From: Kenny Tilton
Subject: You want braces? I'll give you braces...
Date:
Message-ID: <3D76A1A6.2060506@nyc.rr.com>
...and commas and colons and full-stops and...
copyright curl programming bible:
{define-class public AgeException {inherits Exception}
{constructor public {default
value:any,
filename:#Url=null,
line:int=0,
column:int=0
}
{construct-super
value, url=filename, line=line, column=column
}
}
}
{define-proc {set-user-age age:int}
let driver-age:#int
{if age < 1 then
{throw {AgeException "Invalid age"}}
else {if age < 16 then
{throw {AgeException "Still too young to drive"}}
else
set driver-age = age
}
}
}
{define-proc {register-user name:String, age:int}:String
let ret-value = "Success"
{try
{set-user-age age}
catch e:AgeException do
set ret-value = {e.to-String}
}
{return ret-value}
}
{let result:String = {register-user "Jeff", 18}}
Jeff: {value result}
{set result = {register-user "John", -12}}
John: {value result}
{set result = {register-user "David", 12}}
David: {value result}
Curl is apparently derived most closely from Dylan, tho no GFs, but at
least they kept sexprs/prefix, except they changed the parens to braces
and have a /lot/ of other punctuation ala Java/C++. One weird thing is
that in trying to be helpful they made the braces optional in some
circumstances.
The cool thing is they have started out as purely a browser plugin, tho
standalone apps are promised, so a web page can now be a full-blown,
event-handling interactive application programmed in a high-power language.
The other cool thing is not pandering to the C crowd by dumping
sexprs/prefix.
Another cool thing: if it takes off we have one more win for the Lisp
family, which I still maintain will bury the C/C++/Java family.
OK, the bad news is Windows-only, but Mac and Linux are promised (but
not even RSN).
Now where is the Lisp plugin so I do not have to learn Curl?
:)
kenny
ilias wrote:
>
> Kenny Tilton wrote:
> > ....and commas and colons and full-stops and...
> >
> > copyright curl programming bible:
[...]
> pointer to a small introducing document to curl?
http://www.curl.com
--
Pascal Costanza University of Bonn
···············@web.de Institute of Computer Science III
http://www.pascalcostanza.de R�merstr. 164, D-53117 Bonn (Germany)
Pascal Costanza wrote:
> ilias wrote:
>
>>Kenny Tilton wrote:
>>
>>>....and commas and colons and full-stops and...
>>>
>>>copyright curl programming bible:
>>
>
> [...]
>
>
>>pointer to a small introducing document to curl?
>
>
> http://www.curl.com
thank you. academic people (3 out of 10 ar Dr.'s)
http://www.curl.com/html/about/management.jsp
Curl looks like what Paul Graham wants to make with his OpenSource-ARC.
http://www.paulgraham.com/arc.html
The curl code looks like 'mutated LISP'.
For me it seems like they'll be very successfull. Cause they don't have
syntax-romance. That's business.
What is your opinion?
ilias <·······@pontos.net> writes:
...
>
> For me it seems like they'll be very successfull. Cause they don't
> have syntax-romance. That's business.
>
> What is your opinion?
Greenspun's Tenth applies. :)
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th 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 wrote:
> ilias <·······@pontos.net> writes:
>
> ...
>
>>For me it seems like they'll be very successfull. Cause they don't
>>have syntax-romance. That's business.
>>
>>What is your opinion?
>
>
> Greenspun's Tenth applies. :)
>
> Cheers
>
explain me this, please.
Centuries ago, Nostradamus foresaw when ilias <·······@pontos.net> would write:
> Marco Antoniotti wrote:
>> ilias <·······@pontos.net> writes:
>> ...
>>
>>>For me it seems like they'll be very successfull. Cause they don't
>>>have syntax-romance. That's business.
>>>
>>>What is your opinion?
>> Greenspun's Tenth applies. :)
>
> explain me this, please.
No. Get thee to google, and type in "greenspun's tenth." Do your own
research.
--
(reverse (concatenate 'string ··········@" "enworbbc"))
http://www3.sympatico.ca/cbbrowne/multiplexor.html
:FATAL ERROR -- ILLEGAL ERROR
ilias wrote:
> explain me this, please.
Type google.com in the nearest web-browser. Type "greenspun tenth" at
the search prompt and hit the "I'm feeling lucky button." All should
become clear. (Although I won't hold my breath).
:)w
Will Deakin wrote:
> ilias wrote:
>
>> explain me this, please.
>
> Type google.com in the nearest web-browser. Type "greenspun tenth" at
> the search prompt and hit the "I'm feeling lucky button." All should
> become clear. (Although I won't hold my breath).
>
> :)w
>
Token A
http://www.google.com/search?q=
Token B
greenspun+tenth
http://www.google.com/search?q=greenspun+tenth
less typing - more gentleness.
I've constructed it reusable. You can reuse it. Although it seems that
you feel only luck with...
much typing - less gentleness.
Will Deakin wrote:
> ilias wrote:
>
>> less typing - more gentleness.
>
> Uh? Learn to type.
>
>
geentlleness.
i can't type it - i have it.
you can type it.
be happy.
Lars Brinkhoff wrote:
> ilias <·······@pontos.net> writes:
>
>>Marco Antoniotti wrote:
>>
>>>Greenspun's Tenth applies. :)
>>
>>explain me this, please.
>
>
> Greenspun's Tenth Rule of Programming:
> http://philip.greenspun.com/research/
thanx:
> Quotes
> Greenspun's Tenth Rule of Programming: "Any sufficiently complicated C or Fortran program contains an ad-hoc, informally-specified bug-ridden slow implementation of half of Common Lisp."
>
> I've seen this in .signatures on USENET postings but I can no longer remember where I wrote it originally. On the Squeak discussion group, Vassili Bykov helpfully explained the meaning: "that complex systems implemented in low-level languages cannot avoid reinventing and/or reimplementing (poorly on both counts) the facilities built into higher-level languages. Like garbage collection in OLE or keyword arguments in X."
i understand.
something does not fit in all this.
but i don't know what.
ilias <·······@pontos.net> writes:
> Lars Brinkhoff wrote:
> > ilias <·······@pontos.net> writes:
> >
> >>Marco Antoniotti wrote:
> >>
> >>>Greenspun's Tenth applies. :)
> >>
> >>explain me this, please.
> > Greenspun's Tenth Rule of Programming:
> > http://philip.greenspun.com/research/
> thanx:
>
> > Quotes
> > Greenspun's Tenth Rule of Programming: "Any sufficiently complicated
> > C or Fortran program contains an ad-hoc, informally-specified
> > bug-ridden slow implementation of half of Common Lisp." I've seen
> > this in .signatures on USENET postings but I can no longer remember
> > where I wrote it originally. On the Squeak discussion group, Vassili
> > Bykov helpfully explained the meaning: "that complex systems
> > implemented in low-level languages cannot avoid reinventing and/or
> > reimplementing (poorly on both counts) the facilities built into
> > higher-level languages. Like garbage collection in OLE or keyword
> > arguments in X."
>
> i understand.
I don't.
> something does not fit in all this.
How can more mud fit in (into, or onto) a ball of mud?
> but i don't know what.
Do we?
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th 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'.
From: Eduardo Muñoz
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <ubs7ctlzo.fsf@jet.es>
ilias <·······@pontos.net> writes:
> > Quotes
> > Greenspun's Tenth Rule of Programming: "Any sufficiently complicated
> > C or Fortran program contains an ad-hoc, informally-specified
> > bug-ridden slow implementation of half of Common Lisp." I've seen
> > this in .signatures on USENET postings but I can no longer remember
> > where I wrote it originally. On the Squeak discussion group, Vassili
> > Bykov helpfully explained the meaning: "that complex systems
> > implemented in low-level languages cannot avoid reinventing and/or
> > reimplementing (poorly on both counts) the facilities built into
> > higher-level languages. Like garbage collection in OLE or keyword
> > arguments in X."
>
> i understand.
>
> something does not fit in all this.
>
> but i don't know what.
Just to confuse you further.
Greenspun is an ass that made himself rich by
overworking his employees for some time during the
dot.bomb era and later selling them to venture
capitalists. His own "tenth rule" was _explained_
to him by some other guy.
The whole story makes me want to puke.
--
Eduardo Mu�oz
"Eduardo Mu�oz" <···@jet.es> wrote in message ··················@jet.es...
>
> Greenspun is an ass that made himself rich by
> overworking his employees for some time during the
> dot.bomb era and later selling them to venture
> capitalists. His own "tenth rule" was _explained_
> to him by some other guy.
>
You know him personally?
From: Eduardo Muñoz
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <ud6rrvqmp.fsf@jet.es>
"Joe Marshall" <·············@attbi.com> writes:
> "Eduardo Mu�oz" <···@jet.es> wrote in message ··················@jet.es...
>
> >
> > Greenspun is an ass that made himself rich by
> > overworking his employees for some time during the
> > dot.bomb era and later selling them to venture
> > capitalists. His own "tenth rule" was _explained_
> > to him by some other guy.
> >
>
> You know him personally?
No. The quote about Common Lisp "Greenspun tenth"
and the story of ArsDigita made me curious so I
did a bit of research with google.com
A first glance he looks like a nice hacker but
after reading diferent sides of the story and
certain other quotes of him well I reached the
above conclusion.
--
Eduardo Mu�oz
From: Joe Marshall
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <ptvr5ief.fsf@ccs.neu.edu>
"Eduardo Mu�oz" <···@jet.es> writes:
> "Joe Marshall" <·············@attbi.com> writes:
>
> > "Eduardo Mu�oz" <···@jet.es> wrote in message ··················@jet.es...
> >
> > >
> > > Greenspun is an ass that made himself rich by
> > > overworking his employees for some time during the
> > > dot.bomb era and later selling them to venture
> > > capitalists. His own "tenth rule" was _explained_
> > > to him by some other guy.
> > >
> >
> > You know him personally?
>
> No...
Perhaps you would have a more charitable view if you did.
From: Eduardo Muñoz
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <u4rd3xl50.fsf@jet.es>
Joe Marshall <···@ccs.neu.edu> writes:
> "Eduardo Mu�oz" <···@jet.es> writes:
>
> > "Joe Marshall" <·············@attbi.com> writes:
> >
> > > "Eduardo Mu�oz" <···@jet.es> wrote in message ··················@jet.es...
> > >
> > > >
> > > > Greenspun is an ass that made himself rich by
> > > > overworking his employees for some time during the
> > > > dot.bomb era and later selling them to venture
> > > > capitalists. His own "tenth rule" was _explained_
> > > > to him by some other guy.
> > > >
> > >
> > > You know him personally?
> >
> > No...
>
> Perhaps you would have a more charitable view if you did.
Why?. Maybe you would like to explain your POV (by
mail if you will)?
--
Eduardo Mu�oz
In article <···············@octagon.mrl.nyu.edu>,
Marco Antoniotti <·······@cs.nyu.edu> wrote:
> ilias <·······@pontos.net> writes:
>
> ...
> >
> > For me it seems like they'll be very successfull. Cause they don't
> > have syntax-romance. That's business.
> >
> > What is your opinion?
>
> Greenspun's Tenth applies. :)
It seems to me that Greenspun's Tenth applies only to people who don't
*know* that they're implementing half of Common Lisp.
-- Bruce
Bruce Hoult wrote:
> It seems to me that Greenspun's Tenth applies only to people who don't
> *know* that they're implementing half of Common Lisp.
Is that a bit like invoking Hitler?
;)w
From: Tim Bradshaw
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <ey3vg5jxqpy.fsf@cley.com>
To: Bruce Hoult <·····@hoult.org>
Subject: Re: You want braces? I'll give you braces...
References: <················@nyc.rr.com> <················@pontos.net>
<·················@cs.uni-bonn.de> <················@pontos.net>
<···············@octagon.mrl.nyu.edu>
<···························@copper.ipg.tsnz.net>
FCC: ~/Net/outgoing/gnus-mails
From: Tim Bradshaw <···@cley.com>
Date: 06 Sep 2002 13:11:17 +0100
In-Reply-To: <···························@copper.ipg.tsnz.net>
Message-ID: <···············@cley.com>
Lines: 12
User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.1 (Cuyahoga Valley)
--text follows this line--
* Bruce Hoult wrote:
> It seems to me that Greenspun's Tenth applies only to people who don't
> *know* that they're implementing half of Common Lisp.
This must be Hoult's corollary.
Bradshaw's meta-corollary is that all of the results following from
Godwin's law also apply to Greenspun's tenth law after suitable
substitutions.
--tim
>>>>> "Bruce" == Bruce Hoult <·····@hoult.org> writes:
Bruce> It seems to me that Greenspun's Tenth applies only to people who don't
Bruce> *know* that they're implementing half of Common Lisp.
I do not agree.
Even if you know that you are extending you application with a selection
of features from Common Lisp, you are likely to get a slow, poorly
documented and bugridden result.
People who is able to appreciate the features of Common Lisp, and the
subleties of their implementation, are highly unlikely to even attempt to
do it.
In other words, anybody who thinks that choosing an inferior language is
ok because you can always implement the stuff from lisp you need on top of
it, has a serious deficit in the understanding of the implications of
Greenspuns 10. rule.
------------------------+-----------------------------------------------------
Christian Lynbech | Ericsson Telebit, Skanderborgvej 232, DK-8260 Viby J
Phone: +45 8938 5244 | email: ·················@ted.ericsson.se
Fax: +45 8938 5101 | web: www.ericsson.com
------------------------+-----------------------------------------------------
Hit the philistines three times over the head with the Elisp reference manual.
- ·······@hal.com (Michael A. Petonic)
In article <··············@situla.ted.dk.eu.ericsson.se>,
Christian Lynbech <·················@ted.ericsson.se> wrote:
> >>>>> "Bruce" == Bruce Hoult <·····@hoult.org> writes:
>
> Bruce> It seems to me that Greenspun's Tenth applies only to people who don't
> Bruce> *know* that they're implementing half of Common Lisp.
>
> I do not agree.
>
> Even if you know that you are extending you application with a selection
> of features from Common Lisp, you are likely to get a slow, poorly
> documented and bugridden result.
>
> People who is able to appreciate the features of Common Lisp, and the
> subleties of their implementation, are highly unlikely to even attempt to
> do it.
Does this include Paul Graham, with Arc?
What about David Moon et al, with Dylan?
Were these Lisp gurus unable to appreciate the features of Common Lisp
and the subleties of their implementation? I smell a circular argument.
-- Bruce
>>>>> "Bruce" == Bruce Hoult <·····@hoult.org> writes:
Bruce> Does this include Paul Graham, with Arc?
Bruce> What about David Moon et al, with Dylan?
I was unclear. I should of course have said "vastly inferior
programming language". Greenspuns rule relates to Fortran and the
like.
I do not think anybody (here at least) will object to the inclusion of
C, C++, C# and Java in the category of languages covered by Greenspuns
rule, but clearly there are languages that are close enough to Common
Lisp to avoid that particular problem.
------------------------+-----------------------------------------------------
Christian Lynbech | Ericsson Telebit, Skanderborgvej 232, DK-8260 Viby J
Phone: +45 8938 5244 | email: ·················@ted.ericsson.se
Fax: +45 8938 5101 | web: www.ericsson.com
------------------------+-----------------------------------------------------
Hit the philistines three times over the head with the Elisp reference manual.
- ·······@hal.com (Michael A. Petonic)
From: Kenny Tilton
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <3D776EB0.6010708@nyc.rr.com>
ilias wrote:
> Curl looks like what Paul Graham wants to make with his OpenSource-ARC.
> http://www.paulgraham.com/arc.html
What?!!! Curl is verbose and includes OO, both of which are explicit
design anti-goals of Arc.
>
> The curl code looks like 'mutated LISP'.
Indirectly. It is mutated Dylan, which is mutated Lisp.
>
> For me it seems like they'll be very successfull. Cause they don't have
> syntax-romance.
they could use some.
Mind you, I dig Curl for (a) being a member of the Lisp family and (2)
putting it in a Web page (ok, some other stuff too like the built-in 3d
stuff). I was just struck by how the syntax gets in my way as I read the
code.
kenny
From: ilias
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <3D777C4C.70405@pontos.net>
Kenny Tilton wrote:
>
>
> ilias wrote:
>
>> Curl looks like what Paul Graham wants to make with his OpenSource-ARC.
>> http://www.paulgraham.com/arc.html
>
>
> What?!!! Curl is verbose and includes OO, both of which are explicit
> design anti-goals of Arc.
i talk about the whole concept.
languge for web-design.
by the way: OO cannot be antigoal of Arc. It is object-oriented. With a
'different' meaning or with a 'extended' meaning.
verbose: no comment.
>> The curl code looks like 'mutated LISP'.
>
> Indirectly. It is mutated Dylan, which is mutated Lisp.
>
>> For me it seems like they'll be very successfull. Cause they don't
>> have syntax-romance.
>
> they could use some.
>
> Mind you, I dig Curl for (a) being a member of the Lisp family and (2)
> putting it in a Web page (ok, some other stuff too like the built-in 3d
> stuff). I was just struck by how the syntax gets in my way as I read the
> code.
>
> kenny
Lisp family - who cares?
They have no romance.
They are winners.
From: Bill Clementson
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <wk8z2f56ey.fsf@attbi.com>
ilias <·······@pontos.net> writes:
> Curl looks like what Paul Graham wants to make with his OpenSource-ARC.
> http://www.paulgraham.com/arc.html
Arc is a server-based solution, Curl is a client-based solution. I would
say that Curl is more similar (conceptually) to browser support for XSLT
in that both XSLT & Curl represent approaches to moving processing and
rendering logic to the client instead of the server.
FYI - the MITtechnology Insider electronic newsletter for September has
a front page review of Curl (language & company) that talks about why
neither have succeeded very well so far. If you're interested in reading
the article, you can get a 2-issue trial subscription (that will include
the Curl issue) from this site:
https://www.technologyinsider.com/subscribe/sub_soft_mitti.asp?s=A2JPS4
--
Bill Clementson
Bill Clementson wrote:
> ilias <·······@pontos.net> writes:
>
>
>>Curl looks like what Paul Graham wants to make with his OpenSource-ARC.
>>http://www.paulgraham.com/arc.html
>
> Arc is a server-based solution, Curl is a client-based solution. I would
> say that Curl is more similar (conceptually) to browser support for XSLT
> in that both XSLT & Curl represent approaches to moving processing and
> rendering logic to the client instead of the server.
i see. they may have to merge.
> FYI - the MITtechnology Insider electronic newsletter for September has
> a front page review of Curl (language & company) that talks about why
> neither have succeeded very well so far.
one reason i know: the menues on their web-pages are terribly slow.
http://www.curl.com
> If you're interested in reading
> the article, you can get a 2-issue trial subscription (that will include
> the Curl issue) from this site:
> https://www.technologyinsider.com/subscribe/sub_soft_mitti.asp?s=A2JPS4
i'm *very* interested and the free exemplar is on he way via e-mail.
happy: no credid-card needed.
i don't have.
From: Kenny Tilton
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <3D796444.2060105@nyc.rr.com>
Bill Clementson wrote:
> If you're interested in reading
> the article, you can get a 2-issue trial subscription (that will include
> the Curl issue) from this site:
> https://www.technologyinsider.com/subscribe/sub_soft_mitti.asp?s=A2JPS4
I hate to tease the marketing dept. :) Can you give a quick synopsis? My
guesses would be (a) dot.bomb killed their market and (ii) too hard to
learn and (3) not open source.
kenny
From: Kenny Tilton
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <3D7965DA.3050304@nyc.rr.com>
also i just worked out from a warning deep in some example that their GC
apparently works by reference counting; two cross-referencing but
otherwise unreachable objects will not get GCed.
ah, why can't I just have a nice CL plugin? :)
kenny
From: Bill Clementson
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <wkvg5i76ct.fsf@attbi.com>
Kenny Tilton <·······@nyc.rr.com> writes:
> also i just worked out from a warning deep in some example that their
> GC apparently works by reference counting; two cross-referencing but
> otherwise unreachable objects will not get GCed.
>
> ah, why can't I just have a nice CL plugin? :)
Life would be too easy then ;-)
--
Bill Clementson
From: Bill Clementson
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <wkznuu76fm.fsf@attbi.com>
Kenny Tilton <·······@nyc.rr.com> writes:
> Bill Clementson wrote:
> > If you're interested in reading
> > the article, you can get a 2-issue trial subscription (that will include
> > the Curl issue) from this site:
> > https://www.technologyinsider.com/subscribe/sub_soft_mitti.asp?s=A2JPS4
>
> I hate to tease the marketing dept. :) Can you give a quick synopsis?
The Curl vision was "to reinvent the World Wide Web as a more powerful
network through a programming language that would redistribute the
computing workload from over-burdened servers to underutilized client
PCs." The language was meant to be "a successor to HTML that could
express data with the power and flexibility of programming languages
such as C++ or list processing (LISP)."
> My guesses would be (a) dot.bomb killed their market and (ii) too hard
> to learn and (3) not open source.
Not quite, the article indicates that the primary reasons for their lack
of success were:
1. Bad business model: Free developer tools but per-user runtime
charges. When the web advertising bubble burst, this became "an
expensive, and risky, proposition".
2. New language: Hard to convince companies that their developers had to
learn a new programming language. Curl was meant to replace
html instead of simplify the generation of html and it required a
browser plug-in that users had to download.
Although the article doesn't explicitly say so, I would speculate that
your guesses for #a & #3 are also at least partially correct based on
other articles that I've read about Curl.
Curl (the company) is now under new management and they are attempting
to transform the company by doing away with per-user licenses (instead
introducing developer starter kits that start at about $25k - makes the
complaints about commercial Lisp vendor prices seem a bit embarrassing,
eh) and positioning Curl (the language) as a rapid web development tool
rather than as a replacement for html. They have also changed the goals
of the company from "taking over the internet" to just attempting to reach
profitability (common tactic in the post dot.bomb world).
--
Bill Clementson
From: Kenny Tilton
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <3D7A1C33.5000404@nyc.rr.com>
Thx for the detailed summary. I had polled a couple of Web pros re Curl.
One had heard of it (and was hearing more recently), another had not
heard of it. (She liked PHP.) The feeling I got was that Curl Inc had
addressed a problem (overloaded servers) others do not see.
Likewise for the richer Web page interactivity. By now folks think it is
normal for a Web page to just sit there, they don't perceive a problem.
And of course the problem they /do/ perceive, viz not going broke,
pushes adoption of new technology below radar.
My guess is Curl Inc also made the mistake of thinking a better
mousetrap would sell itself and ran thru their investment capital too
fast. Folks tend to get set in their ways; if one wants to sell
something new, one better be prepared for a long haul.
Look at Lisp. It will have taken close to fifty years for Lisp to
prevail when that happens in a year or two.
:)
kenny
Kenny Tilton <·······@nyc.rr.com> wrote:
> Thx for the detailed summary. I had polled a couple of Web pros re
> Curl. One had heard of it (and was hearing more recently), another had
> not heard of it. (She liked PHP.) The feeling I got was that Curl Inc
> had addressed a problem (overloaded servers) others do not see.
"Overloaded servers" have commonly been able to be addressed
assortedly by:
a) Adding great gobs of cheap RAM;
b) Adding more CPUs;
c) Adding more servers;
d) Fixing whatever aspect of the software or configuration caused it
to be desperately broken.
Just had a chat with my brother who's doing some Java architure work;
the _classic_ problem there is that Java servers wind up being poorly
understood and hence poorly configured.
He was trying to do some benchmarking of Apache "Struts" as compared
with some JBoss stuff; all suffered from horrible performance
problems, not particularly due to "open sourceness," but rather
because they're complex masses of software that are hard to use
optimally. The same will be true of other Java systems, and even of
Lisp systems.
The "Transmeta Crusoe" folk apparently thought they'd take the world
by storm by having a lower powered (electricity-wise) CPU for laptops.
They were apparently wrong, too. The CPU wasn't _that_ big a deal,
and _hasn't_ taken the world by storm. It's a really cool design;
"low power" just wasn't compelling enough, just as "overloaded
servers" aren't compelling Curl adoption.
> Likewise for the richer Web page interactivity. By now folks think
> it is normal for a Web page to just sit there, they don't perceive a
> problem.
There _has_ been a /holy grail/ of `local processing;' that's
assortedly where JavaScript, Java plugins, Tcl plugins (yes, that
exists!), and such have /tried/ to fit in. For a not-web-related set
of variations, consider the way that NeXTSTEP could throw Postscript
code at the DPS renderer, and that Sun NeWS did similar for handling
Unix graphics. There is _certainly_ a set of precedents for this.
If you want the _most_ pathological web-related case, consider the way
that Mozilla is basically a set of rendering engines with a horde of
XUL/XML and ECMAScript code running on top of it.
The latter is actually a /really very/ relevant thing to look at in
this regard, despite its klunkiness.
The point in the "dynamic/client-side rendering stuff" that typically
gets /missed/ is that what it amounts to is having a programmable
client-side *environment*. With Mozilla, the XUL 'language' gets used
to describe what sorts of graphical elements are available for
client-side processes to use. In that case, it's a pretty
sophisticated set of elements.
The huge /error/ is to think that it's merely about having a
programming language around. That's not /nearly/ sufficient. You
need to define not only a language, but the whole processing
environment, amounting to defining a 'virtual computer:'
-> There needs to be a CPU, typically handled via either a VM
(as with Java) or an interpreter (as with Tcl, JavaScript, Curl).
It doesn't stop there.
-> The 'memory environment' must be defined. This is usually pretty
easy, as interpreters and VMs can easily create that.
-> There needs to be some defined interfaces to get at the client's
environment, to files, perhaps the network, and such. It needs to
be secured so that "malware" can't come in and do damage.
-> There needs to be a defined GUI, if it is to allow users to
interact with it.
What the Java experience tells us about the 'defining an environment'
process is simple:
_This Means War_
- Sun wanted to define these things /their/ way;
- Microsoft didn't like this, from whence IE and J++ warring came;
- Sun spent money to draw in the Tcl group, which quite nicely
hobbled the likelihood of Tcl influencing any 'de facto standards;'
- Netscape made up Netscape One as a would-be platform;
- Mozilla redid this;
- There are, more than likely, lots of others, including Curl.
The /political/ warring over this has quite nicely fragmented the
browser market to the point that about the nearest thing to 'a
platform' is to write applications that tightly integrate with
Internet Explorer. (Look at Quicken and Intuit's tax products for
examples of this!)
The notion that Curl could have threaded their way through all this to
"define the platform" is _really_ wishful thinking.
> And of course the problem they /do/ perceive, viz not going broke,
> pushes adoption of new technology below radar.
> My guess is Curl Inc also made the mistake of thinking a better
> mousetrap would sell itself and ran thru their investment capital
> too fast. Folks tend to get set in their ways; if one wants to sell
> something new, one better be prepared for a long haul.
Curl Inc may be doing something "different" in using Lisp-like
technology unlike the Java/XML/VB stuff pointed to above, but from the
perspective of users or purchasors or patrons that _won't be writing
Curl code_, it's no different from, say, the Tcl web browser plugin.
> Look at Lisp. It will have taken close to fifty years for Lisp to
> prevail when that happens in a year or two.
>
> :)
Hee.
--
(concatenate 'string "cbbrowne" ·@ntlug.org")
http://cbbrowne.com/info/internet.html
"The problem with the cutting edge is that someone has to bleed."
-- Zalman Stern
From: Bill Clementson
Subject: Re: You want braces? I'll give you braces...
Date:
Message-ID: <wkd6rpiiys.fsf@attbi.com>
Christopher Browne <········@acm.org> writes:
> The point in the "dynamic/client-side rendering stuff" that typically
> gets /missed/ is that what it amounts to is having a programmable
> client-side *environment*. With Mozilla, the XUL 'language' gets used
> to describe what sorts of graphical elements are available for
> client-side processes to use. In that case, it's a pretty
> sophisticated set of elements.
>
> The huge /error/ is to think that it's merely about having a
> programming language around. That's not /nearly/ sufficient. You
> need to define not only a language, but the whole processing
> environment, amounting to defining a 'virtual computer:'
>
> -> There needs to be a CPU, typically handled via either a VM
> (as with Java) or an interpreter (as with Tcl, JavaScript, Curl).
>
> It doesn't stop there.
>
> -> The 'memory environment' must be defined. This is usually pretty
> easy, as interpreters and VMs can easily create that.
>
> -> There needs to be some defined interfaces to get at the client's
> environment, to files, perhaps the network, and such. It needs to
> be secured so that "malware" can't come in and do damage.
>
> -> There needs to be a defined GUI, if it is to allow users to
> interact with it.
This has been a problem for years with with "solutions" that attempt to
address PC browser interactivity via language or plug-in
enhancements. It is only going to get worse as the proliferation of
non-PC clients increases. I don't see any new language or plug-in being
able to grab market share in this scenario.
One of the most promising initiatives (IMHO of course) is the WSIA/WSRP
standards work that Oasis is currently doing. Instead of trying to
create an HTML replacement, they are trying to create a standard way of
representing what an application expects in terms of gui representation,
gui navigation, event handling, etc. for the purposes of providing a
gui-based application as a web service. For example, Amazon currently
provides a SOAP-based interface to their book lookup service that you
could access as a web service from your web site. However, if you want
to order one of the books, you need to directly link to their shopping
cart, losing "your" branding and navigation consistency for what Amazon
provides. If Amazon provided a WSIA/WSRP definition of their shopping
cart
as a web service, you could embed their shopping cart in your
application yet have custom code around their application that allowed
you to add your own "specials", "deals", etc. In addition, you would
maintain the branding of your site and the navigation would be
consistent. Because the Amazon shopping cart definition isn't
language-specific, you could be rendering it in HTML, Curl, XML/XSLT,
SVG, WAP/WML or whatever on the client. This is a win for both Amazon
(who gets the book orders) and you (who provide a richer web site with
increased functionality without the end-user knowing where the
functionality came from).
Although WSRP/WSIA is still very much in the early stages, I think it
holds greater potential for the future than custom browser plug-ins or
proprietary browser environments. There will always be a place for
javascript (and alternatives) on the client; however, most of the real
action will happen on the server.
--
Bill Clementson
Bill Clementson wrote:
>>> If you're interested in reading
>>>the article, you can get a 2-issue trial subscription (that will include
>>>the Curl issue) from this site:
>>>https://www.technologyinsider.com/subscribe/sub_soft_mitti.asp?s=A2JPS4
MIT technology insider - Issue 9.02
Curl's Winding Road
I've read the article.
-
Michael L. Dertouzos.
Greek. A compatriot.
Makes me proud.
Chief of the MIT Laboratory for Computer Science.
One day we will talk together.
In our native language.
Makes me happy.
Proud, happy. Feel. Feelings.
I try to control these feelings with logic.
I'm human. Nationality irrelevant. Equality.
He's human. Positions irrelevant. Equality.
Of course I fail.
.
Left of the article, a small column.
[...]
TECHNICAL FOUNDERS
[...]
Michael L. Dertouzos (d. 2001)
[...]
"(d. 2001)" - ?
I've searched the whole issue.
No further remark.
Only "(d. 2001)"
.
My happiness - is gone.
The falling tears.
So painful.
...
The falling rain.
So beautiful.
Michael Leonidas Dertouzos.
He is a part of the falling rain.
As i will become part of the falling rain.
And then we will talk together.
In our new native language.
The unique language of the falling rain.
.
Equality.
.
-
http://www.lcs.mit.edu/news/dertouzos.html