Are there any plans to develop/foster PowerLisp for the Mac again? The
developer of PowerLisp for the Mac abandoned the Macintosh-platform in
1996 (and is now serving well in the pointless Microsoft-world).
There are often people which are willing to contribute (or searching for
useful projects) to the folks in a way of GNU (or whatever you like
here; GNU is not always the best). PowerLisp has been Shareware all the
time, but maybe the author is prepared to hand in PowerLisp to a
potential developer for the Macintosh platform.
PowerLisp has been not a toy-project (it even had a powerful
PPC-compiler) and hence could serve as a scaffolding. What PowerLisp
misses are GUIs or at least a possibility to link to GUI-tools.
S. Gonzi
In article <·················@kfunigraz.ac.at>, Siegfried Gonzi
<···············@kfunigraz.ac.at> wrote:
> Are there any plans to develop/foster PowerLisp for the Mac again? The
> developer of PowerLisp for the Mac abandoned the Macintosh-platform in
> 1996 (and is now serving well in the pointless Microsoft-world).
>
...
> PowerLisp has been not a toy-project (it even had a powerful
> PPC-compiler) and hence could serve as a scaffolding. What PowerLisp
> misses are GUIs or at least a possibility to link to GUI-tools.
What would be the point? MCL (www.digitool.com) is available for the Mac
and it's a much better implementation of Common Lisp than PowerLisp ever
was. (IMHO, it's the best commercial CL available for _any_ platform
with the possible exception of Genera for the Alpha).
If it's Open Source you're looking for, check out OpenMCL
(openmcl.clozure.com). It's derived from MCL but a) it runs in Linux for
PowerPC, not MacOS, and b) it has no GUI or IDE, so it's not nearly as
nice for development as commercial MCL. But hey, it's open source, so
you can change it!
Shannon Spires
svspire-at-nmia-dot-com
Shannon Spires wrote:
>
>
> What would be the point? MCL (www.digitool.com) is available for the Mac
> and it's a much better implementation of Common Lisp than PowerLisp ever
> was. (IMHO, it's the best commercial CL available for _any_ platform
> with the possible exception of Genera for the Alpha).
www.digitool.com was dead for a week (but I have to see that it works
today). Therefore I thought they are out of business (like Dylan).
S. Gonzi
Siegfried Gonzi <···············@kfunigraz.ac.at> writes:
> Shannon Spires wrote:
> >
>
> >
> > What would be the point? MCL (www.digitool.com) is available for the Mac
> > and it's a much better implementation of Common Lisp than PowerLisp ever
> > was. (IMHO, it's the best commercial CL available for _any_ platform
> > with the possible exception of Genera for the Alpha).
>
> www.digitool.com was dead for a week (but I have to see that it works
> today). Therefore I thought they are out of business (like Dylan).
What do you mean by "like Dylan"? Dylan is a *language*, not a
company. Functional Objects, at http://www.fun-o.com, has a commercial
implementation, and Fun-O appears to be alive and kicking. There is also
an Open-Source implementation (Gwydion Dylan).
--
Raymond Wiker
·············@fast.no
Raymond Wiker wrote:
> > www.digitool.com was dead for a week (but I have to see that it works
> > today). Therefore I thought they are out of business (like Dylan).
>
> What do you mean by "like Dylan"? Dylan is a *language*, not a
> company.
Yes I know. But find any Dylan-implementation for the Mac OS9.
> Functional Objects, at http://www.fun-o.com, has a commercial
> implementation, and Fun-O appears to be alive and kicking. There is also
> an Open-Source implementation (Gwydion Dylan).
I know.
S. Gonzi
In article <·················@kfunigraz.ac.at>, Siegfried Gonzi
<···············@kfunigraz.ac.at> wrote:
> Raymond Wiker wrote:
>
>
> > > www.digitool.com was dead for a week (but I have to see that it works
> > > today). Therefore I thought they are out of business (like Dylan).
> >
> > What do you mean by "like Dylan"? Dylan is a *language*, not a
> > company.
>
> Yes I know. But find any Dylan-implementation for the Mac OS9.
>
> > Functional Objects, at http://www.fun-o.com, has a commercial
> > implementation, and Fun-O appears to be alive and kicking. There is also
> > an Open-Source implementation (Gwydion Dylan).
>
> I know.
So in that case you should know that Gwydion Dylan has a version of
their compiler packaged as a plug-in compiler for the MacOS 9
CodeWarrior IDE.
In fact, the Gwydion project has both the (new) "d2c" compiler which
produces C which is then then compiled to machine code AND the older
"Mindy" bytecode compiler. Both are available on MacOS as both
stand-alone drag-and-drop compilers and as CodeWarrior plug-ins, though
the Mindy plug-in hasn't been maintained for a while...
And besides that, there is still the Apple Technology Release, which
still works on current versions of MacOS and goes *really* fast on the
latest CPUs.
-- Bruce
Bruce Hoult wrote:
> >
> > > Functional Objects, at http://www.fun-o.com, has a commercial
> > > implementation, and Fun-O appears to be alive and kicking. There is also
> > > an Open-Source implementation (Gwydion Dylan).
> >
> > I know.
>
> So in that case you should know that Gwydion Dylan has a version of
> their compiler packaged as a plug-in compiler for the MacOS 9
> CodeWarrior IDE.
The problem is that: I do not have CodeWarrior (some times I use MPW,
though). And even there were CodeWarrior near my side I wouldn't install
Dylan as *plug-in*.
> In fact, the Gwydion project has both the (new) "d2c" compiler which
> produces C which is then then compiled to machine code AND the older
> "Mindy" bytecode compiler. Both are available on MacOS as both
> stand-alone drag-and-drop compilers and as CodeWarrior plug-ins, though
> the Mindy plug-in hasn't been maintained for a while...
Such akward(ness) doesn't get my attention. Believe me it is not a good
deal to go the way via C, especially when somone wants to introduce new
programming (language) paradigms.
S. Gonzi
In article <·················@kfunigraz.ac.at>, Siegfried Gonzi
<···············@kfunigraz.ac.at> wrote:
> Bruce Hoult wrote:
>
> > >
> > > > Functional Objects, at http://www.fun-o.com, has a commercial
> > > > implementation, and Fun-O appears to be alive and kicking. There is
> > > > also
> > > > an Open-Source implementation (Gwydion Dylan).
> > >
> > > I know.
> >
> > So in that case you should know that Gwydion Dylan has a version of
> > their compiler packaged as a plug-in compiler for the MacOS 9
> > CodeWarrior IDE.
>
> The problem is that: I do not have CodeWarrior (some times I use MPW,
> though). And even there were CodeWarrior near my side I wouldn't install
> Dylan as *plug-in*.
So in that case you shouldn't confuse mere preferences with what
actually exists.
> > In fact, the Gwydion project has both the (new) "d2c" compiler which
> > produces C which is then then compiled to machine code AND the older
> > "Mindy" bytecode compiler. Both are available on MacOS as both
> > stand-alone drag-and-drop compilers and as CodeWarrior plug-ins, though
> > the Mindy plug-in hasn't been maintained for a while...
>
> Such akward(ness) doesn't get my attention. Believe me it is not a good
> deal to go the way via C, especially when somone wants to introduce new
> programming (language) paradigms.
It worked for C++.
If it's invisible to the user, and doesn't slow the process down too
much then it's a good way to get on lots of different machine types
quickly.
-- Bruce
Bruce Hoult <·····@hoult.org> writes:
> In article <·················@kfunigraz.ac.at>, Siegfried Gonzi
> <···············@kfunigraz.ac.at> wrote:
>
> > Bruce Hoult wrote:
> >
> > > >
> > > > > Functional Objects, at http://www.fun-o.com, has a commercial
> > > > > implementation, and Fun-O appears to be alive and kicking. There is
> > > > > also
> > > > > an Open-Source implementation (Gwydion Dylan).
> > > >
> > > > I know.
> > >
> > > So in that case you should know that Gwydion Dylan has a version of
> > > their compiler packaged as a plug-in compiler for the MacOS 9
> > > CodeWarrior IDE.
> >
> > The problem is that: I do not have CodeWarrior (some times I use MPW,
> > though). And even there were CodeWarrior near my side I wouldn't install
> > Dylan as *plug-in*.
> So in that case you shouldn't confuse mere preferences with what
> actually exists.
>>> In fact, the Gwydion project has both the (new) "d2c" compiler
>>> which produces C which is then then compiled to machine code AND
>>> the older "Mindy" bytecode compiler. Both are available on MacOS
>>> as both stand-alone drag-and-drop compilers and as CodeWarrior
>>> plug-ins, though the Mindy plug-in hasn't been maintained for a
>>> while...
>> Such akward(ness) doesn't get my attention. Believe me it is not a
>> good deal to go the way via C, especially when somone wants to
>> introduce new programming (language) paradigms.
> It worked for C++.
> If it's invisible to the user, and doesn't slow the process down too
> much then it's a good way to get on lots of different machine types
> quickly.
.. And it means that there's a reasonable way of deploying on
unexpected platforms.
It may have the demerit of not allowing taking direct advantage of
ways that your implementation could be _directly_ translated into
machine language; you wind up using whatever C winds up using.
I seem to recall Erik Naggum indicating that the most likely
implementation scheme for his "Emacs in Common Lisp" proposal would be
to build CL code that would generate, as output, _C_.
As a result, while you'd need to have a CL implementation around,
_somewhere_, it wouldn't forcibly need to run on the target platform.
You might, for instance, generate "CLEmacs" on a PowerMac, using a
PPC-based CL, but then take the resultant C code and compile it
repeatedly on a whole bunch of target platforms.
I was initially fairly surprised by this approach, but it does make
quite a lot of sense.
You could probably get output of a factor "Epsilon" better by directly
generating machine code, but then you're left with either:
a) Only deploying on a single platform, or
b) Creating a Virtual Machine, ala Java or C#, or
c) Having to pick some "lowest common denominator" representation
almost like a VM so that you can write various backends for various
platforms.
Using C has none of the demerits of this...
It could perhaps be further interesting to generate code in another
language; major options being C++ (which doesn't yet have a defacto
standard ABI the way C tends to) or (somewhat vaporish) C--, a
semi-functional, somewhat "dumber" version of C intended specifically
as a form to be generated by other languages.
--
(concatenate 'string "aa454" ·@freenet.carleton.ca")
http://vip.hyperusa.com/~cbbrowne/sgml.html
"Once you accept that the world is a giant computer run by white mice,
all other movies fade into insignificance." -- Mutsumi Takahashi
In article <·················@kfunigraz.ac.at>, Siegfried Gonzi
<···············@kfunigraz.ac.at> wrote:
> Shannon Spires wrote:
> >
>
> >
> > What would be the point? MCL (www.digitool.com) is available for the Mac
> > and it's a much better implementation of Common Lisp than PowerLisp ever
> > was. (IMHO, it's the best commercial CL available for _any_ platform
> > with the possible exception of Genera for the Alpha).
>
> www.digitool.com was dead for a week (but I have to see that it works
> today). Therefore I thought they are out of business (like Dylan).
They seem to have frequent problems with their internet connection, but
the website was working this morning. They're certainly still
in business. I just received an update CD last week. FWIW, MCL now runs
on OS9 and in "classic" mode in OS X. They're working on making
it OS X native.
Shannon Spires
svspire-at-nmia-dot-com
In article <·················@kfunigraz.ac.at>,
Siegfried Gonzi <···············@kfunigraz.ac.at> wrote:
> Are there any plans to develop/foster PowerLisp for the Mac again? The
> developer of PowerLisp for the Mac abandoned the Macintosh-platform in
> 1996 (and is now serving well in the pointless Microsoft-world).
>
> There are often people which are willing to contribute (or searching for
> useful projects) to the folks in a way of GNU (or whatever you like
> here; GNU is not always the best). PowerLisp has been Shareware all the
> time, but maybe the author is prepared to hand in PowerLisp to a
> potential developer for the Macintosh platform.
Why don't you ask Roger Corman directly ?
It is quite possible that he will give his pet (project) into the hands
of someone who can convince him will take good care of it.
I spoke to him about two years ago about further development of his
compiler and he was rather sorry himself that he had abandoned it.
Although he is an idealistic kind of guy (would he have developed it
otherwise in the first place?!), he said he just didn't have the time
and shareware fees didn't justify to spend much time on it.
I have to say though, that in order to stand up against MCL, there needs
to be a lot of work done. As a shareware project I can understand why it
didn't take off. First, let's face it, there are less Lisp interested
people out there than those who are using most other languages. Then, it
would usually be students who would opt for a cheap shareware for
educational purposes. However, a student license for a "newsstand" issue
of MCL is something like 80 USD or so IIRC. Given the quality of MCL,
for that kind of money even the poorest student is likely to go for MCL.
A pity for all that good work that Roger has put into it.
rgds
BK
--
bk <·········@yahoo.com>
ATTENTION! Email to this address *must* contain "USENET" in the subject line
otherwise it will be considered spam, automatically deleted and never reach me.
On Fri, 17 Aug 2001 13:58:34 +0900, BK <·········@yahoo.com> wrote:
>In article <·················@kfunigraz.ac.at>,
> Siegfried Gonzi <···············@kfunigraz.ac.at> wrote:
>
>> Are there any plans to develop/foster PowerLisp for the Mac again? The
>> developer of PowerLisp for the Mac abandoned the Macintosh-platform in
>> 1996 (and is now serving well in the pointless Microsoft-world).
>>
>> There are often people which are willing to contribute (or searching for
>> useful projects) to the folks in a way of GNU (or whatever you like
>> here; GNU is not always the best). PowerLisp has been Shareware all the
>> time, but maybe the author is prepared to hand in PowerLisp to a
>> potential developer for the Macintosh platform.
>
>Why don't you ask Roger Corman directly ?
I am prepared to let anybody work on this who wishes to. I have recently updated
the sources, converted to a newer version of CodeWarrior and all universal
headers. I also ripped out all the 68k support. The new version, complete with
all source code as well as binaries, is available on the web. I haven't had time
to update the PowerLisp web page, so there is no link to it yet. So here it is:
http://corman.net/PowerLisp/PowerLisp_2_02_sea.hqx
>
>It is quite possible that he will give his pet (project) into the hands
>of someone who can convince him will take good care of it.
>
>I spoke to him about two years ago about further development of his
>compiler and he was rather sorry himself that he had abandoned it.
>Although he is an idealistic kind of guy (would he have developed it
>otherwise in the first place?!), he said he just didn't have the time
>and shareware fees didn't justify to spend much time on it.
>
>I have to say though, that in order to stand up against MCL, there needs
>to be a lot of work done. As a shareware project I can understand why it
>didn't take off. First, let's face it, there are less Lisp interested
>people out there than those who are using most other languages. Then, it
>would usually be students who would opt for a cheap shareware for
>educational purposes. However, a student license for a "newsstand" issue
>of MCL is something like 80 USD or so IIRC. Given the quality of MCL,
>for that kind of money even the poorest student is likely to go for MCL.
I never expected it would compete with Digitool's, or that I would make any
serious income from this. Although at the time it was initially developed, and
Apple owned MCL, there was no student version. I think $500 was the minimum
entry, and I thought that a lower priced alternative was a good idea. Quite a
lot of others thought so too. For a year or so, it was the only PowerPC native
Common Lisp.
While I understand the decision to maintain a high price for existing lisp
systems re market size and development costs, I believe even $500 is outside the
range that people are willing to pay unless is clearly fills a professional
need. Many people want to use lisp as a supplementary tool, for prototyping,
playing, etc. Then maybe they will move into serious programming with it. The
high price is a real drawback. I have never been able to afford $500 for such a
tool.
I made the decision to move the the Windows world because
a) at that time (circa 1996) Harlequin had no product on Windows, and
there were some significant deficiencies in Franz's product (IMHO). Also
Franz was quite expensive.
b) Windows (even Windows 95) provided a better platform for what I wanted (with
protected address spaces, preemptive threads, etc).
c) I was moving more to Windows platforms in other work I was doing (because of
the larger market). So I would say the move was only indirectly related to the
market size.
I feel that Corman Lisp is a far superior implementation, and at this point much
richer in features. I can't justify continuing to expend resources on PowerLisp.
If I ever were to release another Mac version I would probably port Corman Lisp
(say to OS X).
>
>A pity for all that good work that Roger has put into it.
Please don't think of it this way. I don't. It was an interesting project,
personally rewarding if not financially, and many individuals and universities
have found it useful. It was an important step to developing Corman Lisp, and
also formed the basis of the decision support system that I wrote for hospital
clinical systems. These are in widespread use, preventing problems and even
saving lives. A huge amount of work went into PowerLisp, I will admit that, but
I don't think it was any kind of a waste of effort.
>
Roger
BK wrote:
> I have to say though, that in order to stand up against MCL, there needs
> to be a lot of work done. As a shareware project I can understand why it
> didn't take off. First, let's face it, there are less Lisp interested
> people out there than those who are using most other languages. Then, it
> would usually be students who would opt for a cheap shareware for
> educational purposes. However, a student license for a "newsstand" issue
> of MCL is something like 80 USD or so IIRC. Given the quality of MCL,
> for that kind of money even the poorest student is likely to go for MCL.
Perhaps if he had made PowerLisp available under the GPL it might have
taken off. Why pay $80 (US) for MCL if you can get a GPL PowerLisp for
free?
In article <·················@mailbag.com>, ·····@callsign.net wrote:
>Perhaps if he had made PowerLisp available under the GPL it might have
>taken off.
It what sense do you mean "it might have taken off"? (1)That developers
would have been attracted to it and turned it into something worth
wanting in the year 2001? Or...
>Why pay $80 (US) for MCL if you can get a GPL PowerLisp for
>free?
(2) If it had the GPL on it, people would use it as is?
If (1), then it's too late so who cares?
If (2), then I must state that MCL is an awesome Lisp implementation
and has been for the past decade. I would hate to see pure cheapness
(why pay $80 when I can have something with 75% of the functionality
for free?) put a great outfit like Digitool out of business. They're
taking a lot of risk providing an implementation for what is, in the
current business environment, a marginal language for a marginal
platform.
Or perhaps you had a third reason in mind (e.g., a belief in the
superiority of free-as-in-GPL software)?
Sashank