cost of paying full time
developers.
To the extent that Active State makes money, they make it by charging
for commercial licences to use their IDE. How is this any different
from, say MCL, where, if you own a licence, you get essentially all the
source code for the compiler and IDE?
In other words, Open Source is not free, as in free beer. It is freedom
to modify the source, if you want to. But that source itself costs money
to develop, and that means it costs you, the IDE/compiler customer money
for a licence. This is how quality work gets done in all fields - people
get *paid* to do it.
Microsoft don't provide free compilers last time I checked, because
Gates was one of the pioneers in recognizing that people who code full
time, can't give the code away free of charge.
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
> In other words, Open Source is not free, as in free beer. It is freedom
> to modify the source, if you want to. But that source itself costs money
> to develop, and that means it costs you, the IDE/compiler customer money
> for a licence. This is how quality work gets done in all fields - people
> get *paid* to do it.
>
> Microsoft don't provide free compilers last time I checked, because
> Gates was one of the pioneers in recognizing that people who code full
> time, can't give the code away free of charge.
But thats not the whole story either. We've been working with a
company that maintains a realtime OS which the US Army paid them to
develop 10 or so years ago. It was done under the condition that the
source remain "open"- free for others to use and modify under
something like a bsd license. OAR is still maintaining the OS, they
make money from it by selling support and by people paying them to
enhance the OS or create new board support packages. There are no
millionaires there, but people earn decent livings; yet anyone can
download and use RTEMS without paying anyone anything.
http://www.oarcorp.com
Gregm
Greg Menke wrote:
>
> > In other words, Open Source is not free, as in free beer. It is freedom
> > to modify the source, if you want to. But that source itself costs money
> > to develop, and that means it costs you, the IDE/compiler customer money
> > for a licence. This is how quality work gets done in all fields - people
> > get *paid* to do it.
> >
> > Microsoft don't provide free compilers last time I checked, because
> > Gates was one of the pioneers in recognizing that people who code full
> > time, can't give the code away free of charge.
>
> But thats not the whole story either. We've been working with a
> company that maintains a realtime OS which the US Army paid them to
> develop 10 or so years ago. It was done under the condition that the
> source remain "open"- free for others to use and modify under
> something like a bsd license. OAR is still maintaining the OS, they
> make money from it by selling support and by people paying them to
> enhance the OS or create new board support packages. There are no
> millionaires there, but people earn decent livings; yet anyone can
> download and use RTEMS without paying anyone anything.
>
> http://www.oarcorp.com
Ada Core Technologies (http://www.gnat.com) does the same thing with the
GNAT Ada 95 compiler, developed in a similar manner. They seem to be
making a pretty good living from "free" software.
Larry
In article <··············@europa.mindspring.com>, Greg Menke
<··········@mindspring.com> wrote:
>they
>make money from it by selling support and by people paying them to
>enhance the OS or create new board support packages.
See my original post. This is one of the only ways people have managed
to make money in the Open Source world - getting people to pay you for
something other than the code itself - such as service contracts.
But if you write tools (like lisp compilers/IDEs) then you have sell
licences to use that compiler/IDE, otherwise, you've just set up shop as
a service supplier, which is *not* what you wanted to do (remember, you
want to write good lisp compilers/IDEs).
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
"Raffael Cavallaro" <·······@mediaone.net> wrote in message
··································@news.ne.mediaone.net...
> [...] This is one of the only ways people have managed
> to make money in the Open Source world - getting people to pay you for
> something other than the code itself - such as service contracts.
Again, Trolltech is a good exception.
They sell good code (Qt) for a good price. And they _also_ give it away
under a different license.
The user is not actually paying for the code; the code is free. The user
pays Trolltech a fair price for the right to use this code for money-making
purposes.
The developers of Qt can afford to eat. They create revenue by selling their
product (not just support contracts). Their business model also benefits the
community by allowing people use their product to make free stuff at no
cost.
In article <·····················@ozemail.com.au>, "John Flynn"
<··········@yahoo.com.au> wrote:
>The user is not actually paying for the code; the code is free. The user
>pays Trolltech a fair price for the right to use this code for money-making
>purposes.
But you miss some more subtle points. If I, as a consumer, buy a product
which includes the Qt libraries, I may get that code, but I can't
redistribute it in my own commercial product without buying a license
from Trolltech. I can only redistribute non-commercially. Note that this
is *not* a restriction of the BSD license, nor even of the GPL. So in
what sense are the Qt libraries free? They're really no different from
commercial libraries that developers pay for.
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
Raffael Cavallaro wrote:
> In article <·····················@ozemail.com.au>, "John Flynn"
> <··········@yahoo.com.au> wrote:
>
>>The user is not actually paying for the code; the code is free. The user
>>pays Trolltech a fair price for the right to use this code for
>>money-making purposes.
>
> But you miss some more subtle points. If I, as a consumer, buy a product
> which includes the Qt libraries, I may get that code, but I can't
> redistribute it in my own commercial product without buying a license
> from Trolltech. I can only redistribute non-commercially. Note that this
> is *not* a restriction of the BSD license, nor even of the GPL. So in
> what sense are the Qt libraries free? They're really no different from
> commercial libraries that developers pay for.
Wrong!
You _cannot_ distribute commercial software that uses GPL'ed Libraries. For
this kind of thing you need the LGPL!
QT Free Edition is a fully conformant _GPL_ licensed Library!!!!
This means you can use it with again GPL'ed software.
If you want to have another license you can buy the commercial license of
QT.
This is IMHO absolutely ok! - Do Trolltech have to allow you to use their
product for what you want without paying anything?
IMHO this licensing model is one of the best available - it is possible to
use the product for free if the result will be free too - and if you want
to do commercial work you can buy a license.
Regards,
Jochen
Jochen Schmidt <···@dataheaven.de> writes:
> Raffael Cavallaro wrote:
>
> > In article <·····················@ozemail.com.au>, "John Flynn"
> > <··········@yahoo.com.au> wrote:
> >
> >>The user is not actually paying for the code; the code is free. The user
> >>pays Trolltech a fair price for the right to use this code for
> >>money-making purposes.
> >
> > But you miss some more subtle points. If I, as a consumer, buy a product
> > which includes the Qt libraries, I may get that code, but I can't
> > redistribute it in my own commercial product without buying a license
> > from Trolltech. I can only redistribute non-commercially. Note that this
> > is *not* a restriction of the BSD license, nor even of the GPL. So in
> > what sense are the Qt libraries free? They're really no different from
> > commercial libraries that developers pay for.
>
> Wrong!
> You _cannot_ distribute commercial software that uses GPL'ed Libraries. For
> this kind of thing you need the LGPL!
You can distribute commercial software that uses GPLed libraries,
the license is quite explicit at allowing commercial sales of the
program. However, if you insist on your code being proprietary/hidden,
you cannot use GPL except inhouse. You can charge how much money you
want for your GPLed program, but you must include source if the
customer wants it.
--
------------------------------------------------------------------
Stig Erik Sandoe ····@ii.uib.no http://www.ii.uib.no/~stig/
In article <··············@palomba.bananos.org>, ····@ii.uib.no (Stig
E. Sandoe) wrote:
>You can distribute commercial software that uses GPLed libraries,
>the license is quite explicit at allowing commercial sales of the
>program. However, if you insist on your code being proprietary/hidden,
>you cannot use GPL except inhouse. You can charge how much money you
>want for your GPLed program, but you must include source if the
>customer wants it.
Your program now falls under the GPL (since the GPLed library and your
program are now one work, not separate works)
Here's the relevant quote from the recently released GPL FAQ:
>If the modules are included in the same executable file, they are
>definitely combined in one program. If modules are designed to run linked
>together in a shared address space, that almost surely means combining
>them into one program.
So now your whole product is GPLed. Anyone who buys it can give copies
away to others free. The GPL gives users the right to the source, *and*
the right to give away copies to others.
This right to distribute commercial products that include GPLed
libraries is a meaningless right, because you really only get one sale.
After that, your first customer is perfectly free to redistribute your
commercial product for free, so you make no money.
It isn't really a commercial distribution right if there's a clause that
says customers can copy your product and give it away to others. That's
the BS part of the GPL. RMS pretends that the GPL is commercially
compatible, but the copying rights make it effectively impossible to
make money for a GPLed product through software sales.
If you want to make money with GPLed software, you need to program
in-house, or do custom contract work. But mass market software can never
make money under the GPL. That's why RedHat and IBM have a support
contract revenue scheme.
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
Raffael Cavallaro <·······@mediaone.net> writes:
> In article <··············@palomba.bananos.org>, ····@ii.uib.no (Stig
> E. Sandoe) wrote:
>
> >You can distribute commercial software that uses GPLed libraries,
> >the license is quite explicit at allowing commercial sales of the
> >program. However, if you insist on your code being proprietary/hidden,
> >you cannot use GPL except inhouse. You can charge how much money you
> >want for your GPLed program, but you must include source if the
> >customer wants it.
>
> Your program now falls under the GPL (since the GPLed library and your
> program are now one work, not separate works)
Yes.
> Here's the relevant quote from the recently released GPL FAQ:
>
> >If the modules are included in the same executable file, they are
> >definitely combined in one program. If modules are designed to run linked
> >together in a shared address space, that almost surely means combining
> >them into one program.
>
> So now your whole product is GPLed. Anyone who buys it can give copies
> away to others free. The GPL gives users the right to the source, *and*
> the right to give away copies to others.
Yes.
> This right to distribute commercial products that include GPLed
> libraries is a meaningless right, because you really only get one sale.
> After that, your first customer is perfectly free to redistribute your
> commercial product for free, so you make no money.
For shrink-wrapped products your sales may be low, but again this
doesn't exclude commercial sales which you said the GPL did. The GPL
separates clearly between commercial and proprietary, and commercial
exploitation with the non-proprietary constraint is allowed. If the
author of some code strapped on a GPL it was his or her choice and it's
your choice whether you'd use his or her code or some other code. Chances
are that he or she[*] might be willing to discuss another license with
you, but it's his or her choice. Using the GPL, you allow new
programmers to stand on the shoulders of old programmers, instead of
on the toes as has been the situation in the industry for a long
time.
[*] Assuming a moderate person and not extremists worshipping
St. iGNUcius.
> It isn't really a commercial distribution right if there's a clause that
> says customers can copy your product and give it away to others.
Sure it is, but it isn't the usual way things are done now.
> If you want to make money with GPLed software, you need to program
> in-house, or do custom contract work. But mass market software can never
> make money under the GPL.
And in-house programming and custom contract-work isn't commercial?
--
------------------------------------------------------------------
Stig Erik Sandoe ····@ii.uib.no http://www.ii.uib.no/~stig/
On Fri, 25 May 2001, Raffael Cavallaro wrote:
> So now your whole product is GPLed. Anyone who buys it can give copies
> away to others free. The GPL gives users the right to the source, *and*
> the right to give away copies to others.
>
> This right to distribute commercial products that include GPLed
> libraries is a meaningless right, because you really only get one sale.
> After that, your first customer is perfectly free to redistribute your
> commercial product for free, so you make no money.
I think you're overestimating the desire of someone who paid for something
to turn around and give it away. For example, if you paid me $100k for
some custom modifications to gcc for your propriatary embedded product,
which presumably give you some advantage, would you put the mods on
your ftp server and share your advantage? I might at some point put the
modifications on my ftp server to ease my support burden, improve gcc in
general, give back to the community, etc. but that's a different matter.
Maybe there's some lower price point at which you'd put the code out
there, but I'm not sure why'd you bother. Maybe you'd put it out there
because you're pissed at me; this is a good part of what drives the whole
warez scene, no doubt. But the real motivation for distributing warez is
that it's forbidden. If it's not forbidden, that takes all the fun out of
it.
> It isn't really a commercial distribution right if there's a clause that
> says customers can copy your product and give it away to others. That's
> the BS part of the GPL. RMS pretends that the GPL is commercially
> compatible, but the copying rights make it effectively impossible to
> make money for a GPLed product through software sales.
Yet that's a large part of Cygnus' business model. Yes, they do ports for
hire which are made free, but that's the chip manufacturer's intent; get
some tools out there so people will buy their stuff. However, they also
sell GNU Pro, the whole gcc/binutils/gdb suite with some fancy window
dressing. GNU Pro obviously falls under the GPL. I once looked around
for a free download of it and failed. Now, that doesn't prove all that
much; I didn't troll in IRC warez channels or anything just made some
search engine queries, but free copies of GNU Pro are hardly ubiquitous.
Indeed, I've been told by Cygnus people that customers don't in general
turn around and release the source of their products.
Tim
In article <············@216.39.145.192>, Tim Moore wrote:
> On Fri, 25 May 2001, Raffael Cavallaro wrote:
>> So now your whole product is GPLed. Anyone who buys it can give copies
>> away to others free. The GPL gives users the right to the source, *and*
>> the right to give away copies to others.
>>
>> This right to distribute commercial products that include GPLed
>> libraries is a meaningless right, because you really only get one sale.
>> After that, your first customer is perfectly free to redistribute your
>> commercial product for free, so you make no money.
>
> I think you're overestimating the desire of someone who paid for something
> to turn around and give it away. For example, if you paid me $100k for
> some custom modifications to gcc for your propriatary embedded product,
> which presumably give you some advantage, would you put the mods on
> your ftp server and share your advantage? I might at some point put the
> modifications on my ftp server to ease my support burden, improve gcc in
> general, give back to the community, etc. but that's a different matter.
What if I write some custom gcc for embeded cpu x27003 and I want to
sell it for $500.00 per developer and it cost me $500,000.00 in time
and money to produce/market. Now I sell 20 copies befor someone puts
it on sourceforge for download and sends it to GNU for inclusion into
gcc. What happens to my investment? How do I recover $490,000.00?
marc
>
> Maybe there's some lower price point at which you'd put the code out
> there, but I'm not sure why'd you bother. Maybe you'd put it out there
> because you're pissed at me; this is a good part of what drives the whole
> warez scene, no doubt. But the real motivation for distributing warez is
> that it's forbidden. If it's not forbidden, that takes all the fun out of
> it.
>
>> It isn't really a commercial distribution right if there's a clause that
>> says customers can copy your product and give it away to others. That's
>> the BS part of the GPL. RMS pretends that the GPL is commercially
>> compatible, but the copying rights make it effectively impossible to
>> make money for a GPLed product through software sales.
>
> Yet that's a large part of Cygnus' business model. Yes, they do ports for
> hire which are made free, but that's the chip manufacturer's intent; get
> some tools out there so people will buy their stuff. However, they also
> sell GNU Pro, the whole gcc/binutils/gdb suite with some fancy window
> dressing. GNU Pro obviously falls under the GPL. I once looked around
> for a free download of it and failed. Now, that doesn't prove all that
> much; I didn't troll in IRC warez channels or anything just made some
> search engine queries, but free copies of GNU Pro are hardly ubiquitous.
> Indeed, I've been told by Cygnus people that customers don't in general
> turn around and release the source of their products.
>
> Tim
>
>
On Fri, 25 May 2001, Marc Spitzer wrote:
> In article <············@216.39.145.192>, Tim Moore wrote:
> > I think you're overestimating the desire of someone who paid for something
> > to turn around and give it away. For example, if you paid me $100k for
> > some custom modifications to gcc for your propriatary embedded product,
> > which presumably give you some advantage, would you put the mods on
> > your ftp server and share your advantage? I might at some point put the
> > modifications on my ftp server to ease my support burden, improve gcc in
> > general, give back to the community, etc. but that's a different matter.
>
> What if I write some custom gcc for embeded cpu x27003 and I want to
> sell it for $500.00 per developer and it cost me $500,000.00 in time
> and money to produce/market. Now I sell 20 copies befor someone puts
> it on sourceforge for download and sends it to GNU for inclusion into
> gcc. What happens to my investment? How do I recover $490,000.00?
>
> marc
I don't know why you're asking me since I've asserted that this does not
in fact happen with gcc unless the development funder wants it to happen.
Nevertheless, I can make several suggestions. First, if you funded
development for the x27003 cpu you are probably the manufacturer of the
cpu and intend to make your money from sales of the cpu, not the
development software, as you want to encourage development on your cpu.
Second, why would someone who paid $500 for your product do that? I could
see them sharing it with their friends, but this isn't Photoshop or Myst
we're talking about here. Third, if you sink (conservatively) 2 man years
of effort into a product, want to make it all back in product sales, and
don't feel that making the source widely available enhances the value of
the product, to you as well as others, perhaps an open source license is
not for you.
Tim
In article <············@216.39.145.192>, Tim Moore wrote:
> On Fri, 25 May 2001, Marc Spitzer wrote:
>
>> In article <············@216.39.145.192>, Tim Moore wrote:
>> > I think you're overestimating the desire of someone who paid for something
>> > to turn around and give it away. For example, if you paid me $100k for
>> > some custom modifications to gcc for your propriatary embedded product,
>> > which presumably give you some advantage, would you put the mods on
>> > your ftp server and share your advantage? I might at some point put the
>> > modifications on my ftp server to ease my support burden, improve gcc in
>> > general, give back to the community, etc. but that's a different matter.
>>
>> What if I write some custom gcc for embeded cpu x27003 and I want to
>> sell it for $500.00 per developer and it cost me $500,000.00 in time
>> and money to produce/market. Now I sell 20 copies befor someone puts
>> it on sourceforge for download and sends it to GNU for inclusion into
>> gcc. What happens to my investment? How do I recover $490,000.00?
>>
>> marc
>
> I don't know why you're asking me since I've asserted that this does not
> in fact happen with gcc unless the development funder wants it to happen.
> Nevertheless, I can make several suggestions. First, if you funded
> development for the x27003 cpu you are probably the manufacturer of the
Let me clarify, think borland not intel. This is a discusion about
tool vendors as the primary business
> cpu and intend to make your money from sales of the cpu, not the
> development software, as you want to encourage development on your cpu.
> Second, why would someone who paid $500 for your product do that? I could
10 colege kids kick in $50 each to buy 1 licence for one of them. The
licenced user gives it to the other 9 people. One of them figures out
that the compiler is based on gcc and is gpled so it is sent to gnu
for inclusion in a future release of gcc and posted in source form as
a sourceforge project.
> see them sharing it with their friends, but this isn't Photoshop or Myst
> we're talking about here. Third, if you sink (conservatively) 2 man years
> of effort into a product, want to make it all back in product sales, and
Do not forget marketing and sales, they cost money also. And I need
produce to ship(stuff in a box) thern there is tech support and the
bookkeeping needed to track the money. You need to consider the WHOLE
bussiness here not just the developers time. And I almost forgot all
the stuff you need, computers, desks, phones, an office... All these
things cost hard cash.
> don't feel that making the source widely available enhances the value of
> the product, to you as well as others, perhaps an open source license is
> not for you.
>
I am not going to get into the GPL is or is not opensource debate
here, I don't think it is. Tools are tools and when I have a tool I
want it to work I do not want to have to make it work. All the time I
spend futzing arond fixing the tools is time spent away from the work
I need to get done and am paid to do. Also how many people realy get
to know the code well enough to fix real bugs, the kind that require
understanding of the product in general and specific? And from my POV
how do I support a product that is not the product I sold you, you
modified the source and strange things happen and you say you did
nothing it just don't work fix it.
Let me ask you a question if you needed a tool that cost $500 to
purchase or you could go to sourceforge and download the exact same
tool for free, what would you do? What would most people do?
Once it get to sourceforge I would be selling sand in the desert and
that is a VERY TOUGH sale.
marc
> Tim
>
>
>>>>> "Marc" == Marc Spitzer <····@oscar.eng.cv.net> writes:
Marc> 10 colege kids kick in $50 each to buy 1 licence for one of
Marc> them. The licenced user gives it to the other 9 people. One of
Marc> them figures out that the compiler is based on gcc and is gpled
Marc> so it is sent to gnu for inclusion in a future release of gcc
Marc> and posted in source form as a sourceforge project.
You might want to look at GNAT (Ada Core Technologies) for an example
where commercial GPL'd software tends *not* to leak out into the
public.
--
Russell Senior ``The two chiefs turned to each other.
·······@aracnet.com Bellison uncorked a flood of horrible
profanity, which, translated meant, `This is
extremely unusual.' ''
In article <··············@coulee.tdb.com>, Russell Senior wrote:
>>>>>> "Marc" == Marc Spitzer <····@oscar.eng.cv.net> writes:
>
>Marc> 10 colege kids kick in $50 each to buy 1 licence for one of
>Marc> them. The licenced user gives it to the other 9 people. One of
>Marc> them figures out that the compiler is based on gcc and is gpled
>Marc> so it is sent to gnu for inclusion in a future release of gcc
>Marc> and posted in source form as a sourceforge project.
>
> You might want to look at GNAT (Ada Core Technologies) for an example
> where commercial GPL'd software tends *not* to leak out into the
> public.
>
I did look at there website and it looks like they sell support. Also
GNAT had no choice but to accept gpl, gnat was based on gcc so gpl
applys. It does look like a complete package though, I know nothing
about ADA, it has a lot of stuff. I downloaded the free version and
found most if not all of there product avalable in it, so it has
leaked out. What probably has not leaked out is the value added stuff
that they might have, I don't know the products at all, that is not
gpled.
marc
>
> --
> Russell Senior ``The two chiefs turned to each other.
> ·······@aracnet.com Bellison uncorked a flood of horrible
> profanity, which, translated meant, `This is
> extremely unusual.' ''
····@oscar.eng.cv.net (Marc Spitzer) writes:
> In article <··············@coulee.tdb.com>, Russell Senior wrote:
> > You might want to look at GNAT (Ada Core Technologies) for an example
> > where commercial GPL'd software tends *not* to leak out into the
> > public.
> I did look at there website and it looks like they sell support. Also
> GNAT had no choice but to accept gpl, gnat was based on gcc so gpl
> applys. It does look like a complete package though, I know nothing
> about ADA, it has a lot of stuff. I downloaded the free version and
> found most if not all of there product avalable in it, so it has
> leaked out. What probably has not leaked out is the value added stuff
> that they might have, I don't know the products at all, that is not
> gpled.
GNAT has not leaked out. It was intentionally public right from the beginning,
and in fact was originally funded by a government contract that mandated
exactly that.
What Ada Core Technologies is doing is maintaining two versions of GNAT: a
cutting edge version for officially supported customers, and a freely
available version. The free version is not dumbed down in any way -- it just
lags somewhat behind (about a year or so, I believe).
ACT has consistently maintained its commitment to a freely available Ada
compiler, for they believe encouraging the use of Ada helps their business.
--
Cheers, The Rhythm is around me,
The Rhythm has control.
Ray Blaak The Rhythm is inside me,
·····@infomatch.com The Rhythm has my soul.
Ray Blaak <·····@infomatch.com> writes:
> ····@oscar.eng.cv.net (Marc Spitzer) writes:
> > In article <··············@coulee.tdb.com>, Russell Senior wrote:
> > > You might want to look at GNAT (Ada Core Technologies) for an example
> > > where commercial GPL'd software tends *not* to leak out into the
> > > public.
>
> > I did look at there website and it looks like they sell support. Also
> > GNAT had no choice but to accept gpl, gnat was based on gcc so gpl
> > applys. It does look like a complete package though, I know nothing
> > about ADA, it has a lot of stuff. I downloaded the free version and
> > found most if not all of there product avalable in it, so it has
> > leaked out. What probably has not leaked out is the value added stuff
> > that they might have, I don't know the products at all, that is not
> > gpled.
>
> GNAT has not leaked out. It was intentionally public right from the beginning,
> and in fact was originally funded by a government contract that mandated
> exactly that.
As a matter of fact many goverment agencies are now mandating "Open
Source" development of software tools.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
In article <·············@infomatch.com>, Ray Blaak wrote:
> ····@oscar.eng.cv.net (Marc Spitzer) writes:
>> In article <··············@coulee.tdb.com>, Russell Senior wrote:
>> > You might want to look at GNAT (Ada Core Technologies) for an example
>> > where commercial GPL'd software tends *not* to leak out into the
>> > public.
>
>> I did look at there website and it looks like they sell support. Also
>> GNAT had no choice but to accept gpl, gnat was based on gcc so gpl
>> applys. It does look like a complete package though, I know nothing
>> about ADA, it has a lot of stuff. I downloaded the free version and
>> found most if not all of there product avalable in it, so it has
>> leaked out. What probably has not leaked out is the value added stuff
>> that they might have, I don't know the products at all, that is not
>> gpled.
>
> GNAT has not leaked out. It was intentionally public right from the beginning,
> and in fact was originally funded by a government contract that mandated
> exactly that.
You are correct, poor choice of words on my part.
>
> What Ada Core Technologies is doing is maintaining two versions of GNAT: a
> cutting edge version for officially supported customers, and a freely
> available version. The free version is not dumbed down in any way -- it just
> lags somewhat behind (about a year or so, I believe).
How do they effectivly do this with gpl? It is based on gcc so gpl
applys, right?
>
> ACT has consistently maintained its commitment to a freely available Ada
> compiler, for they believe encouraging the use of Ada helps their business.
>
I can see there point, we need ada programmers to sell ada compilers.
marc
> --
> Cheers, The Rhythm is around me,
> The Rhythm has control.
> Ray Blaak The Rhythm is inside me,
> ·····@infomatch.com The Rhythm has my soul.
····@oscar.eng.cv.net (Marc Spitzer) writes:
> > What Ada Core Technologies is doing is maintaining two versions of GNAT: a
> > cutting edge version for officially supported customers, and a freely
> > available version. The free version is not dumbed down in any way -- it just
> > lags somewhat behind (about a year or so, I believe).
>
> How do they effectivly do this with gpl? It is based on gcc so gpl
> applys, right?
gcc and the ada compiler code itself are GPL. They release it as LGPL, I
believe, both the supported and the public versions.
They suggest to supported clients to not release the supported version, so as
to ensure that people are always working with properly maintained versions of
things.
However, releasing a supported version is "allowed" of course, since it is
under LGPL. It is just that if you go to ACT with a question about it, and you
are not a supported customer, then they won't help you.
There is nothing to stop another person or company from taking the product and
distributing it again and setting up their own support service for hire.
However, ACT has a good reputation, so people find it more worthwhile to pay
them for a good service.
--
Cheers, The Rhythm is around me,
The Rhythm has control.
Ray Blaak The Rhythm is inside me,
·····@infomatch.com The Rhythm has my soul.
····@oscar.eng.cv.net (Marc Spitzer) writes:
> 10 colege kids kick in $50 each to buy 1 licence for one of them. The
> licenced user gives it to the other 9 people. One of them figures out
> that the compiler is based on gcc and is gpled so it is sent to gnu
> for inclusion in a future release of gcc
which the GCC Steering Committee wouldn't accept, since, in order to
be able to defend the GPL in court, they want a copyright assignment
for all non trivial source that goes into gcc.
--
Lieven Marchand <···@wyrd.be>
Making laws appears to win votes. Enforcing them doesn't.
See Rule One. Roger Burton West in the monastery.
In article <··············@localhost.localdomain>, Lieven Marchand wrote:
> ····@oscar.eng.cv.net (Marc Spitzer) writes:
>
>> 10 colege kids kick in $50 each to buy 1 licence for one of them. The
>> licenced user gives it to the other 9 people. One of them figures out
>> that the compiler is based on gcc and is gpled so it is sent to gnu
>> for inclusion in a future release of gcc
>
> which the GCC Steering Committee wouldn't accept, since, in order to
> be able to defend the GPL in court, they want a copyright assignment
> for all non trivial source that goes into gcc.
I stand corrected, but is there anything that prevents the posting of
the source code on the net?
marc
>
>
> --
> Lieven Marchand <···@wyrd.be>
> Making laws appears to win votes. Enforcing them doesn't.
> See Rule One. Roger Burton West in the monastery.
In article <····················@oscar.eng.cv.net>,
····@oscar.eng.cv.net (Marc Spitzer) wrote:
>I stand corrected, but is there anything that prevents the posting of
>the source code on the net?
No, not according to the GPL there isn't:
>1. You may copy and distribute verbatim copies of the Program's source
>code as you receive it, in any medium, provided that you conspicuously and
>appropriately publish on each copy an appropriate copyright notice and
>disclaimer of warranty; keep intact all the notices that refer to this
>License and to the absence of any warranty; and give any other recipients
>of the Program a copy of this License along with the Program.
That's the first clause of the GPL, right after the Preamble. So as long
as you keep copyright notices and warantee disclaimers intact, you can
publish it on the web, on the back of a box of cereal - anyway you like.
So your original problem with selling high priced GPL software are a
business model stands. You have no way of protecting your investment
because you never know which customer might give it away to the world
for free.
BTW, these same rights to give copies away also apply to binaries too
(as long as you also give away the source code, or make it possible for
those you give binaries to, to request source from you).
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
In article <·····························@news.ne.mediaone.net>,
Raffael Cavallaro <·······@mediaone.net> wrote:
>So your original problem with selling high priced GPL software are a
^^^
as
>business model stands.
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
In article <··············@ID-22205.news.dfncis.de>, Jochen Schmidt
<···@dataheaven.de> wrote:
>If you want to have another license you can buy the commercial license of
>QT.
>This is IMHO absolutely ok! - Do Trolltech have to allow you to use their
>product for what you want without paying anything?
You are arguing *my* point. I'm the one who's advocating that people who
write useful libraries *should* be paid for their code by means of
commercial licenses.
I'm the one who's saying that Open Source screws those who wan't to
write code, not run a service organization.
If you take the split licence route (one commercial, one GPL, like
TrollTech) you've really just joined the commercial camp. If I make a
product that uses Qt, and I want to distribute it commercially, I have
to pay. This is no different than other commercial libraries.
Since it's now become a pay-to-play situation for commercial use, all
the Open Source licencees are stuck in the GPL ghetto, where experience
has shown, you can't make money for writing code, because the market
forces the price of your code toward $0.00 (because, although the GPL
allows you to charge for your product, anyone who buys it also gets
code, which they can simply recompile and distribute for free, with
source - there goes your revenue stream. Essentially, with the GPL, you
only get one commercial customer, and everyone after the first gets your
product free). Thus, people who distribute GPLed software, must make
money for something other than writing code. But if writing code is what
you do best, what you want to do, the GPL screws you economically, by
drastically reducing your revenue options (I would argue, essentially
eliminating them).
The GPL is the darling of people who do custom/in-house work, because
the code sharing makes their jobs easier, and there's no real worry
about losing revenue because they're getting paid a fixed sum for the
product, no matter how many seats it's deployed to.
But people who live in a different market, where revenue comes not from
a flat fee contract, but from sales, have essentially no way to make
money from GPLed software. That's why we see the big GPL product
companies all pursuing a service contract revenue model - because the
software itself can *never* be a revenue stream for GPLed products.
Here's a snippet from an email exchange with Paul Foley:
>On 5/24/01 5:33 PM, Paul Foley, (·······@actrix.gen.nz), wrote:
>
>> The Qt libraries are GPL -- that means you _can_
>> redistribute and use them in your own commercial product, if that
>> product is also GPLed (I don't know whether you can sustain a
>> commercial GPLed product, but that's your problem).
>
>Yes, that's your problem. That's what Kent has been saying, and what the
>Open Source advocates have been ignoring. The fact of the matter is, you
>*can't* sustain a commercial GPLed product, because anyone who buys a
>licence can recompile the source, and distribute it *free of charge*.
>
>Why is this so difficult to see? If the market forces the price of your
>software toward $0.00, then you need to make money some other way. But
>now, you're not really in the software business anymore, you're in the
>support business, which is *not* your area of expertise. Open Source is an
>economic screwing of those whose preference and expertise is writing code
>to be sold in the open market.
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
Raffael Cavallaro <·······@mediaone.net> writes:
> Since it's now become a pay-to-play situation for commercial use,
> all the Open Source licencees are stuck in the GPL ghetto, where
> experience has shown, you can't make money for writing code, because
> the market forces the price of your code toward $0.00 (because,
> although the GPL allows you to charge for your product, anyone who
> buys it also gets code, which they can simply recompile and
> distribute for free, with source - there goes your revenue
> stream. Essentially, with the GPL, you only get one commercial
> customer, and everyone after the first gets your product
> free). Thus, people who distribute GPLed software, must make money
> for something other than writing code. But if writing code is what
> you do best, what you want to do, the GPL screws you economically,
> by drastically reducing your revenue options (I would argue,
> essentially eliminating them).
If we look at RH's annual report (the only one I could find was for
2000) we can see that they were making 56% percent of their revenue on
subscriptions, that's people paying for GPLed code that is available
elsewhere, 23.9 million dollars from GPLed code. Then we can see that
they made in the same period about 17.2 million dollars from custom
development, training and consulting.
This would seem to indicate that your assertion that you have only one
corporate customer with the GPL, and then they distribute the product
for free is patently false. Cygnus' history would also seem to
indicate that assertion is false.
With a 23 million dollar revenue stream from software whose price has
been forced to $0, RH must be working some magic. How many units do
you have to sell at $0 to bring in 23 million?
Then there is Alladin, who do things a bit differently. They release
their commercial product, postscript driver, under the GPL after a
delay. I don't have any financial information for them tho, but they
have been able to drive the development of ghostscript effectively.
It seems there is a disconnect between your theory of software pricing
and reality.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
In article <··············@piracy.red-bean.com>, Craig Brozefsky wrote:
> Raffael Cavallaro <·······@mediaone.net> writes:
>
>> Since it's now become a pay-to-play situation for commercial use,
>> all the Open Source licencees are stuck in the GPL ghetto, where
>> experience has shown, you can't make money for writing code, because
>> the market forces the price of your code toward $0.00 (because,
>> although the GPL allows you to charge for your product, anyone who
>> buys it also gets code, which they can simply recompile and
>> distribute for free, with source - there goes your revenue
>> stream. Essentially, with the GPL, you only get one commercial
>> customer, and everyone after the first gets your product
>> free). Thus, people who distribute GPLed software, must make money
>> for something other than writing code. But if writing code is what
>> you do best, what you want to do, the GPL screws you economically,
>> by drastically reducing your revenue options (I would argue,
>> essentially eliminating them).
>
> If we look at RH's annual report (the only one I could find was for
> 2000) we can see that they were making 56% percent of their revenue on
> subscriptions, that's people paying for GPLed code that is available
> elsewhere, 23.9 million dollars from GPLed code. Then we can see that
> they made in the same period about 17.2 million dollars from custom
> development, training and consulting.
>
> This would seem to indicate that your assertion that you have only one
> corporate customer with the GPL, and then they distribute the product
> for free is patently false. Cygnus' history would also seem to
> indicate that assertion is false.
If I remember correctly the assertion was that after he sells 1 copy
of the product touched by gpl then he would loose all rights to the
code that he wrote that was infected(or embraced) by gpled code he
used. The problem I have with your redhat example is that there are
LOTS of people with slow links so it makes sence not to download
several ISO images at 600 MB each. For a 10 MB tool/lib the equation
is different. Cygnus made its money selling support contracts, if you
need to call for help at 3am well be there, for a fee of course. All
of the money redhat got was from service It was stated(if I remember
correctly) that this was a bussiness model that he did not want to
use, he wanted to make tools for a living not support tools for a
living.
>
> With a 23 million dollar revenue stream from software whose price has
> been forced to $0, RH must be working some magic. How many units do
> you have to sell at $0 to bring in 23 million?
>
Redhat sells CD's, tech support and consulting. If I remember
correctly there is stuff on the CD's that is not avaliable for
download at there public ftp site. This gives you another reason to
buy the cd. One of the reasons that they can do the CD's so cheap is
that most of the software they distribute they do not pay for so they
are making money off the work of many other people and not
compensating them for there work.
> Then there is Alladin, who do things a bit differently. They release
> their commercial product, postscript driver, under the GPL after a
> delay. I don't have any financial information for them tho, but they
> have been able to drive the development of ghostscript effectively.
>
> It seems there is a disconnect between your theory of software pricing
> and reality.
If I write something I should be able to sell it for what ever price I
want. And you have the choice to buy it or not. But please don't
complain about my pricing. A closed wallet is the most effective
method of getting the point across.
marc
>
> --
> Craig Brozefsky <·····@red-bean.com>
> http://www.red-bean.com/~craig
> "Indifference is the dead weight of history." -- Antonio Gramsci
····@oscar.eng.cv.net (Marc Spitzer) writes:
> It was stated(if I remember correctly) that this was a bussiness
> model that he did not want to use, he wanted to make tools for a
> living not support tools for a living.
Not sure who the "he" is here, but certainly this is true of me.
Btw, the kinds of tools I want to write might have the occasional bug
but mostly aren't going to have a lot of bugs. I plan to look over
and test the code reasonably well, and I just can't see hoards of
people paying me a lot to "support" them.
I don't want to get too far into unfair generalizations or casting
aspersions on anyone, but as incentive-based systems go, a system
where you make money on support doesn't encourage QA or good
documentation. Just the opposite, it encourages quick time to market
and donig all those last-minute details after-the-fact since no one
is paying anyway, and since perhaps someone WILL pay you to finish.
Sure, some people won't fall victim to this. But I still believe
that biases like this will occur.
Certainly I don't consider redhat to be easy to use, well-documented,
nor well QA'd. The commonly used stuff works ok, but there is rampant
variability in network card support, for example, since it would be
nearly impossible for them to QA all possible configurations and they
don't go out of their way to push one configuration as well-QA'd.
There's scads of documentation all over the place, but it's not all in
the printed doc--some is in HOWTO files and for many things you have
to get Dummy or O'Reilly books. It's hard to know when something is
missing because it's simply elsewhere or because it's nowhere. This
is no model of anything I'd ever want to produce, and fails in ways that
are at least accidentally aligned very closely with the areas that I
suggested in the prior paragraph would (according to theory) fall short.
Perhaps it's just coincidence, but if revenues drop on support as the
system becomes easier to use, it'll be interesting to see what happens
to the "business model".
> >
> > With a 23 million dollar revenue stream from software whose price has
> > been forced to $0, RH must be working some magic. How many units do
> > you have to sell at $0 to bring in 23 million?
>
> Redhat sells CD's, tech support and consulting.
They also market pretty heavily, too. Enough that I bet a lot of people
buy from them in response to the marketing rather than perceived value.
If they did not do heavy marketing, I bet others who repackage their wares
would be thought to be the original producer (just as redhat is thought
to be the original producer of many gnu-produced tools, since gnu takes
out no ads to protect its image). If RedHat didn't do commercial advertising
and/or non-paid promotional pieces keeping their name visible in the
mainstream media, I bet they'd be closer to gnu in terms of the ratio of
value produced to value acknowledged, and with perhaps correspondingly less
revenues.
It'd be ironic, but not inconceivable, to find that advertising made
the difference, and that the only "free software" that could make it
in business was that which had the advertising backing to do global
marketing campaigns in mainstream media. Maybe we should be pushing
for "free advertising" to go along with "free software". Oops, that's
right, advertising is all that stands between many non-charging services
and bankruptcy... Hmmm.
Kent M Pitman <······@world.std.com> writes:
> It'd be ironic, but not inconceivable, to find that advertising made
> the difference, and that the only "free software" that could make it
> in business was that which had the advertising backing to do global
> marketing campaigns in mainstream media. Maybe we should be pushing
> for "free advertising" to go along with "free software". Oops, that's
> right, advertising is all that stands between many non-charging services
> and bankruptcy... Hmmm.
In case you haven't noticed, most business these days is generated by
advertising. Do you think pepple buy products based on usefulness?
The largest selling products are always the ones that advertise the
most effectively.
But back to topic. GPL proponents bring up companies like RedHat to
demonstrate that people can make money from GPL software. But I can't
believe that many companies can succeed with that business model.
I think there are many other reasons companies would want to finance
GPL software.
If my product is (for an example) a customer database system, when I
sell a system my customers end up having to pay for an OS, OS tools,
backup software, and indirectly the compilers I use to develop the
software. Thats a lots of costs. I may avoid GPL software in my
product, but if I use GPL components for these other things, I may
save a lot of money. It might be worth subsidizing GPL development
groups that keep theses other pieces current.
Unix OS companies like Sun have found supporting the non-kernel parts
of their OS's a real pain. This is not where they get their profits.
They really are a sink, because the require support costs, but nobody
is going to get a Sun computer because it has a great "ls" command.
This is why the Unix community got together and farmed this out to a
"System V" consortium. Sun recently announced that they are going to
finance and use the GPL window system called "Gnome". This is not
because Sun beleves in open software. Its just a way that they can
support a standard window system without having to foot all the bill.
They found that after Sunview, NeWs, and OpenLook, writing their own
latest generation window environment was a loss.
For companies, the GPL supplies a flat playing field, since everyone
has the source, and a large number of people can usually be found to
work with it. You loose any compensative advantage, where you know
your not going to be a leader. You focus your propriety GPL-free code
on where you can find an advantage for yourself.
People complain that this results in fewer programing jobs. At least
the remaining jobs would be more interesting.
Free-Traders, who have control in the US govenment, argue the jobs
lost to overseas manufacturing doen't harm, but helps the US economy.
The same argument should work with the GPL. (I'm not convinced about
free-trade) But, unlike in manufacturing, there is a shortage of good
programmers.
Barry Fishman
Barry Fishman <·············@acm.org> writes:
> If my product is (for an example) a customer database system, when I
> sell a system my customers end up having to pay for an OS, OS tools,
> backup software, and indirectly the compilers I use to develop the
> software. Thats a lots of costs. I may avoid GPL software in my
> product, but if I use GPL components for these other things, I may
> save a lot of money. It might be worth subsidizing GPL development
> groups that keep theses other pieces current.
I don't see how one can avoid having an internal support and QA group
unless one really loves living on the edge. If you've not purchased
the GPL'd component directly, you have no contract. You have only a
gift. A gift has no consideration on both sides, and so you can
really rely on nothing. If you pay into support, you are gifting in
the other direction. The two do not combine (as far as I know) to
make a contract; if they did, it would probably be in violation of
GPL.
My point is that if you are giving away a gift, it's possible to shirk
issues of quality. Certainly, I find the longstanding parts of Linux
to work well, but the more recently added stuff to be middlingly QA'd.
Especially network drivers are bad, missing, and confusing. While I
could imagine making tools that someone else bundled into a layered package,
I could never personally imagine building a shrink-wrapped end-user
layered product that used as its underlying platform something I received
as a gift with no warranty of use of any kind.
You could contract out QA to the people, but that would be a
conflict of interest--they're already giving you the software; they're
not the right people to pay for additional QA.
You could wait until the thing breaks and then rely on a suppport
contract to make after-the-fact repairs. But pro-active work in
support of quality is not "support", and you're essentially proposing
I either not pay for that or that I pay for it separately, partly breaking
down your argument that I'll save.
Sure, I'll save on not doing the development, but the development may not
even be the key soft in ongoing maintenance. Doing the development would
at least assure I had someone on staff who understood the software. Letting
someone else do it means I have to go learn someone else's software, which
in my experience is often much harder than doing software anew. (Your mileage
may vary.)
Plus if the people supporting it are other than your organization and
don't have your same sense of priorities about how fast they need to
fix something, or about how cooperation needs to happen with other
parts of your product not covered by the support, you're caught in a
bind with your customer blaming you but with you unable to have first-hand
control on repair.
Sure, you have the sources, and you might make a fix. But that splits the
sources. You can contribute your fix back, but in my experience that does
not result 100% of the time in the person maintaining the free code actually
inserting your fix. Sometimes they discard it. Sometimes they do something
different and incompatible with what you've done. This creates a burden
(read: "cost") on you to merge sources when you receive new, "free" updates.
All in all, I would never personally advise my company to go this
route. It's fraught with hidden costs, the chief one being a very
doubtful claim that if you somehow relinquish responsibility for
manufacturing the components you need to a specified level of quality,
that they will somehow show up ready for use at the appropriate level
of quality on the appropriate schedule without your needing to
pressure anyone and without your having the leverage to pressure them
even if you wanted to.
There may be companies that can be shown to have done this but I don't think
there is long-term evidence that such companies survive, nor that their
survival has been due to this model, nor even tolerating this model.
> For companies, the GPL supplies a flat playing field, since everyone
> has the source, and a large number of people can usually be found to
> work with it. You loose any compensative advantage, where you know
> your not going to be a leader. You focus your propriety GPL-free code
> on where you can find an advantage for yourself.
You can until the first time the GPL'd code embarrasses you. At that
point, in my experience, the customer doesn't really care that it's
publicly maintained and that the fault is with code you didn't make.
At that point, it's YOUR fault that you went with such a solution and
the burden is on YOU, not the people you chose to buy from, to make
good.
Waiting on free software is a gamble. You might consider it a good
gamble, but it still has to be stated as a gamble because it does not
include safeguards of the kinds that are present in other business
models. In particular, commercial products are often second-sourced
and if someone doesn't deliver a product that works, you often just
switch to a backup source. (Free software seems to assume that a
single, canonical source is best.)
> People complain that this results in fewer programing jobs. At least
> the remaining jobs would be more interesting.
I doubt you can support this claim. Interestingness is surely a subjective
metric.
> Free-Traders, who have control in the US govenment, argue the jobs
> lost to overseas manufacturing doen't harm, but helps the US economy.
> The same argument should work with the GPL. (I'm not convinced about
> free-trade) But, unlike in manufacturing, there is a shortage of good
> programmers.
I don't care a hoot about the shortage of programmers.
What I care about is the direction of computer science. I, and I doubt
I'm unique, have theories of where I'd like things to go. But as long as
free software shows up going one way, it makes it even harder than it
already would be to convince people to go the other way for money...
Kent M Pitman <······@world.std.com> writes:
> I could never personally imagine building a shrink-wrapped end-user
> layered product that used as its underlying platform something I received
> as a gift with no warranty of use of any kind.
But then, commercial software hasn't warranty of use
of any kind
> You could contract out QA to the people, but that would be a
> conflict of interest--they're already giving you the software; they're
> not the right people to pay for additional QA.
Why? Just give a fraction of the cost of a
similar commercial software to this people to
"assure quality". That they are "gpl coders"
doesn't mean that they can't sign a contract.
> Sure, you have the sources, and you might make a fix. But that splits the
> sources. You can contribute your fix back, but in my experience that does
> not result 100% of the time in the person maintaining the free code actually
> inserting your fix.
Create your own branch (remember Xemacs?). It's
not sure that you save any money doing this, but
maybe you do.
> In particular, commercial products are often second-sourced
> and if someone doesn't deliver a product that works, you often just
> switch to a backup source. (Free software seems to assume that a
> single, canonical source is best.)
Sure? So what is this?
http://www.kernel.org/pub/linux/kernel/
I'll tell you. Linux kernels going back to 1.0
> What I care about is the direction of computer
> science.
So do I. And I think that computers science
"general knowledge" should be freely available
(just like with math, phisics or enginnering).
I mean: it's ok that new tecnologies, methods or
programs are closed-source but general tool like
compilers, interpreters, database engines,
newsreaders and even socket libraries should have
at least one GPL representant to be used and
learnt from.
Think this a moment:
Imagine that engineering could be
closed-source. So that you cannot see how any
device works and cannot leverage that knowlege.
Think about the light bulb, transistors, engines,
bridges, buildings, trains, cars ...
Where could we be? Stone age? Maybe not far from
there.
Excuse my bad english.
--
Eduardo Mu�oz
"Eduardo Mu�oz" <···@jet.es> writes:
> Kent M Pitman <······@world.std.com> writes:
> > Sure, you have the sources, and you might make a fix. But that splits the
> > sources. You can contribute your fix back, but in my experience that does
> > not result 100% of the time in the person maintaining the free code actually
> > inserting your fix.
>
> Create your own branch (remember Xemacs?). It's
> not sure that you save any money doing this, but
> maybe you do.
I don't think getting maintainers to insert your fix is as difficult
as you do. Especially if you are helping to fund them.
The problems usually occur around ideology. XEmacs is more of a
political split than functionality. If you just want to fix problems
with the software and you don't want to make a big deal over the
licencing, I don't see why you would get objections. If you have
major problems with the software, why are you using it in the first
place?
If a maintainer puts ideology above working code they won't last very
long. Nobody appointed Linus head of a GNU OS development group.
Maintainers hold their positions only when people like the results.
As far as support, you can't avoid risks. Does MicroSoft give your
problems the highest priority? How much influence does your money
have on them, when their interest is elsewhere. And what happens when
your sofware supplier goes under?
Ultimately major GPL software efforts will live or die based on how
they fit our needs. I don't consider the GPL a magical incantation
which can destroy or save our society.
Barry Fishman
Barry Fishman <·············@acm.org> writes:
> I don't think getting maintainers to insert your fix is as difficult
> as you do. Especially if you are helping to fund them. [...]
> If a maintainer puts ideology above working code they won't last very
> long.
This point, exactly as you've expressed it, is often the source of problem.
You speak as if you mean to say:
1. If a maintainer puts ideology above working code, they won't last very
long.
2. THEREFORE, the maintainer will not put ideology above working code.
3. THEREFORE, I will be happy.
Q.E.D. (Quit Everything's Done)
But there is another scenario consistent with your statement.
1. If a maintainer puts ideology above working code, they won't last very
long.
2. Maintainer puts ideology above working code.
3. THEREFORE, they won't last very long.
4. I am screwed for having relied on them.
Q.E.D. (Quit Everything's Dead)
> As far as support, you can't avoid risks. Does MicroSoft give your
> problems the highest priority?
It used to always be said "No one ever got fired for buying IBM."
Lamentable as it might be, the same could often be said for MicroSoft today.
I have no great love for MicroSoft, but if I had a choice between apologizing
to a boss or a customer because Linux failed or because MicroSoft failed,
I'd take MicroSoft every time. People mostly don't know much about the
technical details of either system, but one thing I'll be statistically
many people do think they know is that MicroSoft is a commercial quality
business solution and that when it lets you down, it can't be helped, while
Linux has not done the work to make the same case to most people. There may
be narrow markets where the reverse is true, and where people trust Linux
more than MicroSoft. But if I had to make the case to a person I knew
nothing about? I'd go with the "MicroSoft screwed me" excuse because
statistically I think I'd still have my job or my customer afterward.
(Incidentally, I say this with Linux running on my web server so please do
not confuse this with Linux-bashing. This is a statement about economic and
political reality.)
Also, the fact is that with a valid Microsoft OS license and no support,
I can call MicroSoft for $35 on my credit card when something advertised
is broken, and they will put a dedicated person with real-time access to
the developers on the phone to me and try to fix it. I don't know a
similarly low-priced or effective way of getting a fix for Linux from a
paid 24/7 connected employee. And I think the reason is that they don't
think they're selling something whose level of quality makes or breaks sales,
BECAUSE they are not selling it. They think that only long-term issues
affect them but they have no apparent commitment to short-term issues.
MicroSoft may be many things, but it certainly knows tehre are short-term
issues directly affecting its position in the marketplace and it has staffed
people to address this problem.
> How much influence does your money have on them, when their interest
> is elsewhere. And what happens when your sofware supplier goes
> under?
My $35 has gone a long way the two or three times I have needed a fix right
now from Microsoft. I can't tell you how many hundreds of kernels I have
laboriously recompiled after reading this and that conflicting advice off
of Deja News trying to do fixes "the Linux way". The gap is enormous between
the level of quality provided by this model "free software" model and the
lamentable but de facto commercial alternative.
> Ultimately major GPL software efforts will live or die based on how
> they fit our needs.
They're not exactly batting 1000 with me. Like anything given away
free, one cannot claim it has no value. But often the value is
deceptive. I would quickly prefer a commercial alternative to
anything I can get for free. I mostly use free software because it
does different things than the commercial variants do. Sadly, that's
what we really stadn to lose. I just don't understand why we can't
just tell those people with a different vision "pursue your dream" AND
"charge me for it so you can stay in business". If they do the "die"
part of your previous statement, it will be only because someone
falsely told them there was something magic about "free" that would
keep them fed. It won't be some accident of having not fulfilled a
societal need. it will be because no number of free copies given up ever
adds up to enough to pay for a cup of coffee.
> I don't consider the GPL a magical incantation
> which can destroy or save our society.
I actually do. I think it is deceptively able to disempower a whole class
of would-be powerful people in our society. It undermines the ability to
accumulate wealth. And while wealth must certainly be dealt with in a
dignified way, if the same energy went into preparing our presently-wealthy
people to be kind and considerate to society as goes into indoctrinating
free software people, we wouldn't have a callous wealthy--or, at least, we'd
have as optimal a wealthy class as we can reasonably expect. Wealth is a
double-edged sword, yes, but at least with it comes freedom for some.
With free software comes the turning of the whole of creation of software in
society into a board game full of "Monopoly(R) money", and the turning of the
people with the talent to manage that monopoly money into peasants because
no amount of winning will ever buy them out of their day job.
>>>>> "KMP" == Kent M Pitman <······@world.std.com> writes:
KMP> [...] I have no great love for MicroSoft, but if
KMP> I had a choice between apologizing to a boss or a customer
KMP> because Linux failed or because MicroSoft failed, I'd take
KMP> MicroSoft every time.
If you'd gone with Microsoft, it would be because either they have some
unique functionality (like talking to other MS stuff) or that you do not
know about anything else. In the first case, there really was no decision
to make so you cannot be blamed for a non-decision. For the second case
you are known to be a Microsoft-only person, possibly hired because of
or despite that so again you are not to blame. The bar is higher for
Linux (BSD, Solaris or whatever) because the implicit promise is that
you will not just install the thing and hope for the best, but you will
actually deliver the solution. So yes, in case of failure I'd much
rather be the MS person too, much the way I'd like to be a kid again
when I am burdened with nasty adult responsibilities. Note that I moved
the goal-posts slightly and posed the question as MS vs. non-MS as
opposed to MS vs. Linux. This not only suits me for my argument, but
also, I believe, it explains the 'social' setting better.
KMP> People mostly don't know much about
KMP> the technical details of either system, but one thing I'll be
KMP> statistically many people do think they know is that
KMP> MicroSoft is a commercial quality business solution and that
KMP> when it lets you down, it can't be helped, while Linux has
KMP> not done the work to make the same case to most people.
You are absolutely right that people mostly don't know much about the
technical details. But if they are bringing in somebody, they already
know they don't know enough and you are undertaking the resposibility
to deliver a solution, not a product. But you do have a good
answer to this below:
KMP> There may be narrow markets where the reverse is true, and
KMP> where people trust Linux more than MicroSoft. But if I had
KMP> to make the case to a person I knew nothing about?
Very good question. But I suspect this might be subsumed by the
platform independent version of the same question: How do you convince
people that there is value in whatever it is that you are selling? If
you are selling something that needs to be on a non-MS platform and
you have to answer the question about why you are not using Microsoft
either you communicated the product description the wrong way or
the platform is relevant for the customer. If the latter, then one
would assume that your decision had sound technical reasons behind it
and defending your choice succesfully is a marketing problem. (I am not
sure we are not making the same observation on this.)
KMP> I'd go
KMP> with the "MicroSoft screwed me" excuse because statistically
KMP> I think I'd still have my job or my customer afterward.
Yes, but my contention is that it would be the kind of job where you did
not have the choice in the first place.
KMP> (Incidentally, I say this with Linux running on my web server
KMP> so please do not confuse this with Linux-bashing. This is a
KMP> statement about economic and political reality.)
I understand.
KMP> Also, the fact is that with a valid Microsoft OS license and
KMP> no support, I can call MicroSoft for $35 on my credit card
KMP> when something advertised is broken, and they will put a
KMP> dedicated person with real-time access to the developers on
KMP> the phone to me and try to fix it.
I'm seems it worked for you. I have no personal experience with the
$35 support. They also had a $200+ support for NT server that I have
seen used -- you basically got to re-install several times with different
click scenarios. Since we are going by anactodal evidence, two fun
pieces come to mind:
http://www.ai.mit.edu/~shivers/rants/msoft.html
http://www.bmug.org/news/articles/MSvsPF.html
MKP> I don't know a similarly
KMP> low-priced or effective way of getting a fix for Linux from a
KMP> paid 24/7 connected employee. [...]
You are correct (unless you are abusing a bright college kid). There
are several attempts to remedy this:
www.hotdispatch.com and www.questionexchange.com
but I am not sure they would be satisfactory for you.
KMP> [...] My $35 has gone a long way the two or three times I have
KMP> needed a fix right now from Microsoft.
The more common way that sometimes works is paying money into this
cottage industry of small consulting houses to fix things for you.
The prices vary between $80-150/hr and these folks mostly reinstall and
tweak. I imagine you are talking about desktop issues, as I think the
MS rate for server stuff is higher.
KMP> I can't tell you how
KMP> many hundreds of kernels I have laboriously recompiled after
KMP> reading this and that conflicting advice off of Deja News
KMP> trying to do fixes "the Linux way".
Were these driver issues or something more serious (syscalls, libc not
behaving etc.)
[the rest elided, I think I understand the rest but do not have a good
opinion on it, thanks for sharing though]
cheers,
BM
Kent M Pitman <······@world.std.com> writes:
> > I don't consider the GPL a magical incantation
> > which can destroy or save our society.
>
> I actually do. I think it is deceptively able to disempower a whole
> class of would-be powerful people in our society. It undermines the
> ability to accumulate wealth. And while wealth must certainly be
> dealt with in a dignified way, if the same energy went into
> preparing our presently-wealthy people to be kind and considerate to
> society as goes into indoctrinating free software people, we
> wouldn't have a callous wealthy--or, at least, we'd have as optimal
> a wealthy class as we can reasonably expect. Wealth is a
> double-edged sword, yes, but at least with it comes freedom for
> some. With free software comes the turning of the whole of creation
> of software in society into a board game full of "Monopoly(R)
> money", and the turning of the people with the talent to manage that
> monopoly money into peasants because no amount of winning will ever
> buy them out of their day job.
Wow.
At least we know where your critique is coming from, well that was
obvious before, but this puts a bow on it.
Don't worry Kent, if I see you up agains the wall, I'll let the Free
Software communards know all the stuff you did and maybe you'll just
get sent to the re-education camps instead of shot.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
Kent M Pitman <······@world.std.com> writes:
> It used to always be said "No one ever got fired for buying IBM."
>
> Lamentable as it might be, the same could often be said for MicroSoft today.
No, not at all.
I don't have the URL handy, but someone _has_ been fired for buying
Microsoft.
--
(let (hh mm ss)
(do () (nil) (multiple-value-setq (ss mm hh) (get-decoded-time))
(format t ····@R ·@R ·@R "
#\return hh mm ss) (sleep 1)))
In article <···············@world.std.com>,
Kent M Pitman <······@world.std.com> wrote:
>
>This point, exactly as you've expressed it, is often the source of problem.
>You speak as if you mean to say:
>
> 1. If a maintainer puts ideology above working code, they won't last very
> long.
>
> 2. THEREFORE, the maintainer will not put ideology above working code.
>
> 3. THEREFORE, I will be happy.
>
> Q.E.D. (Quit Everything's Done)
>
>But there is another scenario consistent with your statement.
>
> 1. If a maintainer puts ideology above working code, they won't last very
> long.
>
> 2. Maintainer puts ideology above working code.
>
> 3. THEREFORE, they won't last very long.
>
> 4. I am screwed for having relied on them.
>
> Q.E.D. (Quit Everything's Dead)
>
This happens in both free and commercial software. When the maintainer
of commercial software goes under for ideological reasons, there is
very often no way to recover. When the maintainer of free software
goes under for ideological reasons, there's a codebase out there.
It is not necessary to restart from scratch.
>Also, the fact is that with a valid Microsoft OS license and no support,
>I can call MicroSoft for $35 on my credit card when something advertised
>is broken, and they will put a dedicated person with real-time access to
>the developers on the phone to me and try to fix it. I don't know a
>similarly low-priced or effective way of getting a fix for Linux from a
>paid 24/7 connected employee.
Is there a market for it? There's no reason why somebody can't set
themselves up in Linux support and run a 24/7 operation that takes
credit cards. It may seem to cost more, since MS might feel it
should lose money on support for whatever reason, but I tend
to think that accurate prices are a good thing for the economy.
>> How much influence does your money have on them, when their interest
>> is elsewhere. And what happens when your sofware supplier goes
>> under?
>
>My $35 has gone a long way the two or three times I have needed a fix right
>now from Microsoft.
Wow. That sort of thing has never worked for me.
>> Ultimately major GPL software efforts will live or die based on how
>> they fit our needs.
>
>They're not exactly batting 1000 with me. Like anything given away
>free, one cannot claim it has no value. But often the value is
>deceptive. I would quickly prefer a commercial alternative to
>anything I can get for free.
Lots of people feel that way. There's nothing wrong with that.
It gives commercial software a competitive advantage.
I mostly use free software because it
>does different things than the commercial variants do.
Yup. So, if there was no free software, you'd have to get by
with the commercial stuff that costs money *and* doesn't do
what you want it to do. I have no faith in the idea that there's
somebody out there just aching to sell me exactly what I need
for a price I can afford, but who's discouraged by free software.
If they had exactly what I needed at a price I could afford, they'd
be out there collecting money.
Sadly, that's
>what we really stadn to lose. I just don't understand why we can't
>just tell those people with a different vision "pursue your dream" AND
>"charge me for it so you can stay in business".
Um, there's two things here.
If the software fulfils your needs as is, then it doesn't matter if
the developer stays in business, you've got the software. If not,
then you can always pay that, or another, developer to make the
changes. If your dream is to give something to the community, then
it really is counterproductive to tell them to pursue their dream
and get paid.
>> I don't consider the GPL a magical incantation
>> which can destroy or save our society.
>
>I actually do. I think it is deceptively able to disempower a whole class
>of would-be powerful people in our society. It undermines the ability to
>accumulate wealth.
Um, huh? It undermines the ability to charge large amounts of money
for a shoddy product and get rich that way. You have no right to
demand that people give you large amounts of money for whatever it
is you want to give them. If you can give people something more
useful than what they already have, you can charge for it and
accumulate as much wealth as you can manage.
The only way this would be destructive would be if the GPL or
some variant thereof were to be mandated by higher authority.
If not, then commercial software can thrive exactly as long as
it can provide more value than the free stuff. It may lower
the amount the commercial people can charge for it, but then it
may lower their costs.
For example, any development effort really does need a version control
system. You can get CVS for free, and it works well. Lots of people
use it. If you had to pay for something, it would be a cost.
Now, this doesn't stop lots of people from selling version control
systems, sometimes for big bucks. The existence of CVS doesn't
seem to be putting Rational out of business, and if they didn't
have to worry about CVS competing with ClearCase they'd have to
worry about hordes of low-priced version control systems. In the
meantime, if you want ClearCase (and I've heard it's *very* nice),
you can pay for it. In the meantime, if you think Rational tools
would break your budget, you don't have to use their stuff.
If the GPL succeeds by making it unprofitable to write commercial
software, then the writers of commercial software simply aren't
contributing the value they need to command the compensation they
want, and from a pure market point of view they don't deserve
to be paid.
And while wealth must certainly be dealt with in a
>dignified way, if the same energy went into preparing our presently-wealthy
>people to be kind and considerate to society as goes into indoctrinating
>free software people, we wouldn't have a callous wealthy--or, at least, we'd
>have as optimal a wealthy class as we can reasonably expect.
It's been tried. Really hard.
For *centuries*, there have been people out there telling the wealthy
to be kind and considerate. These people have been professionals, and
have had great authority. They have claimed, and been acknowledged
to have, immense power over people's destinies.
I refer, of course, to the clergy. They have pleaded. They have
thundered. They have threatened. They have been ignored.
Why do you think Richard Stallman is going to have any better luck
telling the wealthy how to behave than any given Pope?
Wealth is a
>double-edged sword, yes, but at least with it comes freedom for some.
Which you apparently want to gain by denying me the right to write
something and give it away. You want to establish truly fascist
controls over software distribution in order to make it possible
for people to grow fat. (How do you expect to enforce this without
jackbooted accountants breaking down my door and confiscating my
financial records to make sure I have been paid for every copy
that I have allowed out of my possession?)
>With free software comes the turning of the whole of creation of software in
>society into a board game full of "Monopoly(R) money", and the turning of the
>people with the talent to manage that monopoly money into peasants because
>no amount of winning will ever buy them out of their day job.
Look, I'm sorry for all of you that feel you just can't live in a
society with a market economy, but the fact is that if you can't
provide something better than what's already out there you don't
deserve to get paid. If you can provide something better you can
get paid. There are countries out there that don't have a market
economy that you could move to if you wanted, but the fact is that
those countries are poor and probably wouldn't suit you. (On the
other hand, the distribution of wealth tends to be very uneven.)
So, if you want to make it in the *real* world, don't whine for
government operatives to track down and destroy your competition
for you. If you can't push the computing world in the direction
you want because of harsh fiscal realities, *you're no different
from the rest of us*. I do interesting work for a good salary.
It isn't what I would be doing if I didn't need that salary.
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
In article <··············@piracy.red-bean.com>, Craig Brozefsky
<·····@red-bean.com> wrote:
>If we look at RH's annual report (the only one I could find was for
>2000) we can see that they were making 56% percent of their revenue on
>subscriptions, that's people paying for GPLed code that is available
>elsewhere, 23.9 million dollars from GPLed code.
No, no, no, no no!
If you look at your RedHat licence, you are *not* paying for the GPL
code. You are paying for installation tech support - the right to email
RedHat techs, and have them troubleshoot your installation problems,
*not* for the code, *or* the binaries, which you can get for much less
elsewhere (cheapbytes, or downloading and burning an ISO).
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
Raffael Cavallaro <·······@mediaone.net> writes:
> In article <··············@piracy.red-bean.com>, Craig Brozefsky
> <·····@red-bean.com> wrote:
>
> >If we look at RH's annual report (the only one I could find was for
> >2000) we can see that they were making 56% percent of their revenue on
> >subscriptions, that's people paying for GPLed code that is available
> >elsewhere, 23.9 million dollars from GPLed code.
>
> No, no, no, no no!
>
> If you look at your RedHat licence, you are *not* paying for the GPL
> code. You are paying for installation tech support - the right to email
> RedHat techs, and have them troubleshoot your installation problems,
> *not* for the code, *or* the binaries, which you can get for much less
> elsewhere (cheapbytes, or downloading and burning an ISO).
What is the difference between this code shrink-wrapped with
documentation and support contracts, and proprietary software
shrink-wrapped with documentation and support contracts other than the
fact that one is GPLed and the other isn't?
Is it that you an get one free online? For most peices of proprietary
software one can get them free online as well, pirated.
Your objection seems circular to me. People can pay money for it, but
even then, they are not paying money for it because they can get it
elsewhere cheaper, therefor people always are paying the lowest
theoretical price for the software. Since I can get proprietary
software for free, does this mean that I am never paying for
proprietary software either, since the lowest possible price is zero
if I pirate it?
Or do you discount the possibility of paying zero for proprietary
software since it is illegal to pirate it in most countries? This
seems to run contrary to most established software pricing guidelines,
which assume piracy and price such that the market will not turn to
pirate copies in a sufficient number of cases to generate a profit.
If proprietary vendors are allowed to compensate for pirated copies in
their pricing model and still be said to be selling software, then why
can't GPL vendors consider the availability of the GPLed software from
other places in their pricing model as well and yet not be considered
as selling software?
It's only in a theoretical world with no market friction that one
would model the impact of GPLed software availability as reducing your
market price to zero. We don't live in that world tho. Proprietary
software has a legal monopoly which usually, but certianly not always,
causes sufficient friction for non-vendor transactions that a
significant portion of the market goes with the vendor and bears their
price. GPLed software has no such legal monopoly, but there is
certianly friction there, enough to support businesses in many cases,
or multiple businesses in a few.
--
Craig Brozefsky <·····@red-bean.com>
http://www.red-bean.com/~craig
"Indifference is the dead weight of history." -- Antonio Gramsci
In article <············@news.ox.ac.uk>, ·····@comlab.ox.ac.uk (Philip
Armstrong) wrote:
>Larry Wall wrote Perl long before he was employed by O'Reilly.
Yes, because it helped him get his job (for which he was being paid)
done. But he couldn't afford to devote his full time to it until
O'Reilly paid him to do so.
And even after many years of work, Perl still sucks compared to lisp as
a language, but exceeds it in availability of useful libraries.
Again, Common Lisp is *very* difficult to implement well. That means
full time work for months/years. Who's going to pay for that?
Raf
--
Raffael Cavallaro, Ph.D.
·······@mediaone.net
On Mon, 21 May 2001 03:30:28 GMT, Craig Brozefsky <·····@red-bean.com> wrote:
>Agreed. I think that a whole LOT could be done if a FFI library could
>be created that ran efficiently in all of the existing
>implententations and still allowed for the particular features and
>strengths of each implementation to be used.
To me this sounds like a lot more difficult and complicated than picking
one implementation and making sure it has a clean and well-documented FFI.
It seems to me that by it's very nature, an FFI is going to have to
interact with the implementation in a fairly intimate way and doing this in
a portable way is going to be hard. If everyone handles threads and memory
management differently, for example, how do you write FFI code that deals
with threads and object allocation portably?
>From writing applications which build on the open libraries, just like
>in the other langauges with open implementations. Is there really a
>need for these basic services like FFI, sockets, XML tools and the
>like to be proprietary? I don't think so. I think that their
>presence would be an economic boon to users and vendors alike.
Exactly. Everyone needs this stuff. It's not that much fun to write, but
it's really indispensable for most modern applications. Why not write it
once and get on with it? Nobody's going to make any money selling XML
tools in today's software climate anyway.
>I can't see a single implementation ever emerging as dominant either,
>and thank god for that. I think it's important for the long-term
>health of the language to have these different implementations, even
>if it does present a standardization challenge.
I'm really not sure I see the advantage of having all these subtly
divergent implementations. What do we gain from this? Would we all
applaud if the Linux kernel forked into one kernel that supported smp, one
kernel that supported usb, and one kernel that supported a logging
filesystem? I don't think our situation is so different.
--
miles egan
·····@puzl.pixar.com (Miles Egan) writes:
> >I can't see a single implementation ever emerging as dominant either,
> >and thank god for that. I think it's important for the long-term
> >health of the language to have these different implementations, even
> >if it does present a standardization challenge.
>
> I'm really not sure I see the advantage of having all these subtly
> divergent implementations. What do we gain from this? Would we all
> applaud if the Linux kernel forked into one kernel that supported smp, one
> kernel that supported usb, and one kernel that supported a logging
> filesystem? I don't think our situation is so different.
If anything the situation is more comparable to Linux, FreeBSD,
NetBSD, OpenBSD and Darwin (as well as all the commercial Unices) all
existing alongside each other. I don't think that any of those OSs is
going to fold out of its own free will, and join some other camp.
Each of them has its own strengths and weaknesses, and none of them is
going to appeal to all the uses and users out there (not to mention
the developers themselves).
I don't think current users of CLISP will be happy with CMU CL, either
now or anytime soon, and vice-versa. There's even enough difference
in priorities and interests that SBCL and CMU CL will likely never
merge, and I don't consider that a bad thing either.
The hard truth of live is that there are no ideal systems, because a
system is always the result of tradeoffs (e.g. native-code compiler
vs. easy porting, size vs. speed vs. features, etc.). There will
always be as many systems as there are possible tradeoffs, interested
users and resources.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
On 21 May 2001 23:08:05 +0200, Pierre R. Mai <····@acm.org> wrote:
>> I'm really not sure I see the advantage of having all these subtly
>> divergent implementations. What do we gain from this? Would we all
>> applaud if the Linux kernel forked into one kernel that supported smp, one
>> kernel that supported usb, and one kernel that supported a logging
>> filesystem? I don't think our situation is so different.
>
>If anything the situation is more comparable to Linux, FreeBSD,
>NetBSD, OpenBSD and Darwin (as well as all the commercial Unices) all
>existing alongside each other. I don't think that any of those OSs is
>going to fold out of its own free will, and join some other camp.
I agree this is a better analogy. Portability among these platforms, at
least for open source software isn't too painful. I attribute this to the
fact that they all support a fairly precisely defined and extensive systems
programming api, and that for higher level coding toolkits like GTK are a
de facto standard. If Lisp implementations offered the same level of
extensive low-level compatibility then the situation would be much happier.
A quick perusal of the defsystem code, for example, is a good indication of
the obstacles we face now in writing low-level and portable Lisp code.
And, as much as I love FreeBSD I think Linux is eroding its edge fairly
quickly. The concentration of effort brought to bear on Linux development
is pretty hard to beat. Cooperation carries the day in the end.
>I don't think current users of CLISP will be happy with CMU CL, either
>now or anytime soon, and vice-versa. There's even enough difference
>in priorities and interests that SBCL and CMU CL will likely never
>merge, and I don't consider that a bad thing either.
This seems like a fairly small freedom to me that comes at a high price.
Are CLISP and CMUCL really so different? Would it really be so hard to
pool the efforts all the developers and come up with a single system that
combined all the best features of each? I realize there are some
inevitable tradeoffs here but there has to be a sweet spot we can all build
on.
I suspect you're right and that no such merger is on the horizon. I think
my strategy will be to just pick an implementation (probably CMU) and treat
it as if it were the only one.
--
miles egan
·····@puzl.pixar.com (Miles Egan) writes:
> >If anything the situation is more comparable to Linux, FreeBSD,
> >NetBSD, OpenBSD and Darwin (as well as all the commercial Unices) all
> >existing alongside each other. I don't think that any of those OSs is
> >going to fold out of its own free will, and join some other camp.
>
> I agree this is a better analogy. Portability among these platforms, at
> least for open source software isn't too painful. I attribute this to the
> fact that they all support a fairly precisely defined and extensive systems
> programming api, and that for higher level coding toolkits like GTK are a
> de facto standard. If Lisp implementations offered the same level of
> extensive low-level compatibility then the situation would be much happier.
Oh, but we do, actually. Try porting a non-trivial threaded
application between different unices, and you'll see how good
portability is. Most widely-ported Unix software is full of ugly
hacks and portability fixes, that it makes even MK-DEFSYSTEM look
nice.
But maybe you've got a point, in that the new standard for portability
seems to be some form of Linux API/ABI compatibility, with Sun, IBM,
SCO and the *BSDs trying to emulate Linux. Of course this emulation
stuff only works for run-of-the-mill programs most of the time, with
more involved code (like CMU CL) still not working.
> This seems like a fairly small freedom to me that comes at a high price.
> Are CLISP and CMUCL really so different? Would it really be so hard to
> pool the efforts all the developers and come up with a single system that
> combined all the best features of each? I realize there are some
> inevitable tradeoffs here but there has to be a sweet spot we can all build
> on.
I don't see how you will get "portable" native compilation without
going via gcc, i.e. the KCL route (exemplified today mostly in GCL and
ECLS). But since GCC is terribly slow, this will heavily penalize
run-time uses of the compiler (like e.g. to compile effective method
functions in CLOS) to no end. This in itself will very likely satisfy
neither the CLISP nor the CMUCL user bases. And you have to write
terribly clever C code generators in order to achieve the same level
of performance that CMU CL offers for e.g. numerical code. I know of
at least one user/maintainer that might not be happy with this.
And that's only one very obvious place where the compromise solution
will probably suite none of us.
> I suspect you're right and that no such merger is on the horizon. I think
> my strategy will be to just pick an implementation (probably CMU) and treat
> it as if it were the only one.
That's a completely different tactic, and many/most CL users do such.
In fact we have been using CMU CL exclusively in most projects in the
past couple of years.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
On 22 May 2001 01:26:04 +0200, Pierre R. Mai <····@acm.org> wrote:
>I don't see how you will get "portable" native compilation without
>going via gcc, i.e. the KCL route (exemplified today mostly in GCL and
>ECLS).
You won't, of course, but I think Ocaml sets a good example here. The
bytecode interpreter is extremely portable and fast and is decoupled enough
from the native compiler that it can run on platforms for which there is
no native backend. The native compiler is available as a second compiler
on supported platforms. This means it's fairly easy to bootstrap Ocaml on
a new architecture in bytecode and follow up with native support as time
allows. I don't see any technical reason why this couldn't be done for
Lisp.
--
miles egan
"Pierre R. Mai" <····@acm.org> writes:
> The hard truth of live is that there are no ideal systems, because a
> system is always the result of tradeoffs (e.g. native-code compiler
> vs. easy porting, size vs. speed vs. features, etc.). There will
> always be as many systems as there are possible tradeoffs, interested
> users and resources.
I disagree. You don't buy a new brain each time you face different
problems with different tradeoffs. You go in a different mode.
At the meta-level, your one brain can manage the diversity;
it specializes then runs a specialized instance against the problem at hand.
Besides, that's precisely the LISP way: having a generic language
where you define domain-specific languages.
Well, the same principe can, and should apply, at the compile-level:
having different implementation tactics to dynamically choose from
(or to dynamically assemble together), so as to get a compiler
taylored to one's needs, dumping native code when possible, or dumping
C code when needed, etc. Besides, systems like MACISTE or Pliant do it.
OCAML has always had a C bytecode interpreter since caml-light,
and a native code compiler since caml-special-light. Both co-developped,
fully compatible at the source level, mostly compatible at the underlying
FFI level.
Yours freely,
[ Fran�ois-Ren� �VB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
To people who ask what is the business model of Free Software, I reply:
"what is the business model of Free Trade?".
Francois-Rene Rideau <·······@SPAM.tunes.org> writes:
> "Pierre R. Mai" <····@acm.org> writes:
> > The hard truth of live is that there are no ideal systems, because a
> > system is always the result of tradeoffs (e.g. native-code compiler
> > vs. easy porting, size vs. speed vs. features, etc.). There will
> > always be as many systems as there are possible tradeoffs, interested
> > users and resources.
>
> I disagree. You don't buy a new brain each time you face different
> problems with different tradeoffs.
Well, FWIW, I disagree with your disagreement. Companies lay off
employees with one brain to hire people with others exactly for this
reason. Perhaps if the same person could rent different brains,
layoffs of the person would be less necessary.
> You go in a different mode.
> At the meta-level, your one brain can manage the diversity;
In the far distant future, the problems we're working on with
computers may so radically underutilize the computer that they mirror
the way human tasks underutilize the human brain. In cases where
there's plenty of excess capacity, sure, one can substitute for the
other without apparent difference. But whenever the processor is used
to capacity, the details of its architecture will prefer some tasks
over others, just as happens with pre-trained human wetware when used
to capacity.
Kent M Pitman wrote:
> I really reject the notion that open-source is the answer to anyone's
problems,
> though I don't have any problem with someone doing it.
I think most people haven't figured out that open source really only works
if the software in question is:
1. already really good
2. architected by a small group of experts in the field
The majority of "pure" open source software--meaning software written by
gangs of loosely connected programmers--seems to be written by college
students looking for something to occupy their time and who don't have much
software engineering experience. The really good stuff (e.g. OCaml, Corman
Lisp, Squeak) comes from small groups of people who have a great product,
and release the source more as a show of good faith than anything else, not
because they need help with the work.
James
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
"James Hague" <···········@volition-inc.com> writes:
> The really good stuff (e.g. OCaml, Corman
> Lisp, Squeak)
Ocaml is backed by governement money. It's a research project.
Corman Lisp is sold but you can test it out freely.
Squeak was and is supported by someone, because they want to develop
some IIRC multimedia stuff in it
So all are obviously in a position where
a) they are paid for the work on the product
b) they get their income from elsewhere but the product.
> comes from small groups of people who have a great product,
> and release the source more as a show of good faith than anything else, not
> because they need help with the work.
Now, some probably do not need help. But maybe some do. We simply do
not know
Regards
Friedrich
On Sun, 20 May 2001 19:28:17 GMT, Kent M Pitman <······@world.std.com> wrote:
>> ** comments about the lack of standardization in CL deleted ***
>
>While there's some truth to what you say here, I'm not convinced that
>most of these need to be available through the implementation. I
>think your conclusions here would come out at least a bit different if
>you would separate things that "must come from the implementor" from
>things that "could come from anywhere".
I agree and I think I wasn't clear enough on this point. I don't think
it's necessary that CMUCL come with it's own MySQL bindings, for example.
I do think it's essential that such bindings exist and that they easy to
download, install and use. I think Perl and Python provide a reasonable
model here.
>> 3. As a result of this fragmentation, Lisp is actually a poor choice for
>> many important kinds of application development.
>
>This seems an overbroad conclusion. Most applications aren't developed for
>deployment in all lisps initially, but in one. And in individual lisps the
>support is there to do a lot.
Perhaps in the commercial lisps, but the available free lisps are missing
things I've needed in order to do the work I do.
>> More and more it is
>> pushed into the role of a central application intelligence interfaced with
>> more mainstream applications which provide connectivity to the rest of the
>> world. Only its superb core design has allowed to retain even this
>> position. But for many other applications, the burden of maintaining this
>> kind of homogeneous system is too high and it really is easier to develop
>> in a more gregarious language.
>
>This is personal taste and I doubt can be supported. I don't find it easier
>in so uniform a fashion. It's apples and oranges, really.
It may be a personal taste, but it's also certainly a common taste and I
don't think it can be dismissed so simply. A lot of engineers I know
prefer to keep the number of different languages used in a single project
to a minimum for simplicity's sake.
>You've obviously thought hard about this issue. I would encourage you to
>think harder about where to apply pressure and resources to fix it.
>What's depressing is assuming it will remain static. There are enegetic,
>interested people stillin the mix. Things can change. Not everyone can figure
>out what to do to contribute. Leaders are called for.
I don't assume it will remain static and I am doing my best to find ways to
fix some of the problems I perceive in CL now. That instinct prompted my
first message. I do think there is cause to be optimistic. We have an
embarassment of smart hackers, an inspired language design, and decades of
shared experience solving real problems.
>I really reject the notion that open-source is the answer to anyone's problems,
>though I don't have any problem with someone doing it.
Here I disagree. I see open-source as Lisp's greatest hope because of two
inexorable industry trends. One, giants like IBM and Sun and Microsoft are
dumping their development tools for free or at a loss to further their
larger agendas and backing them up with colossal marketing budgets. Two,
there are now several mature and full-featured opensource development
languages. Eking out a living between these two extremes is going to get
harder and harder, I think.
>be created to sustain us. Saying "others are giving stuff away, we must too"
>isn't an economic model. Where are Lisp people expected to make their money?
I don't think anyone is going to be able to make a living selling the kind
of basic plumbing code I'm talking about. I think Lisp people are going to
make money by using a shared, mature and full-featured Lisp to write
reliable, sophisticated complex applications. All Lisp programmers benefit
if we all pool our resources to develop a common platform and toolkit.
Writing this kind of stuff isn't particularly enjoyable, but it's
absolutely essential for getting any work done. If we do it together, we
do it once, make sure it's done well, and then start making money by
solving more interesting problems.
>> 2. A simple and straightforward means of bootstrapping the language on any
>> sane platform with a c compiler. All of these languages are very portable.
>
>Most existing implementations are already built on C cores. That doesn't
>mean they're always portable.
I know, but I think it does make it easier. If you're careful, porting
shouldn't require much more than a recompile.
>> 3. Take advantage of having a single reference implementation to define a
>> clean and straightforward CPAN-like system and a well-documented and
>> straightforward C FFI. I know there are efforts to define something like
>> this now, but, to my eyes, they are all severely limited by the difficulty
>> of supporting so many implementations.
>
>What stops you in what exists now?
>
>> 4. Build full-featured and reliable interfaces to common toolkits and
>> protocols on this infrastructure. Lisp should talk to databases as
>> fluently as Perl, parse XML as well as Python, speak SNMP, HTTP, LDAP,
>> IMAP, and SOAP, and couple with C as easily as Ruby.
>
>What stops you in what exists now?
Like a lot of free software hackers, I'm motivated by the reward of
challenging my mind and creating new things and by the reward of seeing my
software used by others and extended by others into something I could never
have built on my own. Perusing some of the more actively developed free
extension packages I find a morass of implementation-specific code,
duplicated ffi code, and incomplete or nonexistent documentation. As a
prospective free Lisp contributor, this suggests to me that I'm going to
have to waste a lot of time on implementation-specific tweaks and that the
other packages developers are going to need to make anything useful out of
mine are too hard to use so nobody's going to bother with mine either.
This undermines a lot of my enthusiasm.
My inclination at this point is to just proceed as if CMUCL is the only
free Lisp that really matters and start doing what I can to help it grow.
>> 5. Take advantage of having a single implementation to extend and change,
>> when necessary, the Common Lisp standard.
>
>Why? It's VERY expensive to change. It cost just shy of a half a
>million dollars to produce the standard in costs that we can easily
>measure, and maybe again as much in hidden costs.
I don't want to dwell too much on this point, because it's really
peripheral to the main thrust of my argument. I do think we should have a
standard threading api, and a standard networking api. I'm really not
sophisticated enough in Lisp to make any specific criticisms of the
standard. What I'm really advocating is a willingness to extend it.
>> Speaking as an open-source hacker that would love to do what he can to help
>> promote and grow lisp, I think these kinds of changes are necessary. As it
>> stands now, trying to get anything done in the current Babel is even more
>> frustrating than watching Ruby, Perl, and Python slowly reinvent Lisp the
>> hard way.
>
>IMO, you're better off just making eganlisp as a new language,
>dialect, or implementation, whatever you seeok to make, taking anying
>from public sources that you like, and ignoring anything you don't
>like, and working from there.
This is the last thing I want to do. More than anything I'd like to see us
all pool our resources and do what we can to share code and spend less time
reinventing and more time inventing.
>I think success, if it comes, will come from people accepting that not every
>language is the same, and that we are a community which has many bases.
>We should work to have a set of common libraries and stop fussing over
>the base.
I basically agree with you here and I'm mainly trying to address what CL
leaves out and not what it includes. As I see it, right now we're wasting
a lot of energy trying to provide these common libraries for so many
implementations with the result that programmers in other languages take
such libraries for granted while ours grow at a snail's pace. I really do
think it's a huge impediment to our collective productivity. I can cite
specific examples if I'm not making myself clear. We're a small but nimble
community that can't afford to fight a six-front war. United we stand.
By the way, I've noticed that this thread has degenerated into some fairly
uncivil behavior, for which I'd like to apologize by proxy. I'd also like
to thank you for your willingness to participate in these discussions in
spite of such rudenees. I understand your concerns and I think they're
valid. I find the sense of untitlement in younger hackers who regard the
current wealth of free software as fait accompli as unattractive as you do.
--
miles egan
·····@puzl.pixar.com (Miles Egan) writes:
You wrote:
>
> I know, but I think it does make it easier. If you're careful, porting
> shouldn't require much more than a recompile.
And then later
>
> My inclination at this point is to just proceed as if CMUCL is the only
> free Lisp that really matters and start doing what I can to help it grow.
>
I don't think you've looked at CMUCl very carefully. It's a serious
native compiler not built on top of C. Porting it requires that you
write a backend for the machine you are porting to, and then bootstrap
the system from an existing CMUCL port. This is hard work.
--tim
On 21 May 2001 17:19:33 +0100, Tim Bradshaw <···@tfeb.org> wrote:
>You wrote:
>
>>
>> I know, but I think it does make it easier. If you're careful, porting
>> shouldn't require much more than a recompile.
>
>And then later
>
>>
>> My inclination at this point is to just proceed as if CMUCL is the only
>> free Lisp that really matters and start doing what I can to help it grow.
>>
>
>I don't think you've looked at CMUCl very carefully. It's a serious
>native compiler not built on top of C. Porting it requires that you
>write a backend for the machine you are porting to, and then bootstrap
>the system from an existing CMUCL port. This is hard work.
I know. I'm all too painfully aware of this, in fact and this is really
the crux of my argument. CLISP is highly portable, supports unicode and
runs on 64 bit machines, but is relatively slow, has no native compiler and
doesn't support threads. CMUCL has a superb compiler and thread support,
but doesn't (AFAIK) support unicode, 64 bit platforms, or really many
platforms beyond x86 and sparc.
I may soon embark on a large development project on an Alpha server
cluster. I need threading, 64 bit support, unicode, and performance. I'm
out of luck. I have to wonder if we were all working on one implementation
if I'd be able to use a free lisp for this project.
By the way, I'd really like to rephrase my earlier statement that CMU is
"the only free Lisp that matters". I didn't intend this to sound like a
slight to CLISP or SBCL. I think a lot of good work has been done in both
projects. What I really meant to say is that I think CMU is closer to
ready for real production work than any other free Lisp.
--
miles egan
·····@puzl.pixar.com (Miles Egan) writes:
> I know. I'm all too painfully aware of this, in fact and this is really
> the crux of my argument. CLISP is highly portable, supports unicode and
> runs on 64 bit machines, but is relatively slow, has no native compiler and
> doesn't support threads. CMUCL has a superb compiler and thread support,
> but doesn't (AFAIK) support unicode, 64 bit platforms, or really many
> platforms beyond x86 and sparc.
But nothing you can do is going to make porting it trivial, which is
what you seemed to be asking for. Porting native-code compilers just
isn't trivial.
>
> I may soon embark on a large development project on an Alpha server
> cluster. I need threading, 64 bit support, unicode, and performance. I'm
> out of luck. I have to wonder if we were all working on one implementation
> if I'd be able to use a free lisp for this project.
Well, you're out of luck if you want a free Lisp, yes. There are
commercial systems which have all these features on Alpha (assuming
you want single-processor threading only - no extant Lisps do MP
threading that I know of).
--tim
On 21 May 2001 18:23:51 +0100, Tim Bradshaw <···@tfeb.org> wrote:
>·····@puzl.pixar.com (Miles Egan) writes:
>> I know. I'm all too painfully aware of this, in fact and this is really
>> the crux of my argument. CLISP is highly portable, supports unicode and
>> runs on 64 bit machines, but is relatively slow, has no native compiler and
>> doesn't support threads. CMUCL has a superb compiler and thread support,
>> but doesn't (AFAIK) support unicode, 64 bit platforms, or really many
>> platforms beyond x86 and sparc.
>
>But nothing you can do is going to make porting it trivial, which is
>what you seemed to be asking for. Porting native-code compilers just
>isn't trivial.
Yes again. This is my dilemma.
>> I may soon embark on a large development project on an Alpha server
>> cluster. I need threading, 64 bit support, unicode, and performance. I'm
>> out of luck. I have to wonder if we were all working on one implementation
>> if I'd be able to use a free lisp for this project.
>
>Well, you're out of luck if you want a free Lisp, yes. There are
>commercial systems which have all these features on Alpha (assuming
>you want single-processor threading only - no extant Lisps do MP
>threading that I know of).
A commercial lisp is certainly not out of the question, at least as far as I'm
concerned. It's going to be a hard sell to management, though.
--
miles egan
Miles Egan wrote:
> On 21 May 2001 18:23:51 +0100, Tim Bradshaw <···@tfeb.org> wrote:
>>·····@puzl.pixar.com (Miles Egan) writes:
> A commercial lisp is certainly not out of the question, at least as far as
> I'm
> concerned. It's going to be a hard sell to management, though.
If the Project is so large, then the cost of a commercial Lisp is only a
fraction of your monthly payment and will certainly last several months.
It may help to show the management the numbers (and that the Lispsystem
will be a small fraction of the total development cost).
Regards,
Jochen
·····@puzl.pixar.com (Miles Egan) writes:
> On Sun, 20 May 2001 19:28:17 GMT, Kent M Pitman <······@world.std.com> wrote:
> >> ** comments about the lack of standardization in CL deleted ***
> >
> >While there's some truth to what you say here, I'm not convinced that
> >most of these need to be available through the implementation. I
> >think your conclusions here would come out at least a bit different if
> >you would separate things that "must come from the implementor" from
> >things that "could come from anywhere".
>
> I agree and I think I wasn't clear enough on this point. I don't think
> it's necessary that CMUCL come with it's own MySQL bindings, for example.
> I do think it's essential that such bindings exist and that they easy to
> download, install and use. I think Perl and Python provide a reasonable
> model here.
Well, they exist, and are hopefully not that difficult to download,
install and use. And if they are, just bug the author(s), and I'm
sure something can be done about it. BTW contrary to most
run-of-the-mill open-source stuff, they are even fairly-well
documented, even if I say so myself. OTOH I've been neglecting my
maintainer duties for too long, and there has been no release for much
too long a time.
The thing that's IMHO missing vis-a-vis Python and Perl is mostly only
the CPAN-equivalent, i.e. some well-though-out packaging guidelines
and some form of central directory or repository. People are working
on this though, IIUC.
I actually think that providing this is much more important than
"standardising" on one implementation by decree/consent (which won't
happen).
> Perhaps in the commercial lisps, but the available free lisps are missing
> things I've needed in order to do the work I do.
Then maybe obtaining a license for one of the commercial
implementations is the way to go for you, might it not?
> It may be a personal taste, but it's also certainly a common taste and I
> don't think it can be dismissed so simply. A lot of engineers I know
> prefer to keep the number of different languages used in a single project
> to a minimum for simplicity's sake.
Well, but there's often a natural tension between simplicity and
adequacy. If using CL and X in a project will allow me to produce
(and maintain) better software with less resources, then citing
"simplicity" is not going to cut it. If OTOH CL+X results in a more
complex, brittle and harder to maintain software, then either going
all out in CL, X or some other language might be the wisest course of
action.
> Here I disagree. I see open-source as Lisp's greatest hope because of two
> inexorable industry trends. One, giants like IBM and Sun and Microsoft are
> dumping their development tools for free or at a loss to further their
> larger agendas and backing them up with colossal marketing budgets. Two,
> there are now several mature and full-featured opensource development
> languages. Eking out a living between these two extremes is going to get
> harder and harder, I think.
Actually I don't see how open-source is going to magically overcome
these circumstances. It seems to me that Lisp is either able to
corner a niche of the market, or it isn't. If it isn't, then
open-source development isn't going to help either, because we'll be
bleeding both developers and users either way. And if it is, then it
seems to me that this niche should be able to support a couple of
commercial implementors.
> I don't think anyone is going to be able to make a living selling the kind
> of basic plumbing code I'm talking about. I think Lisp people are going to
> make money by using a shared, mature and full-featured Lisp to write
> reliable, sophisticated complex applications. All Lisp programmers benefit
> if we all pool our resources to develop a common platform and toolkit.
> Writing this kind of stuff isn't particularly enjoyable, but it's
> absolutely essential for getting any work done. If we do it together, we
> do it once, make sure it's done well, and then start making money by
> solving more interesting problems.
Well, the same is true of not-for-free software: You get someone to
do it once, and then you can use it for solving more interesting
problems. The difference is only in how you pay for the stuff: In a
"we are all developers" open-source scenario, you pay by providing free
work on the software, whereas in the "commercial tool maker" scenario,
you pay in money. Same difference. The problem for open-source
software is that - since the circle of potential users isn't
restricted - you get many "free riders", which don't pay anything, yet
are able to use the software anyhow. Once those enter into direct
competition with any of the maintainers, they are going to win, since
they can undercut the maintainers, and hence drive them out of
business in their core markets, letting the whole development model
collapse. Of course there are certain social and technical issues
which counteract this effect, and it's still to early to tell, but
IMHO those voicing concerns on the viability of the open-source model
are not to be dismissed out of hand. And to the extent that
open-source software is driving quality tool makers out of business,
I'm getting quite nervous about this.
> I know, but I think it does make it easier. If you're careful, porting
> shouldn't require much more than a recompile.
In theory. In practice, either you get very poor performance, very
poor integration, and few features, or you get easy portability. The
truth of the matter is that porting Perl, Emacs, etc. to a new machine
isn't exactly dead easy, either. It's just that normal users don't
notice this, because someone else has already done it, so that
"reporting" is just a recompile away. Look at gcc for a program
that's very difficult to port to a new machine (so difficult, that
most people normally just pay the people at Cygnus to do it for them),
yet this doesn't present a problem for the normal user, because
someone has already done it for all non-new platforms.
> Like a lot of free software hackers, I'm motivated by the reward of
> challenging my mind and creating new things and by the reward of seeing my
> software used by others and extended by others into something I could never
> have built on my own. Perusing some of the more actively developed free
> extension packages I find a morass of implementation-specific code,
> duplicated ffi code, and incomplete or nonexistent documentation. As a
> prospective free Lisp contributor, this suggests to me that I'm going to
> have to waste a lot of time on implementation-specific tweaks and that the
> other packages developers are going to need to make anything useful out of
> mine are too hard to use so nobody's going to bother with mine either.
> This undermines a lot of my enthusiasm.
Well, that depends entirely on the stuff you are going to develop. I
don't think that MaiSQL is a morass of implementation-specific code,
and it has been ported by at least one person to ACL (IIRC).
Rewriting the FFI declarations for another platform is really not that
hard, and not that much work either. And if you use a parser to
generate FFI declarations, it's even simpler (though I personally
prefer writing CMU CL FFI declarations by hand).
And for other packages, the implementation-specific part is very
small indeed. Our in-house HTTP/1.0+ server toolkit is fairly
full-featured, weighs in at over 5 KLOC, and works out of the box
on CMU CL (both x86/MP and all platforms with serve-event), ACL and
LispWorks Windows/Linux. And the implementation specific code for
each platform is between 40 and 60 LOC. So there.
> My inclination at this point is to just proceed as if CMUCL is the only
> free Lisp that really matters and start doing what I can to help it grow.
Nobody is going to stop you. It's just that when you call on other
people to follow you in this that you might get into problems. Might
I suggest though that a slight bit of code-structuring is going to go
a long way toward keeping the core of your code portable.
Actually I think all this portability-is-a-problem stuff is a red
herring. Without great effort on our part, we've actually always had
more portability in our code -- which includes a number of very
non-trivial frameworks and applications besides such simple stuff as
HTTP serving, etc. -- than we ever really needed.
> I don't want to dwell too much on this point, because it's really
> peripheral to the main thrust of my argument. I do think we should have a
> standard threading api, and a standard networking api. I'm really not
> sophisticated enough in Lisp to make any specific criticisms of the
> standard. What I'm really advocating is a willingness to extend it.
There is a huge difference between extending upon a standard, and
changing a standard, and it was mostly this that people were
vehemently objecting too, and IMHO rightly so.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
On 22 May 2001 01:00:01 +0200, Pierre R. Mai <····@acm.org> wrote:
>> I agree and I think I wasn't clear enough on this point. I don't think
>> it's necessary that CMUCL come with it's own MySQL bindings, for example.
>> I do think it's essential that such bindings exist and that they easy to
>> download, install and use. I think Perl and Python provide a reasonable
>> model here.
>
>Well, they exist, and are hopefully not that difficult to download,
>install and use. And if they are, just bug the author(s), and I'm
>sure something can be done about it.
Individually these kinds of packages may not be that hard to get working,
but plugging them together and getting them running is much more difficult
than it needs to be.
>> Perhaps in the commercial lisps, but the available free lisps are missing
>> things I've needed in order to do the work I do.
>
>Then maybe obtaining a license for one of the commercial
>implementations is the way to go for you, might it not?
Sure. It's certainly an option for the moment. I'm more concerned with
the medium to long-term growth and availability of implementations.
>Actually I don't see how open-source is going to magically overcome
>these circumstances. It seems to me that Lisp is either able to
>corner a niche of the market, or it isn't. If it isn't, then
>open-source development isn't going to help either, because we'll be
>bleeding both developers and users either way. And if it is, then it
>seems to me that this niche should be able to support a couple of
>commercial implementors.
I'm not suggesting open-source magically solves *any* problems. I think it
does give us the *ability* to neutralize some of the otherwise overwhelming
advantages the Java language goliaths wield against us. I'd like nothing
better than to see the commercial Lisps flourish, but, from my vantage
point, Lisp is slowly losing its advantage to other languages and the
comparatively tiny Lisp vendors aren't going to be able to turn the tide.
>Well, the same is true of not-for-free software: You get someone to
>do it once, and then you can use it for solving more interesting
>problems. The difference is only in how you pay for the stuff: In a
>"we are all developers" open-source scenario, you pay by providing free
>work on the software, whereas in the "commercial tool maker" scenario,
>you pay in money.
I think this argument neglects a crucial demographic - the new programmers.
Where will tomorrow's customers of these tools come from? In five years of
professional programming I've basically encountered two kinds of junior
programmers: the Visual Basic monkeys who are in it for the money and don't
know the difference between a compiler and a widget set, and the collegiate
Unix hackers that care about good tools and environments but are used to
getting them for free.
The first group isn't worth worrying about, but it's the second group that
will grow into the architects and implementors of tomorrow's big systems
and these people are used to having free, high-quality tools. They're used
to the kind of facile interoperability these tools provide. They've come
to expect easy access to all these nuts and bolts components and they're
very impatient with the legal and technical limitations of commercial
tools. They've come to expect the kind of freedom and flexibility that
full source access provides. As lamentable as this mindset may or may not
be, it's also endemic and anyone that proposes that they're going to pay
thousands of dollars per seat for a *LISP* compiler, of all things, is
going to be dismissed as a museum piece.
>collapse. Of course there are certain social and technical issues
>which counteract this effect, and it's still to early to tell, but
>IMHO those voicing concerns on the viability of the open-source model
>are not to be dismissed out of hand. And to the extent that
>open-source software is driving quality tool makers out of business,
>I'm getting quite nervous about this.
I appreciate the concerns you and others have expressed about open-source
and I think they're reasonable. I have no desire to drive anyone out of
business. My concern is that there won't be any business left to drive
them out of. I think the real threats to the Lisp vendors' business are
Java and Python and I have to wonder if the time has come to stop trying
use this all this expertise and skill to make money selling proprietary
tools and to start using it to make money selling applications. I think
we're all in agreement that programmers deserve to be paid for their work.
I think we're in disagreement about how they're most likey to achieve that
end in a sustainable way.
>In theory. In practice, either you get very poor performance, very
>poor integration, and few features, or you get easy portability. The
>truth of the matter is that porting Perl, Emacs, etc. to a new machine
>isn't exactly dead easy, either.
Sure, but since the first thing to port to the new architecture is usually
a c compiler, why not leverage it as much as possible? Exotic
architectures will always be difficult but exotic architectures are rare
these days. The CLISP folks have been casually discussing an IPaq port on
the CLISP list and so far it sounds like it's been pretty easy. I don't
imagine CMUCL will follow anytime soon.
>Actually I think all this portability-is-a-problem stuff is a red
>herring. Without great effort on our part, we've actually always had
>more portability in our code -- which includes a number of very
>non-trivial frameworks and applications besides such simple stuff as
>HTTP serving, etc. -- than we ever really needed.
Since you're certainly in a better position to make this judgement than I
am, I'm glad to hear you say so. I'm looking forward to the revision of my
first impressions. I'm sure I've probably made an overly severe impression
of the problem by looking at things like ILISP and defsystem that are bound
to be fairly worst-case.
>There is a huge difference between extending upon a standard, and
>changing a standard, and it was mostly this that people were
>vehemently objecting too, and IMHO rightly so.
Fair enough. I don't think I was very clear on this point. It's really
extensions I'm talking about. My whole goal here is to try to find a way to
minimize the amount of duplicated effort, so throwing out something as hard-won
as the CL standard isn't going to be a very good start.
--
miles egan
unix system administrator
·····@pixar.com
·····@puzl.pixar.com (Miles Egan) writes:
> I think this argument neglects a crucial demographic - the new programmers.
> Where will tomorrow's customers of these tools come from? In five years of
> professional programming I've basically encountered two kinds of junior
> programmers: the Visual Basic monkeys who are in it for the money and don't
> know the difference between a compiler and a widget set, and the collegiate
> Unix hackers that care about good tools and environments but are used to
> getting them for free.
To quote another old Lisp hacker who's done fairly well for himself:
"Linux is only free if your time has no value.".
> The first group isn't worth worrying about, but it's the second group that
> will grow into the architects and implementors of tomorrow's big systems
> and these people are used to having free, high-quality tools. They're used
> to the kind of facile interoperability these tools provide. They've come
> to expect easy access to all these nuts and bolts components and they're
> very impatient with the legal and technical limitations of commercial
> tools. They've come to expect the kind of freedom and flexibility that
> full source access provides. As lamentable as this mindset may or may not
> be, it's also endemic and anyone that proposes that they're going to pay
> thousands of dollars per seat for a *LISP* compiler, of all things, is
> going to be dismissed as a museum piece.
Do you have much experience in corporate development? I've seen
companies pay ten thousands of dollars per seat for crummy UML tools
that were buggy as hell. If you take into account what it costs a
company to employ an experienced programmer, the cost of a CL compiler
license won't cover more than a few weeks of that programmers cost. If
we can make a case to management that a programmer is far more
productive with CL than with <insert competitor of choice> the cost of
a license seat is peanuts.
> >In theory. In practice, either you get very poor performance, very
> >poor integration, and few features, or you get easy portability. The
> >truth of the matter is that porting Perl, Emacs, etc. to a new machine
> >isn't exactly dead easy, either.
>
> Sure, but since the first thing to port to the new architecture is usually
> a c compiler, why not leverage it as much as possible? Exotic
> architectures will always be difficult but exotic architectures are rare
> these days. The CLISP folks have been casually discussing an IPaq port on
> the CLISP list and so far it sounds like it's been pretty easy. I don't
> imagine CMUCL will follow anytime soon.
Compare the performance of CLISP with CMUCL.
> >Actually I think all this portability-is-a-problem stuff is a red
> >herring. Without great effort on our part, we've actually always had
> >more portability in our code -- which includes a number of very
> >non-trivial frameworks and applications besides such simple stuff as
> >HTTP serving, etc. -- than we ever really needed.
>
> Since you're certainly in a better position to make this judgement than I
> am, I'm glad to hear you say so. I'm looking forward to the revision of my
> first impressions. I'm sure I've probably made an overly severe impression
> of the problem by looking at things like ILISP and defsystem that are bound
> to be fairly worst-case.
Some of these tools are trying to remain compatible with a lot of
systems. I've once removed everything that conditionalised on systems
I didn't recognize and only kept ACL/CMUCL/LW and the remaining code
was fairly clean. A lot of old CL code still deals with things like
the TI-explorer, CLUE (whatever that may be, the comment said not to
ask) and other systems long left in the dust of history.
--
Lieven Marchand <···@wyrd.be>
Gla�r ok reifr skyli gumna hverr, unz sinn b��r bana.
On 22 May 2001 19:52:10 +0200, Lieven Marchand <···@wyrd.be> wrote:
>To quote another old Lisp hacker who's done fairly well for himself:
>"Linux is only free if your time has no value.".
This statement was made a long time ago, when getting a linux system
running was a big PITA. These days, Redhat is far easier to install than
any commercial unix. Easier than Windows too, as far as I'm concerned.
It's also interesting to note that JWZ continues to hack on free software,
so I have to assume he thinks the time he spends on it worth it. I know I
do.
>Do you have much experience in corporate development? I've seen
>companies pay ten thousands of dollars per seat for crummy UML tools
>that were buggy as hell. If you take into account what it costs a
>company to employ an experienced programmer, the cost of a CL compiler
>license won't cover more than a few weeks of that programmers cost.
I've been doing corporate software development for five years. I've used
Rational's crappy $15K tools and equally expensive RDBMS tools. The free
software tools I've used have been, for the most part, far more useful and
flexible.
But really, I'm not trying to argue that commercial Lisps aren't worth what
they cost. I'll happily defer to others' experience and assume they are.
If I were making purchasing decisions and I had some good Lisp programmers
on my staff I personally wouldn't hesitate to drop a few grand per seat for
a truly high-quality tool. My point is that, for strictly non-technical
reasons, people like me are increasingly in the minority and it's going to
get harder and harder to charge a reasonable price for any compiler.
>we can make a case to management that a programmer is far more
>productive with CL than with <insert competitor of choice> the cost of
>a license seat is peanuts.
Not in my experience. People roll their eyes when I even mention Lisp.
Call it ignorance but it's also ubiquitous. I think it's even justifiable
to some degree. The average grunt programmer, who is the person that really
counts in most of these decisions, is about as productive with a good Lisp
compiler as he is with an abacus. He doesn't know Lisp and has all the
common biases but knows Perl and Python inside out because he's been using
it for six years on Linux boxes at school and work. Guess which tools get
picked?
>> Sure, but since the first thing to port to the new architecture is usually
>> a c compiler, why not leverage it as much as possible? Exotic
>> architectures will always be difficult but exotic architectures are rare
>> these days. The CLISP folks have been casually discussing an IPaq port on
>> the CLISP list and so far it sounds like it's been pretty easy. I don't
>> imagine CMUCL will follow anytime soon.
>
>Compare the performance of CLISP with CMUCL.
Or watch Ocaml, which does exactly what I'm advocating, walk all over just
about every other HLL compiler out there, including CMUCL in many cases.
I really don't feel like I'm making myself very well understood. I'm not
trying to claim that open-source is morally or techically superior in any
absolute sense at all. It's a *strategic* choice and I think it's the
right choice for any language that can't depend on a billion-dollar life
support system. Some agree and others don't but all these absolutist
philosophical discussions about the menace or manna of open-source are
beside the point. Right now people are free to sell or give away their
software how they choose and thankfully there are no laws that prevent
either. Time will make the strengths and weaknesses of each approach
abundantly clear.
--
miles egan
"Miles Egan" <·····@puzl.pixar.com> wrote in message
·························@puzl.pixar.com...
[...]
> I really don't feel like I'm making myself very well understood. I'm not
> trying to claim that open-source is morally or techically superior in any
> absolute sense at all. It's a *strategic* choice and I think it's the
> right choice for any language that can't depend on a billion-dollar life
> support system. [...]
Your most compelling areguments (IMHO) are being overlooked.
Unless future generations of programmers are encouraged to use Lisp, there
may be a rapidly shrinking market for commercial Lisp vendors in the near
future.
Future generations of programmers will be far more likely to use Lisp if it
plays the role that C once played. As David Thornley mentioned the other
day, C allowed generations of Unix hackers in universities to do (just
about) anything with their machine. They grew up with it, they knew they
could make it bend to their will, they did imaginative things with it, and
they went on to use it successfully in the commercial arena.
The presence of an open source (and free) Lisp implementation that allows
kids to easily write their Go, Yahtzee, MP3 players, IRC clients, networked
games, 3D graphics, etc, need not undercut the profits of commercial Lisp
vendors. Far from it! Stimulating interest in Lisp among new generations of
programmers is (IMO) the best thing that could happen to them.
If Clisp or CMUCL could handle these modern-college-kid projects as well as
C and C++ (with their free GUI toolkits and ability to exploit OS services),
university students are not so stupid that they wouldn't prefer to do it in
a higher level language that's easier, quicker, much more fun and allows
much more scope for pushing their skills to higher levels in the long run.
The popularity of Python more or less proves that people are not hopelessly
tied to trends set decades ago.
I think the importance of access to modern defacto standard GUI toolkit
(like GTK+ or Qt) is being vastly underestimated here. Something as simple
as this can make the difference between a neophyte becoming a long term
addict, or choosing something else for his pet project and getting hooked on
this 'something else'.
I've read in this discussion that some feel "Sun is giving away Java to
sell Severs". If this is the case why did they create a new language?
Couldn't they be giving away their own Lisp and be ahead of the game? IBM
has been mentioned with regard to funding software to sell hardware and
services, why don't they fund a Lisp? I know some people at IBM use Lisp in
their Research, but I don't think they've had a Lisp product for many years?
Why are these two not selling or funding Lisp? Is it because their market
research say's "No Market for Lisp" i.e. it's not going to sell hardware,
help make them money?
"John Flynn" <··········@yahoo.com.au> wrote in message
··························@ozemail.com.au...
>
> "Miles Egan" <·····@puzl.pixar.com> wrote in message
> ·························@puzl.pixar.com...
> [...]
> > I really don't feel like I'm making myself very well understood. I'm
not
> > trying to claim that open-source is morally or techically superior in
any
> > absolute sense at all. It's a *strategic* choice and I think it's the
> > right choice for any language that can't depend on a billion-dollar life
> > support system. [...]
>
> Your most compelling areguments (IMHO) are being overlooked.
>
> Unless future generations of programmers are encouraged to use Lisp, there
> may be a rapidly shrinking market for commercial Lisp vendors in the near
> future.
>
> Future generations of programmers will be far more likely to use Lisp if
it
> plays the role that C once played. As David Thornley mentioned the other
> day, C allowed generations of Unix hackers in universities to do (just
> about) anything with their machine. They grew up with it, they knew they
> could make it bend to their will, they did imaginative things with it, and
> they went on to use it successfully in the commercial arena.
>
> The presence of an open source (and free) Lisp implementation that allows
> kids to easily write their Go, Yahtzee, MP3 players, IRC clients,
networked
> games, 3D graphics, etc, need not undercut the profits of commercial Lisp
> vendors. Far from it! Stimulating interest in Lisp among new generations
of
> programmers is (IMO) the best thing that could happen to them.
>
> If Clisp or CMUCL could handle these modern-college-kid projects as well
as
> C and C++ (with their free GUI toolkits and ability to exploit OS
services),
> university students are not so stupid that they wouldn't prefer to do it
in
> a higher level language that's easier, quicker, much more fun and allows
> much more scope for pushing their skills to higher levels in the long run.
> The popularity of Python more or less proves that people are not
hopelessly
> tied to trends set decades ago.
>
> I think the importance of access to modern defacto standard GUI toolkit
> (like GTK+ or Qt) is being vastly underestimated here. Something as simple
> as this can make the difference between a neophyte becoming a long term
> addict, or choosing something else for his pet project and getting hooked
on
> this 'something else'.
>
>
>
Steven L. Collins <·······@ix.netcom.com> wrote:
> I've read in this discussion that some feel "Sun is giving away Java to
> sell Severs". If this is the case why did they create a new language?
> Couldn't they be giving away their own Lisp and be ahead of the game? IBM
> has been mentioned with regard to funding software to sell hardware and
> services, why don't they fund a Lisp? I know some people at IBM use Lisp in
> their Research, but I don't think they've had a Lisp product for many years?
> Why are these two not selling or funding Lisp? Is it because their market
> research say's "No Market for Lisp" i.e. it's not going to sell hardware,
> help make them money?
Maybe it's "don't give away the gold"?
Ciao,
Oliver
> "John Flynn" <··········@yahoo.com.au> wrote in message
> ··························@ozemail.com.au...
>>
>> "Miles Egan" <·····@puzl.pixar.com> wrote in message
>> ·························@puzl.pixar.com...
>> [...]
>> > I really don't feel like I'm making myself very well understood. I'm
> not
>> > trying to claim that open-source is morally or techically superior in
> any
>> > absolute sense at all. It's a *strategic* choice and I think it's the
>> > right choice for any language that can't depend on a billion-dollar life
>> > support system. [...]
>>
>> Your most compelling areguments (IMHO) are being overlooked.
>>
>> Unless future generations of programmers are encouraged to use Lisp, there
>> may be a rapidly shrinking market for commercial Lisp vendors in the near
>> future.
>>
>> Future generations of programmers will be far more likely to use Lisp if
> it
>> plays the role that C once played. As David Thornley mentioned the other
>> day, C allowed generations of Unix hackers in universities to do (just
>> about) anything with their machine. They grew up with it, they knew they
>> could make it bend to their will, they did imaginative things with it, and
>> they went on to use it successfully in the commercial arena.
>>
>> The presence of an open source (and free) Lisp implementation that allows
>> kids to easily write their Go, Yahtzee, MP3 players, IRC clients,
> networked
>> games, 3D graphics, etc, need not undercut the profits of commercial Lisp
>> vendors. Far from it! Stimulating interest in Lisp among new generations
> of
>> programmers is (IMO) the best thing that could happen to them.
>>
>> If Clisp or CMUCL could handle these modern-college-kid projects as well
> as
>> C and C++ (with their free GUI toolkits and ability to exploit OS
> services),
>> university students are not so stupid that they wouldn't prefer to do it
> in
>> a higher level language that's easier, quicker, much more fun and allows
>> much more scope for pushing their skills to higher levels in the long run.
>> The popularity of Python more or less proves that people are not
> hopelessly
>> tied to trends set decades ago.
>>
>> I think the importance of access to modern defacto standard GUI toolkit
>> (like GTK+ or Qt) is being vastly underestimated here. Something as simple
>> as this can make the difference between a neophyte becoming a long term
>> addict, or choosing something else for his pet project and getting hooked
> on
>> this 'something else'.
>>
>>
>>
On Wed, 23 May 2001 12:29:23 +1000, John Flynn <··········@yahoo.com.au> wrote:
>
>"Miles Egan" <·····@puzl.pixar.com> wrote in message
>·························@puzl.pixar.com...
>[...]
>> I really don't feel like I'm making myself very well understood. I'm not
>> trying to claim that open-source is morally or techically superior in any
>> absolute sense at all. It's a *strategic* choice and I think it's the
>> right choice for any language that can't depend on a billion-dollar life
>> support system. [...]
>
>Your most compelling areguments (IMHO) are being overlooked.
>
>Unless future generations of programmers are encouraged to use Lisp, there
>may be a rapidly shrinking market for commercial Lisp vendors in the near
>future.
Right. Free Lisps and commercial Lisps need not necessarily be
antagonistic. Easier-to-use and more complete free Lisps could be very
powerful teasers for the high-end commercial Lisps. I know that the
commercial Lisps offer free versions, but they don't offer anything close
to the easy hackability that Perl or Python do if only because they don't
plug into all the other tools the typical larval hacker is going to want to
use.
--
miles egan
>>>>> "John" == John Flynn <··········@yahoo.com.au> writes:
John> I think the importance of access to modern defacto standard GUI toolkit
John> (like GTK+ or Qt) is being vastly underestimated here. Something as simple
John> as this can make the difference between a neophyte becoming a long term
John> addict, or choosing something else for his pet project and getting hooked on
John> this 'something else'.
I've always felt that this is the one and only reason tcl is where it
is today: It had tk which made it easy for people to do GUI stuff.
If it didn't it would have gone by the wayside.
Ray
·····@puzl.pixar.com (Miles Egan) writes:
> These days, Redhat is far easier to install than any commercial
> unix.
I've recently installed both the latest RedHat and the latest
Solaris. They're about the same difficulty to install.
> It's also interesting to note that JWZ continues to hack on free software,
> so I have to assume he thinks the time he spends on it worth it.
But he has gotten the time to do it by helping to write a proprietary
program, proving Kent's point.
> >Do you have much experience in corporate development? I've seen
> >companies pay ten thousands of dollars per seat for crummy UML tools
> >that were buggy as hell. If you take into account what it costs a
> >company to employ an experienced programmer, the cost of a CL compiler
> >license won't cover more than a few weeks of that programmers cost.
>
> I've been doing corporate software development for five years. I've used
> Rational's crappy $15K tools and equally expensive RDBMS tools.
And yet you seem to place CL's largest problem in the cost of the
commercial offerings.
> But really, I'm not trying to argue that commercial Lisps aren't worth what
> they cost. I'll happily defer to others' experience and assume they are.
> If I were making purchasing decisions and I had some good Lisp programmers
> on my staff I personally wouldn't hesitate to drop a few grand per seat for
> a truly high-quality tool. My point is that, for strictly non-technical
> reasons, people like me are increasingly in the minority and it's going to
> get harder and harder to charge a reasonable price for any compiler.
Microsoft, Sun and a bunch of others aren't doing to bad.
> >we can make a case to management that a programmer is far more
> >productive with CL than with <insert competitor of choice> the cost of
> >a license seat is peanuts.
>
> Not in my experience. People roll their eyes when I even mention Lisp.
> Call it ignorance but it's also ubiquitous. I think it's even justifiable
> to some degree. The average grunt programmer, who is the person that really
> counts in most of these decisions, is about as productive with a good Lisp
> compiler as he is with an abacus.
There I disagree. Lisp is an amplifier, so a good programmer will be
far more productive with Lisp than with Perl, but even for the average
programmer, Lisp is a better choice than C or Java.
> He doesn't know Lisp and has all the common biases but knows Perl
> and Python inside out because he's been using it for six years on
> Linux boxes at school and work. Guess which tools get picked?
Do you know of any major project that uses Perl or Python? And the
average grunt programmer doesn't understand Perl but can do a bit of
cargo cult programming at best. With this argument Clearcase wouldn't
get used because everyone's been using CVS and Solaris wouldn't get
picked for a production Enterprise 10K because everyone's running
Linux. My experience is different. At that level, cost isn't an issue
and managers like paying for accountability.
> Or watch Ocaml, which does exactly what I'm advocating, walk all over just
> about every other HLL compiler out there, including CMUCL in many cases.
I've used Ocaml. It reminds me fondly of those days in the early 90's
when I was resident C++ standards guru on a project and could work out
the changes that came with every revision of the compiler as it
aligned itself with yet another standard draft. It gave me a firm
appreciation of a standardized language with a thorough description as
a contract between multiple implementors and clients.
> It's a *strategic* choice and I think it's the right choice for any
> language that can't depend on a billion-dollar life support system.
Writing an optimizing CL implementation together with all the tools
that modern corporate development needs is a large and complex
undertaking. You are free to surprise me with your efforts on CMUCL or
any other implementation of your choice but IMHO are the commercial CL
vendors absolutely vital for CL and I regard baseless attacks on them
- which is not the same as constructive criticism - from the community
as irresponsable.
--
Lieven Marchand <···@wyrd.be>
Making laws appears to win votes. Enforcing them doesn't.
See Rule One. Roger Burton West in the monastery.
Lieven Marchand <···@wyrd.be> writes:
> ·····@puzl.pixar.com (Miles Egan) writes:
> > I've been doing corporate software development for five years. I've used
> > Rational's crappy $15K tools and equally expensive RDBMS tools.
>
> And yet you seem to place CL's largest problem in the cost of the
> commercial offerings.
You completely misunderstand me.
> Microsoft, Sun and a bunch of others aren't doing to bad.
Yep. The billion dollar budget helps. I wonder how much money they
make from sales of software tools. I bet they don't break even.
> There I disagree. Lisp is an amplifier, so a good programmer will be
> far more productive with Lisp than with Perl, but even for the average
> programmer, Lisp is a better choice than C or Java.
Lisp is an amplifier for those who know it. It's about as useful as a
Palm V to a caveman for a programmer that doesn't know it.
> Do you know of any major project that uses Perl or Python? And the
> average grunt programmer doesn't understand Perl but can do a bit of
> cargo cult programming at best.
I'm not arguing that these tools are good for large scale development.
My point is that they are used for large scale development because
they're so nice for smaller scale development that people use them
when they're not appropriate out of familiarity. It's important for a
language to appeal to the junior programmer and facilitate small-scale
development because it's those same junior programmers that turn into
managers later and choose implementation tools later. I've seen this
happen many times now.
Lisp is, by design, optimized for more advanced programmers and larger
projects, which is why we all like it so much. That means we have to
try that much harder to compete with languages like Python that are
optimized by design for beginning programmers.
> > Or watch Ocaml, which does exactly what I'm advocating, walk all over just
> > about every other HLL compiler out there, including CMUCL in many cases.
>
> I've used Ocaml. It reminds me fondly of those days in the early 90's
> when I was resident C++ standards guru on a project...
We're really not communicating here, are we? I'm not holding up Ocaml
as the pinnacle of language or compiler design. I'm using it as a
concrete example of a two mode bytecode/native compilation strategy
that's both easy to port and fast that I think would make sense for
Lisp. That's all.
> > It's a *strategic* choice and I think it's the right choice for any
> > language that can't depend on a billion-dollar life support system.
>
> Writing an optimizing CL implementation together with all the tools
> that modern corporate development needs is a large and complex
> undertaking.
Larger and more complex than an entire operating system? I doubt it.
> You are free to surprise me with your efforts on CMUCL or
> any other implementation of your choice but IMHO are the commercial CL
> vendors absolutely vital for CL and I regard baseless attacks on them
> - which is not the same as constructive criticism - from the community
> as irresponsable.
You're either confusing me with someone else or not paying much
attention. *I* am making no attacks of any kind on any commercial
vendors and I think I've made that abundantly clear. I absolutely
support their right to sell whatever they want at whatever price they
want. My concern is that I don't think any small vendor can support a
language in today's software climate, especially something already
viewed as exotic and old.
--
miles egan
Miles Egan <·····@puzl.pixar.com> wrote in message news:<···············@puzl.pixar.com>...
> We're really not communicating here, are we? I'm not holding up Ocaml
> as the pinnacle of language or compiler design. I'm using it as a
> concrete example of a two mode bytecode/native compilation strategy
> that's both easy to port and fast that I think would make sense for
> Lisp. That's all.
Umm, such a common lisp exists, or at least it did. It was spun out of
Xerox[1] and John Sybalsky tried to make a living supporting and selling it
and delivering applications. I don't think that he succeeded, so I'm
skeptical of the idea that one can necessarily make a living supporting
a good product.
That particular implementation ran well in very little memory. As a
result, applications using it were far less likely to swap than applications
running in competitive implementations. At that point, it was
significantly faster. (If you did manage to start swapping, it was
still faster because it swapped so much less; I remember that it
used 2-3x less memory, even when it played jit games, long before
Java. For those keeping score, some of the initial Java folk joined
Sun after Lucid died.)
It was, however, slower when running when the competition had "enough"
memory. It's hard to say where it would come out now. (It would run
more from cache.)
I wonder if any of the claims for a byte-coded lisp are confirmed by
the experiences with byte-coded lisps.
-andy
[1] - At the time, Lucid was still alive, as well as today's survivors.
Hello!
In article <····························@posting.google.com>,
Andy Freeman <······@earthlink.net> wrote:
>Miles Egan <·····@puzl.pixar.com> wrote in message
>news:<···············@puzl.pixar.com>...
>> We're really not communicating here, are we? I'm not holding up Ocaml
>> as the pinnacle of language or compiler design. I'm using it as a
>> concrete example of a two mode bytecode/native compilation strategy
>> that's both easy to port and fast that I think would make sense for
>> Lisp. That's all.
>Umm, such a common lisp exists, or at least it did. It was spun out of
>Xerox[1] and John Sybalsky tried to make a living supporting and selling it
>and delivering applications. I don't think that he succeeded, so I'm
>skeptical of the idea that one can necessarily make a living supporting
>a good product.
Was that bytecode-only or like Ocaml (on nearly all platforms you have
the bytecode system, on selected platforms you have the native
code system in addition - simple bootstrapping by compiling a
C-based VM + shipped ocaml->bytecode compiler as bytecode image;
say, the advantages of e.g. Clisp [bytecode, very portable, easy
bootstrap] and CMUCL [good native code on selected platforms] combined)?
In Ocaml, you can of course choose to compile your applications to
bytecode also on native-supported platforms (e.g. if you prefer
the more compact memory footprint or the fact that you can ship
your bytecode image to other execution platforms without recompiling).
>[...]
Kind regards,
Hannah.
······@schlund.de (Hannah Schroeter) wrote in message news:<············@c3po.schlund.de>...
> >Umm, such a common lisp exists, or at least it did. It was spun out of
> >Xerox[1] and John Sybalsky tried to make a living supporting and selling it
> >and delivering applications. I don't think that he succeeded, so I'm
> >skeptical of the idea that one can necessarily make a living supporting
> >a good product.
>
> Was that bytecode-only or like Ocaml (on nearly all platforms you have
> the bytecode system, on selected platforms you have the native
> code system in addition - simple bootstrapping by compiling a
> C-based VM + shipped ocaml->bytecode compiler as bytecode image;
> say, the advantages of e.g. Clisp [bytecode, very portable, easy
> bootstrap] and CMUCL [good native code on selected platforms] combined)?
The bytecodes could be further compiled, but I don't remember if the
target was platform specific. (One would think that it would be, but
the Parc folks, including Sybalsky, were quite clever. There may have
been both platform-independent and platform-dependent targets.) I
suppose that one could have invoked the beyond-bytecode compiler "by
hand", but they put a lot of effort into making that invisible. (It
would incrementally recompile, throw out translations that weren't
"hot", and everything was always fully-debuggable.)
Of course, I'm obviously delusional, because everyone knows that the
Java folks invented 50% of this stuff in 95.
I wonder what cool technology Parc is pissing way this year?
-andy
Lieven Marchand <···@wyrd.be> wrote in
<··············@localhost.localdomain>:
>·····@puzl.pixar.com (Miles Egan) writes:
[snip]
>> He doesn't know Lisp and has all the common biases but knows Perl
>> and Python inside out because he's been using it for six years on
>> Linux boxes at school and work. Guess which tools get picked?
>
>Do you know of any major project that uses Perl or Python? And the
Without knowing what qualifies as a "major project", I can toss out a few
factoids without anyone accusing me of having clue:
-- Google (http://www.google.com/) was (and as far as I know still is)
written almost entirely in Python with a few bits in C where performance
demanded. The Google team claims that Python was their competitive
advantage, by the way.
-- Walt Disney Imagineering is actively recruiting Python developers,
although I don't know of any projects they've completed.
-- Johns Hopkins University is recruiting developers for post-processing
information from the Hubble Space Telescope. They're a bit more catholic in
that they're willing to accept either Python or Perl developers.
That aside, I still wish I could justify writing my next project(s) in Lisp
(or Python) instead of the Java+Javascript that will almost assuredly be
mandated. In my role as a developer/manager, however, the largest metrics
of interest are "lines of code not written" and the nebulous "playing-well-
with-others quotient". I opine (with my uneducated opinion) that it will be
very difficult for any technology to catch up to Java's huge and rapidly
increasing set of libraries, both from the language vendor as well as
reputable third parties. Sure, I can do more in ten lines of raw Lisp than
I can in ten lines of raw Java, when one of those Java lines is "import
org.apache.soap.rpc.SOAPContext;" then I'm going to have to write a non-
trivial amount of Lisp to catch up again.
Or, I could be wrong, and Lisp has all that and then some. I'd still have
to spend a good bit of political capital re-educating people that Lisp
should be used instead of Java. Unfortunately, I will probably feel that
that capital might better be spent fighting the "All Rational All RUP All
The Time" meme that seems to be percolating through the culture.
without-clue-or-rancor-ly y'rs
+Mitchell
On Sun, 20 May 2001 19:28:17 GMT, Kent M Pitman <······@world.std.com>
wrote:
>
>I mention this because I think one of the problems is that the Lisp community
>as a whole is used to relying on the VENDORS for answers and they must,
>as the Java community has, start relying on the USERS to provide some stuff,
>leaning only on the vendors for what they cannot make themselves. There are
>or ought to be more users than vendors, so it's the users who are empowered
>to keep up. They must do their fair share and blame must be attributed
>in proper proportions.
>
Absolutely. Btw, I searched the Internet recently and got these two
documents. The first,
"Modernizing Common Lisp: Recommended Extensions",
July 7, 1999"
was written by people at MIT - a copy of the text is at the URL:
http://lists.tunes.org/archives/tunes/2001-February/003058.html
The second was this J13 document:
"DRAFT Minutes of the NCITS/J13 meeting on 10/14/99 at SRI"
found at
http://www.franz.com/~smh/J13/99-010.txt
>.....
>I think success, if it comes, will come from people accepting that not every
>language is the same, and that we are a community which has many bases.
>We should work to have a set of common libraries and stop fussing over
>the base. CL is done. Leave it. More resources spent on that are a commitment
>individuals can reasonably make, but not the community. Our community has
>paid once for its base and cannot afford to pay again for it, hoping for
>a magic bullet fix.
Both texts show the basic dilemma which can be described, I think, as
"the absence of an Influential CL User Group". You mention "our
community", but this community must be narrowed a bit first.
To begin, when surfing the ALU web pages, I cannot see how ALU, as a
body, can take an active part. Could it? One ALU page mentions a lot
of mailing lists, but where is a visible group?
I see other very good Lisp-related pages by dedicated people, and some
CL source code collections, but there is still a long way to go ...
So I ask these two simple questions:
- Who (which body, group, etc) can possibly put together a set of
documents to become widely acknowledged and accepted as _the_ defacto
CL extension(s)?
- If no such body exists, can one be established and if so, how?
/Lars
·············@era.ericsson.se (Lars Lundback) writes:
> I ask these two simple questions:
[These are at some level good and natural questions, and I understand your
reason for asking them, but they have been asked over and over again and
I'll try through my answer to help you see why I, at least, think these
are the wrong questions to ask. In some sense, I think it's the focus
on these questions that keeps us holding in place.]
> - Who (which body, group, etc) can possibly put together a set of
> documents to become widely acknowledged and accepted as _the_ defacto
> CL extension(s)?
I think the answer is either "no one" or "anyone". The problem is that the
necessary next step is not "documents", it is "crisis". Vendors told that
they should produce "documents" will "put it on the list". Vendors respond
to crisis and what will make them do something will be the perception that
sales will drop precipitously off, especially to another vendor (I feel as
if they don't measure drop-off to other languages the same way, and I can't
say why), if they don't take immediate action.
So IF j13 got together and somehow made a new standard, I do think the
vendors would react to it. But I don't think it's wise for J13 to do this
because its processes are so long that it would only either make another
outmoded standard (one that was good at the time the approval/review processes
began, but no longer current at the time the approval/review processes ended),
or it would make incompatible changes in the existing standard or... well,
suffice to say that I didn't drop my J13 membership casually. I think the
vehicle has had its day but I don't see how to proceed from here in that
vehicle.
I also think J13 would respond to users if they got together and operated
as a block. For example, if the ALU created a mechanism for coming up with
recommendations and people as a block had the will to say "we're only buying
Lisps that correspond to that recommendation", then the vendors would react.
In effect, this is why the previous paragraph would work, and did work when
X3.226-1994 was published. People said they only wanted ANSI CL, accept
no substitutes, and the vendors did get to it and provide that.
But honestly, I think most people are practical enough that they're
not going to turn down a perfectly good Lisp that just has a few
implementation-specific packages for getting the job done. So they're
not going to not use Franz just because its implementation of multiprocessing
is vendor-specific--what's their option? Use LispWorks, whose implementation
is also vendor-specific? That won't do much. You could sit down and write
a standard and hope the vendors will migrate toward it. And I think with
some areas you'd have a big win. DEFSYSTEM for example. Geez, there are so
many DEFSYSTEM's around and they're all incompatible and my sense is that
even the lumbering, slow-moving vendors would love to get this sorted out.
But MP? Subtle details affect efficiency, so you might have a tougher time
with that unless you carefully involve the vendors.
It could be done. But who's to do it? I don't know that it's waiting on
anyone in specific. Maybe any of us could just start writing standards
and then seeing if someone bought in. Maybe it needs compatibility packages
so the vendors' work to accept the standard is initially null. But then,
the ALU told the vendors about 5 or more years ago that they'd be happy with
the start of just adding CLIM-SYS to the distributed language. It *has*
such compatibility packages already implemented for the major vendors, and
it's already something that is known to support at least one well-known
biggish system across multiple implementations. But the vendors never did
offer even that as standard--who knows why. I think they just didn't see
it standing between them and sales.
So you can try to orchestrate more fear in them if you want by getting
users to operate as a block, but I think it's been amply demonstrated
that mere "this would be the right thing" reasoning does not work.
> - If no such body exists, can one be established and if so, how?
No one is stopping you. But it may do no good.
The more rational solution, I think, is to not care. I don't mean to not care
about the language or its communities, but rather not care to care about
standards and language synchronization in their
traditional form. Lisp has required a single way of doing things for a long
time, but other languages have not. The whole notion of "library structure"
is regarded by Lisp as an optimization, but underlying this is an assumption
that as a community we will all go in one direction, and that it is even more
efficient for us to do so. You might think competition among equivalent
libraries was bad, but one thing it does is to remove the need for
synchronization, which standards bodies are. They are very expensive to
synchronize, and while the ANSI CL standard provides us with a good strong
synchronized base, I think a continued desire to follow that particular
synchronization model in the future could spell even worse troubles for
our community.
I do have a theory about how to confront this directly, but I'm still
doing implementation work on that right now and I want to hold off
talking about my personal plan until I have some other details in
place. In essence, I don't want to defend my theory until I'm sure I
know all of the details of how it would work. Maybe it will help, in due
time--though there's no guarantee anyone will like it at all.
Certainly no one should assume that I will necessarily succeed in my
personal plans to address this issue. I just hope you'll be heartened to
know someone is even actively trying. I write code for and do design work
toward this project on a very regular basis.
But if others have suggestions to air on the matter, or want to try their
own approaches, they shouldn't feel like I'm saying I have the only right
to go after this. One thing about this community is that it's always
open to ideas, no matter who they're from.
Kent M Pitman <······@world.std.com> writes:
> But honestly, I think most people are practical enough that they're
> not going to turn down a perfectly good Lisp that just has a few
> implementation-specific packages for getting the job done. So they're
> not going to not use Franz just because its implementation of multiprocessing
> is vendor-specific--what's their option? Use LispWorks, whose implementation
> is also vendor-specific? That won't do much. You could sit down and write
> a standard and hope the vendors will migrate toward it. And I think with
> some areas you'd have a big win. DEFSYSTEM for example. Geez, there are so
> many DEFSYSTEM's around and they're all incompatible and my sense is that
> even the lumbering, slow-moving vendors would love to get this
> sorted out.
I think you are in error here. I have a personal stake in this, but,
as far as I can tell there is currently only one DEFSYSTEM that runs
(sputtering, but it runs) on most known CL implementations:
MK:DEFSYSTEM. It runs under ACL, LW, CMUCL, CLisp, MCL, etc etc.
I have not done it, but, as an example, the people porting AServe to
LW would greatly benefit from having a single DEFSYSTEM specifiction.
They have the tool to do so. The tool is MK:DEFSYSTEM.
Now the question is (since I gather that you don't): why don't you use
MK:DEFSYSTEM? It is more than a curiosity.
> It could be done. But who's to do it? I don't know that it's waiting on
> anyone in specific. Maybe any of us could just start writing standards
> and then seeing if someone bought in. Maybe it needs compatibility packages
> so the vendors' work to accept the standard is initially null. But then,
> the ALU told the vendors about 5 or more years ago that they'd be happy with
> the start of just adding CLIM-SYS to the distributed language. It *has*
> such compatibility packages already implemented for the major vendors, and
> it's already something that is known to support at least one well-known
> biggish system across multiple implementations. But the vendors never did
> offer even that as standard--who knows why. I think they just didn't see
> it standing between them and sales.
As a matter of fact it *did* stand between them and sales. Adopting a
standard (CLIM-SYS in this case) would have forced them to compete on
a different level, while having users locked in with Common Graphics
and CAPI.
This was a perfectly sane business decision (taken rationally or not,
maybe not even historically begotten that way) on the vendors' part.
It is a decision that in the long term has hurt Common Lisp and the
users, but nobody is claiming that many business decisions have a
particular long term view of the world. After all in the long term,
we are all going to be dead (this isn't mine: it's J.M. Keynes' :) ).
...
> But if others have suggestions to air on the matter, or want to try their
> own approaches, they shouldn't feel like I'm saying I have the only right
> to go after this. One thing about this community is that it's always
> open to ideas, no matter who they're from.
As it should be.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
Kent M Pitman <······@world.std.com> writes:
> But honestly, I think most people are practical enough that they're
> not going to turn down a perfectly good Lisp that just has a few
> implementation-specific packages for getting the job done. So they're
> not going to not use Franz just because its implementation of multiprocessing
> is vendor-specific--what's their option? Use LispWorks, whose implementation
> is also vendor-specific? That won't do much. You could sit down and write
> a standard and hope the vendors will migrate toward it. And I think with
> some areas you'd have a big win. DEFSYSTEM for example. Geez, there are so
> many DEFSYSTEM's around and they're all incompatible and my sense is that
> even the lumbering, slow-moving vendors would love to get this
> sorted out.
I think you are in error here. I have a personal stake in this, but,
as far as I can tell there is currently only one DEFSYSTEM that runs
(sputtering, but it runs) on most known CL implementations:
MK:DEFSYSTEM. It runs under ACL, LW, CMUCL, CLisp, MCL, etc etc.
I have not done it, but, as an example, the people porting AServe to
LW would greatly benefit from having a single DEFSYSTEM specifiction.
They have the tool to do so. The tool is MK:DEFSYSTEM.
Now the question is (since I gather that you don't): why don't you use
MK:DEFSYSTEM? It is more than a curiosity.
> It could be done. But who's to do it? I don't know that it's waiting on
> anyone in specific. Maybe any of us could just start writing standards
> and then seeing if someone bought in. Maybe it needs compatibility packages
> so the vendors' work to accept the standard is initially null. But then,
> the ALU told the vendors about 5 or more years ago that they'd be happy with
> the start of just adding CLIM-SYS to the distributed language. It *has*
> such compatibility packages already implemented for the major vendors, and
> it's already something that is known to support at least one well-known
> biggish system across multiple implementations. But the vendors never did
> offer even that as standard--who knows why. I think they just didn't see
> it standing between them and sales.
As a matter of fact it *did* stand between them and sales. Adopting a
standard (CLIM-SYS in this case) would have forced them to compete on
a different level. Instead having users locked in with Common Graphics
and CAPI is a definite advantage for them.
This was a perfectly sane business decision (taken rationally or not,
maybe not even historically begotten that way) on the vendors' part.
It is a decision that in the long term has hurt Common Lisp and the
users, but nobody is claiming that many business decisions have a
particular long term view of the world. After all in the long term,
we are all going to be dead (this isn't mine: it's J.M. Keynes' :) ).
...
> But if others have suggestions to air on the matter, or want to try their
> own approaches, they shouldn't feel like I'm saying I have the only right
> to go after this. One thing about this community is that it's always
> open to ideas, no matter who they're from.
As it should be.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
Marco Antoniotti <·······@cs.nyu.edu> writes:
> Kent M Pitman <······@world.std.com> writes:
>
> > But honestly, I think most people are practical enough that
> > they're not going to turn down a perfectly good Lisp that just has
> > a few implementation-specific packages for getting the job done.
> > So they're not going to not use Franz just because its
> > implementation of multiprocessing is vendor-specific--what's their
> > option? Use LispWorks, whose implementation is also
> > vendor-specific? That won't do much. You could sit down and
> > write a standard and hope the vendors will migrate toward it. And
> > I think with some areas you'd have a big win. DEFSYSTEM for
> > example. Geez, there are so many DEFSYSTEM's around and they're
> > all incompatible and my sense is that even the lumbering,
> > slow-moving vendors would love to get this sorted out.
>
> I think you are in error here. I have a personal stake in this, but,
> as far as I can tell there is currently only one DEFSYSTEM that runs
> (sputtering, but it runs) on most known CL implementations:
> MK:DEFSYSTEM. It runs under ACL, LW, CMUCL, CLisp, MCL, etc etc.
>
> I have not done it, but, as an example, the people porting AServe to
> LW would greatly benefit from having a single DEFSYSTEM specifiction.
> They have the tool to do so. The tool is MK:DEFSYSTEM.
>
> Now the question is (since I gather that you don't): why don't you use
> MK:DEFSYSTEM? It is more than a curiosity.
To be clear: I did not say the vendors would love to find a DEFSYSTEM
that ran in everyone else's implementation. I said the vendors would
love to find a DEFSYSTEM they didn't get beaten up about.
You have yet to make the case (by which I don't mean you have yet to
make logical arguments, but rather I mean you have yet to cause
screaming hoards to descend, axe or spear in hand at vendor's door)
that MK:DEFSYSTEM is that implementation.
The point of my remarks were that this has everything to do with politics
and nothing to do with technical issues.
Further, if you think politics is the gentlemanly debate that you see
recorded in the CL Cleanup Issues, you're confused. That stuff is like
the record left by C-SPAN or BBC when politicians get up to pontificate.
Sure, it's the stuff people voted on, but it never captures the WHY of
people voting. People voted no on perfectly good things and yes on silly
things. Overall, we didn't do badly, but the record of debates and speeches
and logical arguments does not sum up the quality of what was in the room.
There was a genuine since of urgency about CL at that time--that we needed
a standard to build useful applications. There is not that now because,
no matter how you use the word "intolerable", it is not IN FACT intolerable
unless you refuse to tolerate it. Once you lay your money on the table and
say "I will buy this implementation anyway, in spite of its non-adherence
to what I have asked for", you have lost your leverage, and the vendors know
it. They know you probably better than you know you yourself.
It is just not a commercially problemsome thing the absence of defsystem.
The workarounds are trivial. It is an annoyance to be sure, but you have
not made a business case that it is worth a vendor's time.
> > It could be done. But who's to do it? I don't know that it's waiting on
> > anyone in specific. Maybe any of us could just start writing standards
> > and then seeing if someone bought in. Maybe it needs compatibility packages
> > so the vendors' work to accept the standard is initially null. But then,
> > the ALU told the vendors about 5 or more years ago that they'd be happy with
> > the start of just adding CLIM-SYS to the distributed language. It *has*
> > such compatibility packages already implemented for the major vendors, and
> > it's already something that is known to support at least one well-known
> > biggish system across multiple implementations. But the vendors never did
> > offer even that as standard--who knows why. I think they just didn't see
> > it standing between them and sales.
>
> As a matter of fact it *did* stand between them and sales. Adopting a
> standard (CLIM-SYS in this case) would have forced them to compete on
> a different level. Instead having users locked in with Common Graphics
> and CAPI is a definite advantage for them.
As I understand it, and it's been years since I looked so I could
easily be suffering from neuron misfirings, CLIM-SYS does not have the
grpahics component. CLIM-SYS is the life support component that has
DEFSYSTEM and MP and a few things like that in it.
The layer you're thinking about either doesn't exist or is higher up.
Kent M Pitman <······@world.std.com> writes:
>
> As I understand it, and it's been years since I looked so I could
> easily be suffering from neuron misfirings, CLIM-SYS does not have the
> grpahics component. CLIM-SYS is the life support component that has
> DEFSYSTEM and MP and a few things like that in it.
>
Yes, that's correct. I don't think CLIM has a coherent, extractable,
graphics layer.
--tim
Kent M Pitman <······@world.std.com> writes:
> Marco Antoniotti <·······@cs.nyu.edu> writes:
>
> > Kent M Pitman <······@world.std.com> writes:
> >
> > > But honestly, I think most people are practical enough that
...
> > > example. Geez, there are so many DEFSYSTEM's around and they're
> > > all incompatible and my sense is that even the lumbering,
> > > slow-moving vendors would love to get this sorted out.
> >
> > I think you are in error here. I have a personal stake in this, but,
> > as far as I can tell there is currently only one DEFSYSTEM that runs
> > (sputtering, but it runs) on most known CL implementations:
> > MK:DEFSYSTEM. It runs under ACL, LW, CMUCL, CLisp, MCL, etc etc.
> >
> > I have not done it, but, as an example, the people porting AServe to
> > LW would greatly benefit from having a single DEFSYSTEM specifiction.
> > They have the tool to do so. The tool is MK:DEFSYSTEM.
> >
> > Now the question is (since I gather that you don't): why don't you use
> > MK:DEFSYSTEM? It is more than a curiosity.
>
> To be clear: I did not say the vendors would love to find a DEFSYSTEM
> that ran in everyone else's implementation. I said the vendors would
> love to find a DEFSYSTEM they didn't get beaten up about.
>
> You have yet to make the case (by which I don't mean you have yet to
> make logical arguments, but rather I mean you have yet to cause
> screaming hoards to descend, axe or spear in hand at vendor's door)
> that MK:DEFSYSTEM is that implementation.
It runs on LW and ACL. The respective DEFSYSTEMs (LW and ACL) are
incompatible. To me this is a good enough argument. Moreover,
MK:DEFSYSTEM is a little more than an implementation. There is (not
complete by all means) document that describes most of its
functionality. In this case the choice is clear cut: if you write a
complex CL system, it is in everybody's best interest that you use
MK:DEFSYSTEM. I can tell the vendors to adopt MK:DEFYSTEM (at least
to bundle it), but my message is better directed to programmers.
Each of us has the choice to do what s/he pleases. I can merely argue
that some choices have better long terms benefit. Again I can only
argue that.
Which leaves the question hanging... I gather that you do not use
MK:DEFSYSTEM. Isn't it so?
> The point of my remarks were that this has everything to do with politics
> and nothing to do with technical issues.
I am very well aware of this. The issue about case is very clear and
I think you - personally :) - were, although in perfect good faith, on
the Dark Side of the Force then.
...
> It is just not a commercially problemsome thing the absence of defsystem.
> The workarounds are trivial. It is an annoyance to be sure, but you have
> not made a business case that it is worth a vendor's time.
I do nto think I have the arguments to make that case. It is obvious
that the vendors are not (yet) complying. My case is to the CL
programmers.
...
> > As a matter of fact it *did* stand between them and sales. Adopting a
> > standard (CLIM-SYS in this case) would have forced them to compete on
> > a different level. Instead having users locked in with Common Graphics
> > and CAPI is a definite advantage for them.
>
> As I understand it, and it's been years since I looked so I could
> easily be suffering from neuron misfirings, CLIM-SYS does not have the
> grpahics component. CLIM-SYS is the life support component that has
> DEFSYSTEM and MP and a few things like that in it.
>
> The layer you're thinking about either doesn't exist or is higher
> up.
My (partly intended :) ) mistake. I know. But the reasoning applies.
Cheers.
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
Marco Antoniotti <·······@cs.nyu.edu> writes:
> Kent M Pitman <······@world.std.com> writes:
> > To be clear: I did not say the vendors would love to find a DEFSYSTEM
> > that ran in everyone else's implementation. I said the vendors would
> > love to find a DEFSYSTEM they didn't get beaten up about.
> >
> > You have yet to make the case (by which I don't mean you have yet to
> > make logical arguments, but rather I mean you have yet to cause
> > screaming hoards to descend, axe or spear in hand at vendor's door)
> > that MK:DEFSYSTEM is that implementation.
>
> It runs on LW and ACL. The respective DEFSYSTEMs (LW and ACL) are
> incompatible. To me this is a good enough argument. Moreover,
> MK:DEFSYSTEM is a little more than an implementation. There is (not
> complete by all means) document that describes most of its
> functionality. In this case the choice is clear cut: if you write a
> complex CL system, it is in everybody's best interest that you use
> MK:DEFSYSTEM.
Marco,
Can you with certainty say that mk:defsystem provides all of the
features of the various defsystems out there?
If so, can you point me to some html I can look at?
If not, then how do you expect users of the current systems (including
us vendors) to make a switch if mk:defsystem doesn't at least have
compatibility packages to aid the transition?
> > It is just not a commercially problemsome thing the absence of defsystem.
> > The workarounds are trivial. It is an annoyance to be sure, but you have
> > not made a business case that it is worth a vendor's time.
>
> I do nto think I have the arguments to make that case. It is obvious
> that the vendors are not (yet) complying. My case is to the CL
> programmers.
What does this "compliance" mean? I've never used mk:defsystem, hence
the request for documentation, but my understanding is that Allegro's
defsystem has features that mk:defsystem doesn't have and vice-versa.
The fact that both defsystems work on Allegro CL provides a choice for
users with different needs.
If by "compliance" you really mean "exclusive use", then as the current
vendor of mk:defsystem you should make a case to us. We are committed
to de-facto standards and would be happy to eventually move to such a
defsystem _iff_ it is able to satisfy all of our needs as well as our
customers' needs. The reason why we have not done so is because of
our (possibly incorrect) understanding that neither defsystem version
encloses the other functionally. So we continue with what we have,
because it does what we need it to do.
--
Duane Rettig Franz Inc. http://www.franz.com/ (www)
1995 University Ave Suite 275 Berkeley, CA 94704
Phone: (510) 548-3600; FAX: (510) 548-8253 ·····@Franz.COM (internet)
Duane Rettig <·····@franz.com> writes:
> Marco Antoniotti <·······@cs.nyu.edu> writes:
>
...
> > It runs on LW and ACL. The respective DEFSYSTEMs (LW and ACL) are
> > incompatible. To me this is a good enough argument. Moreover,
> > MK:DEFSYSTEM is a little more than an implementation. There is (not
> > complete by all means) document that describes most of its
> > functionality. In this case the choice is clear cut: if you write a
> > complex CL system, it is in everybody's best interest that you use
> > MK:DEFSYSTEM.
>
> Marco,
>
> Can you with certainty say that mk:defsystem provides all of the
> features of the various defsystems out there?
No. It definitively provides "similar" features. It definitively does
not (yet) provide extensibility through CLOS as ACL defsystem does.
> If so, can you point me to some html I can look at?
There is no HTML, but there is a CMU TR report that describes
MK:DEFSYSTEM and other utilities as well. That is the reliable
documentation about it. It should be available in the CMU
AI.Repository. It is not part of the MK:DEFSYSTEM distribution
because of some copyright issues (it is a CMU TR).
> If not, then how do you expect users of the current systems (including
> us vendors) to make a switch if mk:defsystem doesn't at least have
> compatibility packages to aid the transition?
That is a very good point and I must say that I underestimated such a need.
> > > It is just not a commercially problemsome thing the absence of defsystem.
> > > The workarounds are trivial. It is an annoyance to be sure, but you have
> > > not made a business case that it is worth a vendor's time.
> >
> > I do nto think I have the arguments to make that case. It is obvious
> > that the vendors are not (yet) complying. My case is to the CL
> > programmers.
>
> What does this "compliance" mean?
Forgive my poor choice of words. I was merely saying that the vendors
have not shown interest in MK:DEFSYSTEM, although it has been around
for a long time. Much before I became its caretaker.
> I've never used mk:defsystem, hence
> the request for documentation, but my understanding is that Allegro's
> defsystem has features that mk:defsystem doesn't have and vice-versa.
> The fact that both defsystems work on Allegro CL provides a choice for
> users with different needs.
>
> If by "compliance" you really mean "exclusive use", then as the current
> vendor of mk:defsystem you should make a case to us. We are committed
> to de-facto standards and would be happy to eventually move to such a
> defsystem _iff_ it is able to satisfy all of our needs as well as our
> customers' needs. The reason why we have not done so is because of
> our (possibly incorrect) understanding that neither defsystem version
> encloses the other functionally. So we continue with what we have,
> because it does what we need it to do.
As I said in response to KMP, right now I cannot make that case to you
using the "functionality arguments". My only argument is - as weak as
it is - MK:DEFSYSTEM's portability. As a user, I need a DEFSYSTEM that
runs on ACL and (say) CLisp. MK:DEFSYSTEM fits the bill.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
Marco Antoniotti <·······@cs.nyu.edu> writes:
> I think you are in error here. I have a personal stake in this, but,
> as far as I can tell there is currently only one DEFSYSTEM that runs
> (sputtering, but it runs) on most known CL implementations:
> MK:DEFSYSTEM. It runs under ACL, LW, CMUCL, CLisp, MCL, etc etc.
Yes, but which vendor would want to do support on MK:DEFSYSTEM? There
is no specification[1], one implementation which is not the cleanest code
under the sky, and in fact very few people who are intimately familiar
with it. If I were a vendor that had to (i.e. was legally obliged to)
offer support for MK:DEFSYSTEM, I wouldn't be able to sleep at night.
And since no specification exists, vendors wouldn't be able to
reasonably reimplement it on their own either, so there is no way out
of this trap.
> I have not done it, but, as an example, the people porting AServe to
> LW would greatly benefit from having a single DEFSYSTEM specifiction.
> They have the tool to do so. The tool is MK:DEFSYSTEM.
You are IMHO greatly overestimating the benefit here. Translating
most system definitions from one DEFSYSTEM to another is often a work
of minutes, maybe hours at most. That doesn't translate to "greatly"
in my eyes. And that's the whole problem with the lack of layered
standards: The cost (both to current users and vendors) of not having
them currently mostly exceeds the cost creating them, for each single
item proposed.
If we are to have any realistic chance of getting to a DEFSYSTEM
standard, we will need a clear, precise specification first, and then
maybe a portable _clean_ implementation. But code without a
specification is just plain not going to convince any vendors to drop
their proven, documented and accepted DEFSYSTEM facilities (which is
really what we are after).
Personally I think that implementing something that incorporates in
some form the ideas behind Kent's DEFSYSTEM proposal is the way to
go. But ultimately unless I find the time to sit down and do
something more constructive than posting about this, I don't think
anyone is going to care what I think (and rightly so).
Regs, Pierre.
Footnotes:
[1] Actually you put the beginnings of a specification for a version
4.x on CLOCC, but since there is no corresponding implementation,
the only MK:DEFSYSTEM you can be talking about is 3.x, which as
we all know has no form of specification, no up-to-date
documentation, and is in fact only specifiable as "whatever the
code does now".
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein
"Pierre R. Mai" <····@acm.org> writes:
> Marco Antoniotti <·······@cs.nyu.edu> writes:
>
...
>
> Personally I think that implementing something that incorporates in
> some form the ideas behind Kent's DEFSYSTEM proposal is the way to
> go. But ultimately unless I find the time to sit down and do
> something more constructive than posting about this, I don't think
> anyone is going to care what I think (and rightly so).
>
> Regs, Pierre.
>
> Footnotes:
> [1] Actually you put the beginnings of a specification for a version
> 4.x on CLOCC, but since there is no corresponding implementation,
> the only MK:DEFSYSTEM you can be talking about is 3.x, which as
> we all know has no form of specification, no up-to-date
> documentation, and is in fact only specifiable as "whatever the
> code does now".
You are right all along the way. What can I say? Probably just shut
up and try to write up something coherent.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
719 Broadway 12th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
·····@puzl.pixar.com (Miles Egan) writes:
> Like many of the readers of this group, I've spent a lot of time pondering
> the popularity and evolution of lisp. I've also spent a lot of time
> playing with the various implementations of lisp out there and I've come to
> a few tentative conclusions:
<snip>
(score -1: redundant):
short:
If it was easy to use CL for small ``solved problems'' more people would
know it enough to consider it for bigger projects (from the start, instead
of despairing when the application grows big and unmanagable). Readily
available developer environment in linux dists, gui/network/db bindings as
easy as other languages would help get wanna-be programmers started when
they are young and their brains ready for molding :)
long:
A lot of people are coming to linux, and want to learn programming. Many
start or experiment with (after hello world) tetris, an mp3 player
front-end, a front-end to a command line tool, an irc-client, etc.
Starting such a project in C/C++ or one of the scripting languages
(perl/python/tcl/ruby) is easy because there is plenty tutorials, similar
programs with source available, easy-to-get-started-with gui toolkits,
networking libraries, etc.
With a small program up and running, it starts evolving, is rewritten from
scratch and dumped for more interesting projects and so on. During this
process nitty gritty details of language, libraries and toolkits are
absorbed, and idioms/work-arounds/standard techniques learned. Not to
mention that understanding of the basics of programming, compiling, how a
computer works grow - perhaps from almost nothing, in small steps.
Out of this group, some go on to be professional programmers, and are
exposed to different languages in a college/university setting. In my
case, common lisp was never mentioned as a multi-paradigm language. Rather
it was bundled with scheme (which was used to do the trivial exercises) and
functional languages, and while it seemed interesting I didn't see the
beauty in it. Java and C++ was used for bigger programming projects.
Now, impressed with P. Norvig's PAIP, I'd like to do *small* hobby free
software projects in CL. Right now I use C++, which is much more than
sufficient for the modest tasks - but after seeing how beautiful the code
*could* have been in CL, I'm constantly annoyed.
How would the situation have been if, say, KDE had been written CL? A lot
of programming hours have gone into KDE related libraries and programs
since it's inception. Many problems have been overcome. A lot of
infrastructure is made, removing a lot of boilerplate from small K*
utilities, and making it easy to start programming. (Similar with the gnome
project I assume.)
I'm not sure how much money the CL vendors could have made if freshmeat.net
and sourceforge.net were filled with CL programs, but it would have made my
life happier :). And with the demand of programmers at every skill-level
it shouldn't hurt to have a lot of people familiar with CL. For example
using lisp for web services, dynamic or just generated web pages, might
have been done on a bigger scale, with more solutions/frame-works
available. A rolling snow ball?
Enough rambling, I'll end with noting that the failure of universities to
give students a realistic picture of CL along with many programmers way of
sticking with the first language they got some amount of experience with
doesn't help grown CL's market share.
-- Rene
On 20 May 2001, Miles Egan wrote:
> Like many of the readers of this group, I've spent a lot of time pondering
> the popularity and evolution of lisp. I've also spent a lot of time
> playing with the various implementations of lisp out there and I've come to
> a few tentative conclusions:
I'll just briefly join the others who are piling on in saying that
pondering directions for open source projects, which is what you're doing
here, is a pretty useless activity. Only one thing talks in open source,
and that's code. Code you write cuz "it scratches an itch," code you
write cuz it's kool, code you write to make a statement and a reputation,
it doesn't matter. All else is fluff.
The Lisp vendors have other mechanisms for progress, but they probably
involve words like "runtime license" and "support contract."
> 1. Althouth Lisp users often criticize Scheme for it's overly conservative
> and incomplete definition, in many respects Lisp now suffers from the same
> incompleteness for modern application development. Many of the tools
> needed to develop typical modern-day applications are non-standard among
> implementations, crude and often only partially functional, or simply
> unavailable. FFI, database connectivity, network protocol bindings, gui
> toolkit bindings, xml tools etc. are fragmented and unstandardized among
> all CL implementations.
I'll just comment on FFI. There's a very strong standard for FFI. It's
called ANSI C. Yeah, yeah, people want to do Java to, but C FFI gets you
most of the way there and most things that you might want to do outside of
Lisp have a C interface.
I wanted to call XLib and OpenGL functions, but I wasn't willing to accept
the license on the packages I've seen out there. Nor did I want to
retype thousands of lines of interface declarations. So I wrote cparse,
http://www.bricoworks.com/~moore/cparse/. This parses C header files and
produces FFI declarations for CMUCL's Alien facility. I don't mention
this because I did it first (doubtful) or I believe it's unique, but I
know most about it. It's very easy to write the methods to produce FFI
declarations for other implementations, but I haven't done that because it
doesn't particularly push any of my buttons.
So you care about FFI portability? Jump in and fix the problems in cparse
that I haven't yet gotten to. Write an interface for your favorite Lisp
implementation. Define a general set of C types and mappings between
those and the implementations' FFI types. Send me your code. Problem
solved.
Tim
On 21 May 2001 07:01:25 GMT, Tim Moore <·····@herschel.bricoworks.com> wrote:
>I'll just briefly join the others who are piling on in saying that
>pondering directions for open source projects, which is what you're doing
>here, is a pretty useless activity. Only one thing talks in open source,
>and that's code.
Not all problems are technical. The issue I'm trying to address here is
social and can't be fixed with a patch.
>I'll just comment on FFI. There's a very strong standard for FFI. It's
>called ANSI C. Yeah, yeah, people want to do Java to, but C FFI gets you
>most of the way there and most things that you might want to do outside of
>Lisp have a C interface.
C is not an FFI. It's a programming langauge. An FFI is a bridge from one
language to another. The issue here is what it takes to glue Lisp and C
code together. Right now it's a hassle in free lisps. Take a look at the
clg code (http://ww.telent.net/cliki/clg, file cl-gtk.c) if you want to see
how gross this is in practice right now.
>So you care about FFI portability? Jump in and fix the problems in cparse
>that I haven't yet gotten to. Write an interface for your favorite Lisp
>implementation. Define a general set of C types and mappings between
>those and the implementations' FFI types. Send me your code. Problem
>solved.
cparse sounds like a nice tool and I think this is a smart way to attack
these kinds of problems, but you can't just dismiss the effort it takes to
make this work in five totally different FFIs. I'm like spending time
hacking on free software to make it better. I don't like wasting time
fixing hairy problems like this that shouldn't exist in the first place.
--
miles egan
On 21 May 2001, Miles Egan wrote:
> On 21 May 2001 07:01:25 GMT, Tim Moore <·····@herschel.bricoworks.com> wrote:
> >I'll just briefly join the others who are piling on in saying that
> >pondering directions for open source projects, which is what you're doing
> >here, is a pretty useless activity. Only one thing talks in open source,
> >and that's code.
>
> Not all problems are technical. The issue I'm trying to address here is
> social and can't be fixed with a patch.
With all due respect, the technical tends to drive the social in open
source; because coercion to do work isn't possible, other aspects must
flow from the work.
> >I'll just comment on FFI. There's a very strong standard for FFI. It's
> >called ANSI C. Yeah, yeah, people want to do Java to, but C FFI gets you
> >most of the way there and most things that you might want to do outside of
> >Lisp have a C interface.
>
> C is not an FFI. It's a programming langauge. An FFI is a bridge from one
> language to another. The issue here is what it takes to glue Lisp and C
> code together. Right now it's a hassle in free lisps. Take a look at the
> clg code (http://ww.telent.net/cliki/clg, file cl-gtk.c) if you want to see
> how gross this is in practice right now.
I know it's gross. I know what an FFI is. I know that some free Lisps,
e.g. CMUCL, have grave deficiencies (e.g., callable Lisp functions from
C) that need to be fixed. My point is that in most cases we are presented
with an authoritative declaration and specification of an interface: the C
header file. And, in most cases, what we want to do with that interface
is make function calls with simple data types or opaque pointers as
arguments. It's silly to wring one's hands about FFI incompatability when
instead one could write the code *once* to turn C declarations into Lisp
functions using a particular vendor's FFI.
> >So you care about FFI portability? Jump in and fix the problems in cparse
> >that I haven't yet gotten to. Write an interface for your favorite Lisp
> >implementation. Define a general set of C types and mappings between
> >those and the implementations' FFI types. Send me your code. Problem
> >solved.
>
> cparse sounds like a nice tool and I think this is a smart way to attack
> these kinds of problems, but you can't just dismiss the effort it takes to
> make this work in five totally different FFIs. I'm like spending time
> hacking on free software to make it better. I don't like wasting time
> fixing hairy problems like this that shouldn't exist in the first place.
I know I'm being a bit flip about the effort involved, but only a few
people need to make the coding effort to get to a solution that works
well; standardizing an FFI will involve unbounded agonizing and log
rolling that Lisp vendors and free Lisp hackers are unlikely to find
rewarding.
Tim
On 21 May 2001 17:22:47 GMT, Tim Moore <·····@herschel.bricoworks.com> wrote:
>> cparse sounds like a nice tool and I think this is a smart way to attack
>> these kinds of problems, but you can't just dismiss the effort it takes to
>> make this work in five totally different FFIs. I'm like spending time
>> hacking on free software to make it better. I don't like wasting time
>> fixing hairy problems like this that shouldn't exist in the first place.
>
>I know I'm being a bit flip about the effort involved, but only a few
>people need to make the coding effort to get to a solution that works
>well; standardizing an FFI will involve unbounded agonizing and log
>rolling that Lisp vendors and free Lisp hackers are unlikely to find
>rewarding.
I agree completely, which is why I'm not advocating a standard FFI
in ANSI sense of the word standard. I agree that the road to such a
standard would be completely hellish.
What I am suggesting is that those of us interested in advancing free lisp
focus on a single implementation, polish its FFI, and start using it to
build bridges to the rest of the world and abandon any idea of supporting
any other CL implementations.
--
miles egan
·····@puzl.pixar.com (Miles Egan) wrote in message news:<····················@puzl.pixar.com>...
> Like many of the readers of this group, I've spent a lot of time pondering
> the popularity and evolution of lisp. I've also spent a lot of time
> playing with the various implementations of lisp out there and I've come to
> a few tentative conclusions:
>
> 1. Althouth Lisp users often criticize Scheme for it's overly conservative
> and incomplete definition, in many respects Lisp now suffers from the same
> incompleteness for modern application development. Many of the tools
> needed to develop typical modern-day applications are non-standard among
> implementations, crude and often only partially functional, or simply
> unavailable. FFI, database connectivity, network protocol bindings, gui
> toolkit bindings, xml tools etc. are fragmented and unstandardized among
> all CL implementations.
This is an important point, but see, the only languages that don't
have this problem are the ones that are proprietary: Java, Perl,
Python, TCL. You may not think of these language as proprietary, but
(except maybe for Java) each one has one and only one implementation.
With only one implementation, there is only 1 way to do things.
Non-proprietary languages like C/C++ and Lisp suffer from this
problem. There are 2 solutions:
(1) You have a small language core that's standardized (like C and
Scheme). You rely on 3rd party libraries for everything else!
(2) You have a large language core that's standardized (like C++ and
Lisp). You still rely on 3rd party libraries for everything else. No
matter how huge your core language is, there's always going to be
stuff that falls outside. I wish sockets had been a part of the C.L.
standard, but they aren't; oh well...
Lisp is at a disadvantage here because it's not a proprietary
language. It's a standard that has evolved out of usage, with a small
number of surviving implementations.
It's also at an advantage for the same reason. A lot of the stuff that
new languages are still dealing with has been addressed in Lisp.
[...]
> 3. No langauge standardization bureaucracy. The langauge designers are
> free to evolve the langauge when necessary. These langauges have evolved
> quickly in response to user requests and technological shifts.
There's always some sort of bureaucracy (or dictatorship, in the case
of Perl). Somebody or some group has to be there to keep the interests
of the users in mind.
The other things you want are, IMHO, already available in CLISP. Or,
if you're doing Windows development, Corman Lisp. It's not that hard
to interface either one of these with C libraries or DLLs. (Try it!)
[...]
g
glauber wrote:
> This is an important point, but see, the only languages that don't
> have this problem are the ones that are proprietary: Java, Perl,
> Python, TCL. You may not think of these language as proprietary, but
> (except maybe for Java) each one has one and only one implementation.
> With only one implementation, there is only 1 way to do things.
Python has been completely re-implemented in Java -- it's called "Jython".
There was also an experimental branch/fork of Python known as "Stackless
Python", and perhaps other branches/forks I'm not aware of.
Topaz was an effort to rewrite Perl5 in C++ but I think it's been abandoned.
And of course, there are many different JVMs for Java (some of them
open-source, some proprietary).
The reason most of these languages have only 1 implementation is because
the 1 reference implementation does everything that is needed. What need is
there for 10 different implementations of a language, each with different
bugs and each costing a lot of time to develop, when 1 high-quality
implementation would do just as well? To call a language with a single
mainstream implementation "proprietary" is bizarre.
>> 3. No langauge standardization bureaucracy. The langauge designers are
>> free to evolve the langauge when necessary. These langauges have evolved
>> quickly in response to user requests and technological shifts.
>
> There's always some sort of bureaucracy (or dictatorship, in the case
> of Perl). Somebody or some group has to be there to keep the interests
> of the users in mind.
I'd hardly describe Perl's design/maintainance as a dictatorship. The code
is all free; you're free to fork it and maintain your own version. Of
course, all the main developers happen to agree on a single implementation
and work together designing and improving things. And in this
implementation, they control what goes in. That's hardly dictatorship.
Cheers,
Neil
Miles Egan wrote:
> ... My impression of the popular open-source
> languages (Perl, Python, Ruby, Tcl and, to a lesser extent, Ocaml) is that
> they all have the following in common:
>
> 1. A single implementation. This means any extensions or add-on modules
> are developed only once and that their authors need only worry about
> architecture-related portability issues. All developer effort goes toward
> maintaining and improving a single common implementation.
>
...
>
> 2. A simple and straightforward means of bootstrapping the language on any
> sane platform with a c compiler. All of these languages are very portable.
>
...
> 3. No langauge standardization bureaucracy. The langauge designers are
> free to evolve the langauge when necessary. These langauges have evolved
> quickly in response to user requests and technological shifts.
GCL would be pretty close to meeting points 1 and 2, if it were still
being maintained adequately. The problem is we as Lispers are spoiled
by true dynamic compilation - printing C source code to a gcc/as pipe
and dynamically loading the file that falls out of the other end is
awfully klunky.
Also, the flip side of 3 above is that the language specification isn't
a moving target, and doesn't intentionally contain platform- or
implementation-specific junk. Having a single implementation means you
can change quickly, but it limits how fast you can ultimately be. I
predict it will be a long time before any of the languages you
mention[1] compile to native code, anywhere.
[1] I vaguely recall soemthing about good compilers for Ocaml, so I may
be wrong about it.
--
Remove obvious stuff to e-mail me.
Bob Bane
On Mon, 21 May 2001 16:36:43 -0400, Bob Bane <····@removeme.gst.com> wrote:
>> 3. No langauge standardization bureaucracy. The langauge designers are
>> free to evolve the langauge when necessary. These langauges have evolved
>> quickly in response to user requests and technological shifts.
>
>Also, the flip side of 3 above is that the language specification isn't
>a moving target, and doesn't intentionally contain platform- or
>implementation-specific junk. Having a single implementation means you
>can change quickly, but it limits how fast you can ultimately be. I
>predict it will be a long time before any of the languages you
>mention[1] compile to native code, anywhere.
The idea is that there is no target, the language IS the target and the
designers are the implementors. I think open-source language development
changes all the weights in the old language standardization equation.
>[1] I vaguely recall soemthing about good compilers for Ocaml, so I may
>be wrong about it.
Ocaml comes with a very fast bytecode interpreter and a superb compiler
that generates very good native code for x86, Alpha, Sparc, Mips, HP
PA-Risc, PowerPC, Strong ARM, and IA64.
Ocaml is an interesting example and is fairly illustrative of the point I'm
trying to make. Ocaml has grown by leaps and bounds in the past few years.
Major language features have been added and the compiler has been refined
to the point that it generates code that often compares favorably to C++
code but lets the programmer work at a much, much higher level. In
contrast, the various Standard ML implementations all struggle along
seperately and continue to lose ground.
I think Ocaml owes a lot of its success to the exceptional skill of its
implementors, but I think the freedom of working within a single
implementation is also an important factor. Ocaml is starting to build
momentum and they're making real headway on the kind of essential toolkit
we're talking about for Lisp. The freedom and directness with which they
can carry out this work is a lesson, I think.
--
miles egan
Miles Egan wrote:
> Like many of the readers of this group, I've spent a lot of time pondering
> the popularity and evolution of lisp. I've also spent a lot of time
> playing with the various implementations of lisp out there and I've come
> to a few tentative conclusions:
>
> 1. Althouth Lisp users often criticize Scheme for it's overly
> conservative and incomplete definition, in many respects Lisp now suffers
> from the same
> incompleteness for modern application development. Many of the tools
> needed to develop typical modern-day applications are non-standard among
> implementations, crude and often only partially functional, or simply
> unavailable. FFI, database connectivity, network protocol bindings, gui
> toolkit bindings, xml tools etc. are fragmented and unstandardized among
> all CL implementations.
This may count for socket, DB-interfaces and GUIs - but I don't see where
this problem is related to xml tools.
Yes it would be nice if we had a common FFI but I don't think that this
point will change next time. It may work to make the interfaces at least a
little bit more similar to make porting a bit easier. But I think the
possibility of having a tool that reads header-files and produces FFI code
could really be a feasible solution to that problem (as others mentioned).
> 2. Ironically, the diversity of CL implementations seems to exacerbate
> the
> problem of extending CL. The implementations I've seen of some of the
> above tools clearly expend a lot of effort trying to support them all,
> with code which is a mess of implementation-specific switches and parallel
> FFI
> code. I have to imagine that time spent dealing with cross-implementation
> compatibility is time not spend adding robustness, functionality, or
> documentation. The irony here, of course, is that in some ways it is the
> very standardization of Lisp that allows for this kind of forking because
> there is a reference standard upon which to base new and otherwise
> divergent implementations.
No - you don't need a ANSI standard to standarize interfaces. I don't know
why some people get the impression that a ANSI-Standard is a drawback
compared to not standardized languages.
As Kent has mentioned it seems to be impossible actually to enhance the
ANSI-Standard - but this doesn't mean that we could not create
community-standards that bind maybe as strong as the ANSI-Standard.
> A survey of the handful of languages that are growing reveals some obvious
> commonalities. Living languages today are either funded and marketed by
> huge corporations with vast advertising and development budgets or are
> open-source languages supported by volunteers. Since a Java or C# push
> isn't likely for Lisp, the only alternative seems to be a strong
> open-source development effort. My impression of the popular open-source
> languages (Perl, Python, Ruby, Tcl and, to a lesser extent, Ocaml) is that
> they all have the following in common:
>
> 1. A single implementation. This means any extensions or add-on modules
> are developed only once and that their authors need only worry about
> architecture-related portability issues. All developer effort goes toward
> maintaining and improving a single common implementation.
Not true Python has at least CPython, Jython and Stackless and I don't
think that this is a drawback to it.
> 2. A simple and straightforward means of bootstrapping the language on
> any
> sane platform with a c compiler. All of these languages are very
> portable.
CLISP seems to be rather portable to me...
But a question:
We have CMUCL on UNIX/Linux, Corman Lisp on Windows, the free MCL on Mac
and CLISP on a huge amount of platforms, and not to forget the commercial
Lisps LW + ACL - I don't think that the community should concentrate on a
more portable Common Lisp.
> 3. No langauge standardization bureaucracy. The langauge designers are
> free to evolve the langauge when necessary. These langauges have evolved
> quickly in response to user requests and technological shifts.
I see no real need to change the language "Common Lisp" - I think we need
to concentrate on the library-parts and common interfaces.
> What can be done about this? I think it's imperative that the open-source
> Lisp community should learn from the Pythonistas:
>
> 1. Pick a single open-source Lisp and focus all development effort on
> that
> implementation. We have more than our share of brilliant implementors,
> but splitting our efforts among CLISP, CMUCL and SBCL is lethal.
No! - I think diversity is an important thing. I use both CLISP and CMUCL
and I think each of them has it's own capabilities. CLISP is very small,
and portable - it is the ideal scripting-language (and more!). CMUCL is a
nice fast native-compiling system. I don't want to have CMUCL as small as
CLISP because I think it will loose much of it's capabilities and I don't
want CLISP to get so big as CMUCL is.
SBCL is IMHO a good idea - often it is better do a fork so you can risk to
get incompatible with your roots and maybe to fail without destroying the
root. I think SBCL was a amazing success so far.
> 2. Do what it takes to make porting this implementation as simple as
> possible. This implementation should compile and run on any new
> posix-compliant platform with nothing more than a C compiler. It should
> also be trivial to make executables with this Lisp and it should function
> as an executable script interpreter with no extra gymnastics.
We don't critically need more platforms - we have more than some other
languages have.
> 3. Take advantage of having a single reference implementation to define a
> clean and straightforward CPAN-like system and a well-documented and
> straightforward C FFI. I know there are efforts to define something like
> this now, but, to my eyes, they are all severely limited by the difficulty
> of supporting so many implementations.
I think it's in the beginnings.
> 4. Build full-featured and reliable interfaces to common toolkits and
> protocols on this infrastructure. Lisp should talk to databases as
> fluently as Perl, parse XML as well as Python, speak SNMP, HTTP, LDAP,
> IMAP, and SOAP, and couple with C as easily as Ruby.
I've seen XML-parsers, SNMP-interfaces and HTTP-Servers and I'm sure the
other things will come too.
> 5. Take advantage of having a single implementation to extend and change,
> when necessary, the Common Lisp standard. Obviously this should be done
> extremely carefully, but I think it's a mistake at this point to consider
> the CL spec to be either complete or sacrosanct.
There's no critical need to change the ANSI-Standard - create
community-standards.
> Speaking as an open-source hacker that would love to do what he can to
> help
> promote and grow lisp, I think these kinds of changes are necessary. As
> it stands now, trying to get anything done in the current Babel is even
> more frustrating than watching Ruby, Perl, and Python slowly reinvent Lisp
> the hard way.
You're welcome to contribute - there's a lot things to do.
Regards,
Jochen
Jochen Schmidt <···@dataheaven.de> writes:
> No - you don't need a ANSI standard to standarize interfaces. I don't know
> why some people get the impression that a ANSI-Standard is a drawback
> compared to not standardized languages.
> As Kent has mentioned it seems to be impossible actually to enhance the
> ANSI-Standard - but this doesn't mean that we could not create
> community-standards that bind maybe as strong as the ANSI-Standard.
And we actually do have them already, the MOP (described in AMOP) and
e.g. gray-streams. Some of those de-facto community standards are of
lower quality than the ANSI standard, which IMHO (after having read
and worked with many, many standards) is of exceptionally high
quality, but they do exist, and most implementors do take them quite
seriously.
It seems to me that new layered standards will probably come from
single vendors and users that are willing and able to do the upfront
work required, and where those proposals are acceptable to other
vendors/users with slight changes.
Regs, Pierre.
--
Pierre R. Mai <····@acm.org> http://www.pmsf.de/pmai/
The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We cause accidents. -- Nathaniel Borenstein