From: Alain Picard
Subject: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <86ofw1wotf.fsf@optushome.com.au>
Dear Lispers,
The company at which I work is currently speaking with
potential investors. These investors would like to know
why we are using Lisp to implement our product, and,
who else uses it, and for what applications.
We have explained to them the technical advantages of Lisp
(stability, programmer productivity, cross-platform development, etc)
and while they appreciate these reasons, they wonder why they
have not heard of anyone else using it.(*)
So, what do you use Lisp for? Real life stories of
successful systems built in Lisp would be greatly appreciated.
I think this subject is of general interest, (I certainly wanted to
know of such endeavours before *we* chose to go with Lisp) so post to
the group if possible, but if not, anything you send me via e-mail
will be considered confidential(*).
Oh, I know about the lists maintained on www.franz.com,
I am asking for your personal testimonials and estimation
of how successful lisp has been as a choice of technology
in your commercial endeavour.
Alain Picard
(*) My theory is that companies using Lisp keep it under wraps,
because they consider it a strategic advantage (to outcompete
the C++/Java crowd). That is why you cannot discuss this publicly,
I assure you that any communications will be held in confidentiality.
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
In article <··············@optushome.com.au>, Alain Picard
<·······@optushome.com.au> wrote:
[snip]
>So, what do you use Lisp for? Real life stories of
>successful systems built in Lisp would be greatly appreciated.
[snip]
Carnegie Learning (http://www.carnegielearning.com/) develops and sells
computer tutors for a variety of elementary, middle, and secondary school
subjects. Their tutors and tutor development systems are an outgrowth of
basic research done at Carnegie Mellon. Their systems have traditionally
been built in Common Lisp, MCL initially and then Franz when they started
developing for the PC platform. As I understand it, they have also been
using Java of late. You should contact them for whatever details they're
willing to provide.
Sashank
On Sat, 17 Feb 2001 10:46:14 GMT, Alain Picard <·······@optushome.com.au>
wrote:
>
>Dear Lispers,
>
>
>The company at which I work is currently speaking with
>potential investors. These investors would like to know
>why we are using Lisp to implement our product, and,
>who else uses it, and for what applications.
Viaweb seems like a good example to convince investors. :-)
//-----------------------------------------------
// Fernando Rodriguez Romero
//
// frr at mindless dot com
//------------------------------------------------
hey,
I'm pretty sure that HotDispatch.com uses it. They have an awesome
site -- one of those in the same class as guru.com, where they broker
work. I believe they use MCL, and CL-HTTP. Please -- I've only heard
this from at least two independent sources, so it should probably be
verified. But I've used HotDispatch.com, and it's great.
dave
They seem to be hiring Java/Corba/XML people:
http://corporate.hotdispatch.com/current-openings.html
David Bakhash wrote:
>
> hey,
>
> I'm pretty sure that HotDispatch.com uses it. They have an awesome
> site -- one of those in the same class as guru.com, where they broker
> work. I believe they use MCL, and CL-HTTP. Please -- I've only heard
> this from at least two independent sources, so it should probably be
> verified. But I've used HotDispatch.com, and it's great.
>
> dave
From: Eugene Zaikonnikov
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <6y8zn5xjwz.fsf@viking.cit>
David> hey, I'm pretty sure that HotDispatch.com uses it. They have
David> an awesome site -- one of those in the same class as guru.com,
David> where they broker work. I believe they use MCL, and CL-HTTP.
I heard they used Lispworks, CL-HTTP and Oracle as a database
backend. I also head that they are about to move to Java (after all,
HotDispatch is a Java zoo), and, considering that recently I observed
Apache error message when connecting to there, probably they moved
already.
--
Eugene
So you use Lisp to get something working quickly, raise capital, and
then spend the money paying people to ossify it in Java. But I'd only
put the first half in the business plan ;-)
Eugene Zaikonnikov wrote:
>
> David> hey, I'm pretty sure that HotDispatch.com uses it. They have
> David> an awesome site -- one of those in the same class as guru.com,
> David> where they broker work. I believe they use MCL, and CL-HTTP.
>
> I heard they used Lispworks, CL-HTTP and Oracle as a database
> backend. I also head that they are about to move to Java (after all,
> HotDispatch is a Java zoo), and, considering that recently I observed
> Apache error message when connecting to there, probably they moved
> already.
>
> --
> Eugene
From: ······@comac.com
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <96mplh$s69$1@news.netmar.com>
In article <··············@optushome.com.au>, Alain Picard
<·······@optushome.com.au> writes:
>
>Dear Lispers,
>
>
>The company at which I work is currently speaking with
>potential investors. These investors would like to know
>why we are using Lisp to implement our product, and,
>who else uses it, and for what applications.
>
>We have explained to them the technical advantages of Lisp
>(stability, programmer productivity, cross-platform development, etc)
>and while they appreciate these reasons, they wonder why they
>have not heard of anyone else using it.(*)
>
>So, what do you use Lisp for? Real life stories of
>successful systems built in Lisp would be greatly appreciated.
>
>I think this subject is of general interest, (I certainly wanted to
>know of such endeavours before *we* chose to go with Lisp) so post to
>the group if possible, but if not, anything you send me via e-mail
>will be considered confidential(*).
>
>Oh, I know about the lists maintained on www.franz.com,
>I am asking for your personal testimonials and estimation
>of how successful lisp has been as a choice of technology
>in your commercial endeavour.
>
>
> Alain Picard
>
>
>(*) My theory is that companies using Lisp keep it under wraps,
> because they consider it a strategic advantage (to outcompete
> the C++/Java crowd). That is why you cannot discuss this publicly,
> I assure you that any communications will be held in confidentiality.
>
>
>--
>It would be difficult to construe Larry Wall, in article
>this as a feature. <·····················@netlabs.com>
Yahoo!Stores is built in Lisp
CommerceOne uses it in one aspect of their business
You will find others at http://www.lisp.org http://www.franz.com etc.
Alex
----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web -----
http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups
NewsOne.Net prohibits users from posting spam. If this or other posts
made through NewsOne.Net violate posting guidelines, email ·····@newsone.net
······@comac.com writes:
> Yahoo!Stores is built in Lisp
I find this surprising, given that it's Yahoo!, unless they went with
a non-commercial implementation.
dave
From: Bulent Murtezaoglu
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <87r90tj6ob.fsf@kapi.internal>
DB> [...]I find this surprising, given that it's Yahoo!, unless they
DB> went with a non-commercial implementation.
They bought viaweb which was Paul Graham's company. I think the original
code was developed with Allegro.
BM
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjofvwnx8r.fsf@tfeb.org>
Bulent Murtezaoglu <··@acm.org> writes:
> They bought viaweb which was Paul Graham's company. I think the original
> code was developed with Allegro.
>
CLISP according to Paul Graham in 1998.
--tim
From: Bulent Murtezaoglu
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <877l2khy7y.fsf@kapi.internal>
>> [me:] I think
>> the original code was developed with Allegro.
>>
TimB> CLISP according to Paul Graham in 1998.
I stand corrected. I seem to have assumed that since Jim Veitch
inteviewed him for CACM about his use of Lisp, he must have used Allegro.
Wrong inference. To redeem myself I'll provide the reference ($$ to
get pdf from acm.org if not member):
A conversation with Paul Graham ; Jim Veitch; Commun. ACM 41, 5 (May. 1998),
Pages 52 - 54.
cheers,
BM
David Bakhash <·····@alum.mit.edu> writes:
> ······@comac.com writes:
>
> > Yahoo!Stores is built in Lisp
>
> I find this surprising, given that it's Yahoo!, unless they went with
> a non-commercial implementation.
>
> dave
According to this page, they use CLISP:
http://www.podval.org/~sds/tool.html
Peter
From: Howard Stearns
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A93EB05.8DF36C2@curl.com>
Erik Naggum wrote:
>
> * David Bakhash
> > I find this surprising, given that it's Yahoo!, unless they went with a
> > non-commercial implementation.
>
> If memory serves, Yahoo!Stores made Paul Graham's fortune and uses CLISP.
> The Lisp software generates static web pages and is not involved in
> serving them to the users.
My recollection/understanding is slightly different (and could as easily
be wrong):
The "editor" used for creating and customizing stores uses (used???)
CLISP to generate HTML on the fly -- i.e., Lisp "serves" the pages.
I do not recollect any info on whether shoppers (end users of the
stores) see pages that are served by Lisp.
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjpugc55oj.fsf@tfeb.org>
Howard Stearns <·······@curl.com> writes:
>
> I do not recollect any info on whether shoppers (end users of the
> stores) see pages that are served by Lisp.
I'nm almost sure they do not -- files are generated and then served as
static pages. He makes a big point somewhere that static pages, or
things that look like static pages, are a win as indexing systems
like them.
--tim
In article <···············@tfeb.org>, Tim Bradshaw wrote:
>Howard Stearns <·······@curl.com> writes:
>>
>> I do not recollect any info on whether shoppers (end users of the
>> stores) see pages that are served by Lisp.
>
>I'nm almost sure they do not -- files are generated and then served as
>static pages. He makes a big point somewhere that static pages, or
>things that look like static pages, are a win as indexing systems
>like them.
>
>--tim
>
>
From what I read about it it was a good old fashioned batch job.
Every so often it would just rebuild the site and publish it.
And static pages with lots of text are a big win when the webbots
come by to get your stuff, if it looks like a script they will
probably ignore it. There is an interesting article about this
somewhere on arsdigita.com, no url handy.
marc
From: Tim Lavoie
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <44Vk6.852$vh7.29894@news1.mts.net>
In article <····················@oscar.eng.cv.net>, Marc Spitzer wrote:
>From what I read about it it was a good old fashioned batch job.
>Every so often it would just rebuild the site and publish it.
>And static pages with lots of text are a big win when the webbots
>come by to get your stuff, if it looks like a script they will
>probably ignore it. There is an interesting article about this
>somewhere on arsdigita.com, no url handy.
For what it's worth, the open-source Zope app server has nice,
static-looking URLs, so much of the info looks like a path instead. On the
other hand, you have to watch out for the default inheritance, as relative
paths can get extended forever, so a bot may crawl /foo/bar, /foo/foo/bar,
/foo/foo/foo/bar ....
Neat system though, and mostly in Python.
From: Will Deakin
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A953442.6030903@pindar.com>
Tim Bradshaw wrote:
> I'nm almost sure they do not -- files are generated and then served as
> static pages. He makes a big point somewhere that static pages, or
> things that look like static pages, are a win as indexing systems
> like them.
Is point five big enough? on www.paulgraham.com/mistakes.html,
:)w
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkj1ysqiu2r.fsf@tfeb.org>
Will Deakin <········@pindar.com> writes:
>
> Is point five big enough? on www.paulgraham.com/mistakes.html,
>
.5 is way too small, at least a few hundred I should think.
From: Rob Warnock
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <97o04v$7r30e$1@fido.engr.sgi.com>
Will Deakin <········@pindar.com> wrote:
+---------------
| Tim Bradshaw wrote:
| > I'nm almost sure they do not -- files are generated and then served as
| > static pages. He makes a big point somewhere that static pages, or
| > things that look like static pages, are a win as indexing systems
| > like them.
|
| Is point five big enough? on www.paulgraham.com/mistakes.html,
+---------------
Philip Greenspun shows one antidote to this in his book "Philip and
Alex's Guide to Web Publishing", "Chapter 7: Publicizing Your Site
(Without Irritating Everyone on the Net)" <URL:http://www.arsdigita.com/
books/panda/publicizing>, in the section named "Hiding Your Content
from Search Engines (By Mistake)". Since the search engines won't
index URLs like "http://photo.net/bboard/fetch-msg.tcl?msg_id=000037",
he chose to:
Write another AOLServer TCL program that presents all
the messages from URLs that look like static files, e.g.,
"/fetch-msg-000037.html" and point the search engines to a
huge page of links like that. The text of the Q&A forum postings
will get indexed out of these pseudo-static files and yet I can
retain the user pages with their *.tcl URLs.
-Rob
-----
Rob Warnock, 31-2-510 ····@sgi.com
SGI Network Engineering <URL:http://reality.sgi.com/rpw3/>
1600 Amphitheatre Pkwy. Phone: 650-933-1673
Mountain View, CA 94043 PP-ASEL-IA
From: Bulent Murtezaoglu
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <87y9uogqp8.fsf@kapi.internal>
RW> [...] <URL:http://www.arsdigita.com/ books/panda/publicizing>,
RW> in the section named "Hiding Your Content from Search Engines
RW> (By Mistake)". Since the search engines won't index URLs like
RW> "http://photo.net/bboard/fetch-msg.tcl?msg_id=000037", [...]
And lots of people believe that. I don't know how true that statement
is anymore, as the search
http://www.google.com/search?q=philip+site%3Awww.arsdigita.com&lr=&safe=off
is certainly getting a lot of hits of the kind he mentions. Maybe
the engines are now using http "expires" info instead of guessing by URL.
The .tcl and .adp URL's I clicked from the list returned by google for that
query did not provide any modification date or expiration info.
cheers,
BM
From: Chris Double
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <wk4rxo7gwm.fsf@double.co.nz>
David Bakhash <·····@alum.mit.edu> writes:
> ······@comac.com writes:
>
> > Yahoo!Stores is built in Lisp
>
> I find this surprising, given that it's Yahoo!, unless they went with
> a non-commercial implementation.
Briefly mentioned here:
http://www.ai.mit.edu/projects/dynlangs/Talks/server-based-lisp.htm
Chris.
--
http://www.double.co.nz/cl
From: Kent M Pitman
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <sfwg0hc26dy.fsf@world.std.com>
Alain Picard <·······@optushome.com.au> writes:
> (*) My theory is that companies using Lisp keep it under wraps,
> because they consider it a strategic advantage (to outcompete
> the C++/Java crowd). That is why you cannot discuss this publicly,
> I assure you that any communications will be held in confidentiality.
They also keep it under wraps because it got some inappropriately bad
press back in the mid to late 1980's (the beginning of the so-called
"AI Winter", when the every-three-year cycle of rampant investing in
the latest cool thing shifted from AI to other endeavors, like flashy
graphics). Some management failures (e.g., at MCC) were blamed on the
choice of Lisp as a programming language. (I said then and I still
think now, "Oh, right. If only they'd done that large AI project in
C++, it'd be winning away now. Ha!" I seriously doubt the language
was the reason for whatever failures, but it became fashionable for
various failing AI projects to blame their failure on Lisp.) I think
vendors (possibly even legitimately) fear that their customers will get
weirded out by the presence of Lisp inside, and so simply choose not to
mention it. (I'm not saying the customer fear is based in something
solid, just that it's a real fear, and that can still affect sales..)
Add to this the fact that Lisp has a Large Footprint(TM). That is, it
takes up MANY megabytes. I say "(TM)" to underscore the fact that
there are some factoids (or false-oids) that get enshrined as lemmas
and not re-evaluated with time. Lisp WAS big back in the early 90's
but it so panicked customers that Lisp images have hardly grown in
size for the last decade. During the same interval, every release of
every other language has grown by leaps and bounds. A Lisp image is
now competitive with, and often tiny, compared to images of other
languages. The difference is often a distribution of .dll's that make
it hard to see the size of the other things. And even there, the
commercial vendors have dll solutions that do similar tricks. So this
is just a myth perpetuated by people not going back to find out what the
real truth is.
And, finally, some people use Lisp for prototyping and then deploy in
some other language. This isn't to say the other language could have
been used at all to evolve the concept. And Lisp rarely gets the
benefit of any noise about that. People mostly chalk this up to "Lisp
didn't make the grade and had to be replaced at the last minute" but
it often isn't about that at all. It's pretty plain that Lisp is not
a "commodity" language; there are more programmers out there for Java
or C++, and they are cheaper--as are all things that are bought in a
mass market fashion. If Lisp programmers were as plentiful as Java or
C++ programmers, they might be as cheap, too. But a lot of it comes
down to what a job shop will tolerate. And if a company buy's
another's technology, it may insist on a refitting of the product into
another language "just for good measure".
I do think ultimately that language choice should be a private matter and
that the vendors should work as hard as possible to reduce the places
where the choice of language matters. Often it does become an all-or-nothing
thing, either to allow programmers to mix around or because two modules in
the wrong languages won't connect. And that ends up with people making
"C++ shops" or "Java shops", eschewing all else. That's a pity.
Btw, in the space of material to convince someone of Lisp's technical
value, if you haven't seen my Lisp Pointers article on why Lisp is
good for Rapid Prototyping, I think it's at:
http://world.std.com/~pitman/PS/Hindsight.html
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <ey3bss0glk8.fsf@tfeb.org>
* Kent M Pitman wrote:
> It's pretty plain that Lisp is not
> a "commodity" language; there are more programmers out there for Java
> or C++, and they are cheaper--as are all things that are bought in a
> mass market fashion. If Lisp programmers were as plentiful as Java or
> C++ programmers, they might be as cheap, too. But a lot of it comes
> down to what a job shop will tolerate. And if a company buy's
> another's technology, it may insist on a refitting of the product into
> another language "just for good measure".
I think there are other points about this -- lisp programmers are
*seriously* hard to find, and often when you do find one they turn out
to be really ex-academics with an AI background who have a superb
combination of cluelessness about producing maintainable &
comprehensible code, arrogance about their ability, and often they
just aren't very good programmers at all, because academia just
doesn't think mundane craft skills like programming are in any way
important. This leads to friction when trying to work with them
because their heads are just so vast; they also produce terrible code,
and here Lisp hurts you because it's possible to get Lisp systems to
run which are so badly-written that they simply would not work at all
in many other languages because they'd spend their entire life
following dead pointers &c. So you end up with awful, marginally
maintainable code, difficult programmers *and* huge problems hiring
people, so you can get held to ransom by the people who produced the
stuff. This is a bad situation.
It's now being made worse for many people because -- I presume because
the Lisp market is small and thus the lisp vendors can't afford to
invest -- there's a serious shortage of Lisp systems which support the
kind of hardware that people want to use. For instance there is, to
my knowledge, no commercial system which supports the most common
`enterprise' 64bit platform (Sun), and no commercial system which
supports any multiprocessor system at all. This is really frightening
to people with multithreaded applications whose data are creeping up
towards the 2Gb/4Gb point.
I guess this looks like several reasons not to use Lisp, and it really
is, unfortunately. It would be nice to say that this all myth and
lore, but I'm working for a client right now which is in this exact
situation, and I've seen enough other people end up in this bad place
that I think that it's a reasonably common scenario.
Of course I'm only giving one side of the story here, there are a lot
of good reasons *to* use Lisp too, and many of the problems I mention
above can be resolved -- for instance it's pretty easy to train people
to write reasonable Lisp if their brains haven't been rotted by too
long in academia, so once you have one good Lisp person you can
cultivate others, and so on.
--tim
From: Craig Brozefsky
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <87wvan63sj.fsf@piracy.red-bean.com>
Tim Bradshaw <···@tfeb.org> writes:
> I think there are other points about this -- lisp programmers are
> *seriously* hard to find, and often when you do find one they turn
> out to be really ex-academics with an AI background who have a
> superb combination of cluelessness about producing maintainable &
> comprehensible code, arrogance about their ability, and often they
> just aren't very good programmers at all, because academia just
> doesn't think mundane craft skills like programming are in any way
> important.
My employer has gone thru several hiring rounds looking for Lisp
programmers and we did not find them seriously hard to find. We
actually didn't get any of the academics you talk about. Our core
team is all non-AI people, in fact none of our current programmers
have degrees in anything related to CompSci, most are musicians (I'm
the exception) and in general it's a very craft oriented team. The
two coders we hired as contractors to help us out had degrees, one was
an experienced Lisp Machine hacker, and the other was a UofCHicago
grad who was pretty much a lisp newbie but learned quick.
--
Craig Brozefsky <·····@red-bean.com>
In the rich man's house there is nowhere to spit but in his face
-- Diogenes
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <ey33ddbhaqf.fsf@cley.com>
* Craig Brozefsky wrote:
> My employer has gone thru several hiring rounds looking for Lisp
> programmers and we did not find them seriously hard to find. We
> actually didn't get any of the academics you talk about. Our core
> team is all non-AI people, in fact none of our current programmers
> have degrees in anything related to CompSci, most are musicians (I'm
> the exception) and in general it's a very craft oriented team. The
> two coders we hired as contractors to help us out had degrees, one was
> an experienced Lisp Machine hacker, and the other was a UofCHicago
> grad who was pretty much a lisp newbie but learned quick.
This may be a UK/US thing I guess. Certainly in the UK it's *very*
hard to find anyone.
--tim
From: Will Deakin
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A910026.3030701@pindar.com>
Erik Naggum wrote:
> Other that that, I seem to recall Margaret Thatcher killing Lisp and AI
> in the UK with a few immensely harsh blows. Fairly annoying that she'd
> go and do that.
Yup. To add to the list of all those other intangibles that her
government had a go at killing (like manufacturing industry,
higher education, the NHS, the railway...)
:(will
From: Douglas T. Crosher
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A8FE299.E75B749E@scieneer.com>
Tim Bradshaw wrote:
...
> It's now being made worse for many people because -- I presume because
> the Lisp market is small and thus the lisp vendors can't afford to
> invest -- there's a serious shortage of Lisp systems which support the
> kind of hardware that people want to use. For instance there is, to
> my knowledge, no commercial system which supports the most common
> `enterprise' 64bit platform (Sun), and no commercial system which
> supports any multiprocessor system at all. This is really frightening
> to people with multithreaded applications whose data are creeping up
> towards the 2Gb/4Gb point.
Development of a multi-threaded implementation is underway. The
immediate target is x86/linux. We are looking for a small number of
developers that would be interested in evaluating and providing
feedback on the implementation, so please contact me if interested.
Regards
Douglas Crosher
Director
Scieneer Pty Ltd
Melbourne Australia
"Tim Bradshaw" <···@tfeb.org> wrote in message
····················@tfeb.org...
> * Kent M Pitman wrote:
> It's now being made worse for many people because -- I presume because
> the Lisp market is small and thus the lisp vendors can't afford to
> invest -- there's a serious shortage of Lisp systems which support the
> kind of hardware that people want to use. For instance there is, to
> my knowledge, no commercial system which supports the most common
> `enterprise' 64bit platform (Sun), and no commercial system which
> supports any multiprocessor system at all. This is really frightening
> to people with multithreaded applications whose data are creeping up
> towards the 2Gb/4Gb point.
What is the status of projects like CLISP, do they support SMP?
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjg0hakgds.fsf@tfeb.org>
"Joshua Boyd" <······@cs.millersville.edu> writes:
>
> What is the status of projects like CLISP, do they support SMP?
No.
--tim
* Kent M Pitman wrote:
> > It's pretty plain that Lisp is not a "commodity" language; there
> > are more programmers out there for Java or C++, and they are
> > cheaper--as are all things that are bought in a mass market
> > fashion. If Lisp programmers were as plentiful as Java or C++
> > programmers, they might be as cheap, too.
Really? I didn't know that CL people were so expensive. The supply
of Java people might be massively more, but so is the demand.
> I think there are other points about this -- lisp programmers are
> *seriously* hard to find, and often when you do find one they turn
> out to be really ex-academics with an AI background who have a
> superb combination of cluelessness about producing maintainable &
> comprehensible code, arrogance about their ability, and often they
> just aren't very good programmers at all, because academia just
> doesn't think mundane craft skills like programming are in any way
> important.
that's true. That's why it's better to hire people who have used CL in
industry successfully than those who just used it in academia.
> This leads to friction when trying to work with them because their
> heads are just so vast; they also produce terrible code, and here
> Lisp hurts you because it's possible to get Lisp systems to run
> which are so badly-written that they simply would not work at all in
> many other languages because they'd spend their entire life
> following dead pointers &c. So you end up with awful, marginally
> maintainable code, difficult programmers *and* huge problems hiring
> people, so you can get held to ransom by the people who produced the
> stuff. This is a bad situation.
yes. but easily avoided. I know the kinds of programmers that you're
talking about, but this is only a problem for companies that just hire
anyone who has Lisp on their resume.
As for the ego thing...yes. Common Lisp programmers have quite a bit
of ego.
Sometimes I think of the religiosity and ego of Common Lisp
programmers, and how I'm part of it. I boast of the excellence of the
language more than of my own abilities, but programmers often
represent their choice languages, and if you're boasting that the
language is better, and you're using it while few others are, then
you're implicitly saying "I'm better" -- at least from some people's
perspective. I think this is a bad thing. But opinions are
legitimately important, and mine is that CL is better for many
problems, and so the best I can do is to simply state that "I feel I
am using a better tool for the job."
Note that in academia, many students just use what they know, often
don't care so much about speed, and only care about minimal working
order. If they made it work in Common Lisp, then it just sounds
cooler sometimes than if it were written in C. But yeah -- often
it'll be just as sloppy, much slower, and virtually unreadable. They
might think that most people would say "Wow! Common Lisp. That must
be intense." while they know that it was just one big shortcut. If
their egos are indeed inflated, it's because they actually managed to
fool others. Bottom line is that different programming tasks are
judged by different metrics, which have lots of variables:
o speed
o startup speed
o memory resources
o cost (i.e. for development tools)
o implementation time
o code readability
o code extensibility
o platform independence
o flexibility -- ability to change aspects of the design
o LOC
o usability by other systems (i.e. as libraries, servers, etc.)
> Of course I'm only giving one side of the story here, there are a
> lot of good reasons *to* use Lisp too, and many of the problems I
> mention above can be resolved -- for instance it's pretty easy to
> train people to write reasonable Lisp if their brains haven't been
> rotted by too long in academia, so once you have one good Lisp
> person you can cultivate others, and so on.
rotting in academia are some strong words. I've seen people rot in
industry too, and really badly. I know people who have been living in
gdb (doing C) for years now at his company. So how is this better
than academia, where at least interesting ideas are often being
developed and attempted.
It's just that academia doesn't have the bottom-line "does it cut it"
specs. I recall some results from early graduate work that were very
good, but which required my Lisp program (using ACL) on a Sun to crank
for several seconds to recognize a single target, where useful
commercial implementation would be completely worthless and
unmarketable unless it could do the same in a few hundred
milliseconds. But for the sake of getting results, and trying new
methods, it was interesting nonetheless. When it comes to research, I
think that a lot of interesting ideas might not be attemped because of
how financially unvaluable they are.
dave
Tim Bradshaw <···@tfeb.org> writes:
> I think there are other points about this -- lisp programmers are
> *seriously* hard to find, and often when you do find one they turn out
My impression is that it's much harder to find perl and java programmers
*with a clue*. And if you find them, it's even harder to keep them.
> supports any multiprocessor system at all. This is really frightening
> to people with multithreaded applications whose data are creeping up
> towards the 2Gb/4Gb point.
Yes, but are there actually any kind of system with automatic memory
management (except hand-craftet ones) which is usable in this range?
> Of course I'm only giving one side of the story here, there are a lot
> of good reasons *to* use Lisp too, and many of the problems I mention
> above can be resolved -- for instance it's pretty easy to train people
> to write reasonable Lisp if their brains haven't been rotted by too
> long in academia, so once you have one good Lisp person you can
> cultivate others, and so on.
Your attitude towards academia astounds me. Yes: I *do* know what you
mean, but on the other side it's equally easy to rot in industry and
end up as a buzzword-repeating device.
--
(espen)
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> Your attitude towards academia astounds me. Yes: I *do* know what you
> mean, but on the other side it's equally easy to rot in industry and
> end up as a buzzword-repeating device.
a little add-on: I think the thing is that some people are Natural
Born Hackers and some are not. And some of the talented programmers
also have theoretical skills, and some have not. Unfortunately, there are
much too many people working as programmers and system architects out
there that have neither theoretical nor programming skills. And these
typically quickly evolve into *very* annoying buzzword repeating
leaders. The Dilbert principle at work.
--
(espen)
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjelwukfs4.fsf@tfeb.org>
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> Your attitude towards academia astounds me. Yes: I *do* know what you
> mean, but on the other side it's equally easy to rot in industry and
> end up as a buzzword-repeating device.
yes, of course it is, although academia seems to be pretty anxious to
compete here too -- the Lisp course at the place I used to work got
eventually nuked to make room for Java. I guess my attitude towards
academia (or rather towards the CS/AI/CogSci bits of it) comes from
having spent 10 years working in it (not as an academic), so I'm
probably biased.
--tim
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjy9v27qk3.fsf@tfeb.org>
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
>
> Yes, but are there actually any kind of system with automatic memory
> management (except hand-craftet ones) which is usable in this range?
>
If you mean in terms of memory, then I guess Franz's 64bit Lisps are
(I don't really see any reasons they should not be anyway). If you
mean in terms of multiprocessors then I don't know, probably not,
altough there are Boehm-GC derivatives which make noises about this,
and there were recently articles here (I think) about Java systems
which looked like they would usefully run on multiprocessors. I guess
my feeling is that it's these big, small-production-run, systems where
Lisp stands to win, and so it *needs* to have a story about running on
the kinds of machines people buy for this stuff before Java or
something cleans up.
(sorry to have split the thread, I forgot this bit in my previous
article)
--tim
Tim Bradshaw <···@tfeb.org> writes:
> If you mean in terms of memory, then I guess Franz's 64bit Lisps are
> (I don't really see any reasons they should not be anyway).
I guess it depends on the application and how you're able to optimize
GC. Straightforwardly using Allegro CLs 'global GC' tends to be
impossible for big applications.
--
(espen)
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjwvam7nvn.fsf@tfeb.org>
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> I guess it depends on the application and how you're able to optimize
> GC. Straightforwardly using Allegro CLs 'global GC' tends to be
> impossible for big applications.
>
I agree that you need to tune GC. On the other hand, how bad is a
global GC for say a 4Gb heap if you have the physical memory? It
shouldn't be *awful*, though probably it starts getting too slow to do
in `real-time' (not in any formal sense of real-time).
I think it's like fsck -- when filesystems (memory spaces) get big you
can't do the trivial walk-the-while-filesystem (memory) thing any more
because it takes hours, you have to rely on cleverer techniques.
--tim
Tim Bradshaw <···@tfeb.org> writes:
> I agree that you need to tune GC. On the other hand, how bad is a
> global GC for say a 4Gb heap if you have the physical memory?
I don't know, but based on my experiences with GC on a ACL 5.0.1-
based applications with 400-800MB heaps, which, in some cases, could
take *several* hours, I would assume that some 4GB-applications could
spend several days in GC... I don't know what others have found out,
but I ended up tuning the garbage collector to use big 'newspace',
do *no* global GC, and restart the program once a week when its pile
of garbage had reached about half a gig.
--
(espen)
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjvgq67mpz.fsf@tfeb.org>
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> I don't know, but based on my experiences with GC on a ACL 5.0.1-
> based applications with 400-800MB heaps, which, in some cases, could
> take *several* hours, I would assume that some 4GB-applications could
> spend several days in GC... I don't know what others have found out,
> but I ended up tuning the garbage collector to use big 'newspace',
> do *no* global GC, and restart the program once a week when its pile
> of garbage had reached about half a gig.
I've never seem times anything like that on heaps of 400Mb, but I'm
sureit depends very much on the nature of the data. I think the trick
of rebooting rather than doing a global GC is cool -- I have a scheme
where you have two running systems, the master (which does all the
work) dribbles stuff to the slave, and then at some point when it's
time to GC you cut over to the slave instead and start a new system
which then loads from the dribble file and becomes the new slave. (I
don't think this is original to me.)
--tim
From: Andrew K. Wolven
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A924EA8.3C5ACDDC@redfernlane.org>
I am sure that is how David Moon would do it.
Tim Bradshaw wrote:
> Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
>
> > I don't know, but based on my experiences with GC on a ACL 5.0.1-
> > based applications with 400-800MB heaps, which, in some cases, could
> > take *several* hours, I would assume that some 4GB-applications could
> > spend several days in GC... I don't know what others have found out,
> > but I ended up tuning the garbage collector to use big 'newspace',
> > do *no* global GC, and restart the program once a week when its pile
> > of garbage had reached about half a gig.
>
> I've never seem times anything like that on heaps of 400Mb, but I'm
> sureit depends very much on the nature of the data. I think the trick
> of rebooting rather than doing a global GC is cool -- I have a scheme
> where you have two running systems, the master (which does all the
> work) dribbles stuff to the slave, and then at some point when it's
> time to GC you cut over to the slave instead and start a new system
> which then loads from the dribble file and becomes the new slave. (I
> don't think this is original to me.)
>
> --tim
On 19 Feb 2001 14:48:24 +0000, Tim Bradshaw <···@tfeb.org> wrote:
>I've never seem times anything like that on heaps of 400Mb, but I'm
What kind of app needs heaps of 4 Gb, or even 400Mb? O:-)
//-----------------------------------------------
// Fernando Rodriguez Romero
//
// frr at mindless dot com
//------------------------------------------------
From: Andrew K. Wolven
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A9251F7.4B4B9FCC@redfernlane.org>
"Fernando Rodr�guez" wrote:
> On 19 Feb 2001 14:48:24 +0000, Tim Bradshaw <···@tfeb.org> wrote:
>
> >I've never seem times anything like that on heaps of 400Mb, but I'm
>
> What kind of app needs heaps of 4 Gb, or even 400Mb? O:-)
Cycorp comes to mind.
>
> //-----------------------------------------------
> // Fernando Rodriguez Romero
> //
> // frr at mindless dot com
> //------------------------------------------------
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkjsnl9tvtr.fsf@tfeb.org>
Fernando Rodr�guez <·······@must.die> writes:
> >I've never seem times anything like that on heaps of 400Mb, but I'm
>
> What kind of app needs heaps of 4 Gb, or even 400Mb? O:-)
Many many applications. Consider something that wants to think about
phone calls for a reasonable-sized telco -- 1E7 customers, 10 calls a
day, 100 bytes a call record perhaps, that's nearly 10Gb, per day. Or
something that wants to look at a day's worth of data coming from
supermarket tills. In the scientific field, the human genome is 3+
Gb, high energy physics experiments produce Gb of data a second.
Image processing (which could be scientific or could be very
commercial) deals with Gb of data.
Of course most of the commercial-type apps traditionally work by
stuffing things into a database, which is fine if you want to do
simple computations, like computing a bill or something, but is
definitely not fine if you want to do something more hairy like
looking for patterns of various kinds: there you want big data in
memory.
--tim
From: Will Deakin
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A92776E.1060509@pindar.com>
Fernando Rodríguez wrote:
>> I've never seem times anything like that on heaps of 400Mb, but I'm
> What kind of app needs heaps of 4 Gb, or even 400Mb? O:-)
Uhhhh, big ones...
;)w
From: David Thornley
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <szAk6.606$S7.51774@ruti.visi.com>
In article <··································@4ax.com>,
Fernando Rodr�guez <·······@must.die> wrote:
>On 19 Feb 2001 14:48:24 +0000, Tim Bradshaw <···@tfeb.org> wrote:
>
>
>>I've never seem times anything like that on heaps of 400Mb, but I'm
>
>What kind of app needs heaps of 4 Gb, or even 400Mb? O:-)
>
Yeah, any well-written program can be made to fit in 4K.
(Or am I dating myself?)
--
David H. Thornley | If you want my opinion, ask.
·····@thornley.net | If you don't, flee.
http://www.thornley.net/~thornley/david/ | O-
On Sat, 17 Feb 2001 10:46:14 GMT, Alain Picard <·······@optushome.com.au>
wrote:
> potential investors. These investors would like to know
> why we are using Lisp to implement our product, and,
> who else uses it, and for what applications.
[...]
> Oh, I know about the lists maintained on www.franz.com,
I can offer you just another list, sorry. You might have a look at the
industrial applications section of the Association of Lisp Users site. I
recently submitted a few tens of new entries to the section maintainer, I
don't know whether they have been included.
> I am asking for your personal testimonials and estimation
> of how successful lisp has been as a choice of technology
> in your commercial endeavour.
The proceedings of recent Lisp conferences may be a good starting point.
Contact Franz to get copies.
Paolo
--
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://cvs2.cons.org:8000/cmucl/doc/EncyCMUCLopedia/
On Sat, 17 Feb 2001 10:46:14 GMT, Alain Picard <·······@optushome.com.au>
wrote:
>The company at which I work is currently speaking with
>potential investors. These investors would like to know
>why we are using Lisp to implement our product, and,
>who else uses it, and for what applications.
Scientia (http://www.scientia.com/) use Lisp to produce timetabling and
scheduling applications. Our flagship product, Syllabus Plus, is in use at
approximately 400 universities and colleges in 20 countries.
Mark
From: Will Deakin
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A9106F7.6030802@pindar.com>
Try www.aiai.ed.ac.uk/project/formation/
:) will
From: Reini Urban
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3a915d5e.636682060@judy>
AutoCAD developers still love AutoLISP because it is so easy and
elegant, compared to the other possibilities. There're much more
AutoLisper's than VB'lers or C++ people by far. The learning curve is
higher.
Emacs hackers love elisp for the same reasons, though there's only C as
alternative. The elisp library is horrible, nevertheless they love it.
--
Reini Urban
http://xarch.tu-graz.ac.at/autocad/news/faq/autolisp.html
From: Andrew K. Wolven
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A91804D.3E2B3B13@redfernlane.org>
Erik Naggum wrote:
> * Reini Urban
> > AutoCAD developers still love AutoLISP because it is so easy and elegant,
> > compared to the other possibilities. There're much more AutoLisper's
> > than VB'lers or C++ people by far. The learning curve is higher.
>
> Speaking of which, I have heard some unsubstantiated rumors that AutoCAD
> is going to support VB as its application language and will leave Lisp.
Are you a Zen Master, Erik? You sometimes say so many things in such compact
poetic ways.
You never know.
>
>
> #:Erik
> --
> If you are concerned about netiquette, you are either concerned about
> your own and _follow_ good netiquette, or you are concerned about others
> and _violate_ good netiquette by bothering people with your "concern", as
> the only netiquette you can _actually_ affect is your own.
>>>>> "Erik" == Erik Naggum <····@naggum.net> writes:
Erik> Speaking of which, I have heard some unsubstantiated rumors
Erik> that AutoCAD is going to support VB as its application
Erik> language and will leave Lisp.
I believe that in addition to AutoLisp, VB is supported at present. I
have not seen anything remotely close to official that AutoLisp is to
be left behind.
--
Matt Curtin, Founder Interhack Corporation http://www.interhack.net/
"Building the Internet, Securely." research | development | consulting
From: Craig Brozefsky
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <87zofil1bo.fsf@piracy.red-bean.com>
Erik Naggum <····@naggum.net> writes:
> The reason I still stand behind Common Lisp, despite a lack of some
> crucial tools to penetrate the Web (and Franz's pricing of "servers" is
> such a horrible disincentive I could cry) is that it is still so far
> superior to achieving practically error-free (and if not, extremely fault
> tolerant), 100% operational systems that nothing is likely to replace it
> in the coming decade, either.
The tools are there. Franz's pricing for web applications is indeed
problematic, reminding me of Oracle's old pricing policies for web
applications. There are other options tho. Lispwork's has very
attractive pricing, and CMUCL is workeable as well. Most of these web
tools are Free Software of various strains so licensing costs for them
is a non-issue.
--
Craig Brozefsky <·····@red-bean.com>
In the rich man's house there is nowhere to spit but in his face
-- Diogenes
>>>>> Erik Naggum writes:
Erik> Why do they need to know who else uses it and for what? That
Erik> is, what is the _real_ question behind those (pointless)
Erik> questions?
Well, let's see. Firstly, it's their money. :-)
Secondly, if *I* were an investor, and investing into a company
using technologies I'd never heard from, or had horrible prejudices
about, I'd want to know why I should stick my neck out. Dunno.
To me, their questions seem reasonable. I know why *I* use Lisp,
they want to make sure that I'm not a lunatic. Fair's fair.
Erik> But why is this important to them? Again, what is the _real_ question
Erik> they are asking?
The real question(s) are:
* are these guys for real?
* will this thing they're building work?
* will they be able to support/maintain it
(i.e. find lisp programmers in 3 years?)
* is this a good bet for me to place my money on?
Erik> Software that can't fail. For software that can fail, we have
Erik> any number of other options, naturally.
[snip]
This, and all other good reasons explained there, fell under the
category of "what I explained to them". Having you and others
confirm it will strenghten in their minds that I'm not a lunatic,
which is a Good Thing (TM). Thanks.
Erik> The reason I still stand behind Common Lisp, despite a lack of some
Erik> crucial tools to penetrate the Web (and Franz's pricing of "servers" is
Erik> such a horrible disincentive I could cry)
Boy, am I with you there. What a disaster. :-(
Fortunately, there *were* excellent alternatives.
Erik> Other than this, I think you should dissuade your investors from asking
Erik> such questions.
Well, if wishes were horses then beggars would ride.
We're begging, so we're answering.
Erik> If I
Erik> told people who didn't know it was possible that we have had servers run
Erik> from late January to the end of the year, and were shut down only to
>>> ^^^^^^^^^^^^^^^^^^^^^^^(!!! Wow! I'm duly impressed.)
Erik> and ignore
Erik> the hype from Redmond and other places who don't know what it means.
Well, naturally, I do *that* as a matter of course. ;-)
Thanks for your thoughtful reply!
--Alain Picard
From: Tim Bradshaw
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <nkju25ptwc5.fsf@tfeb.org>
Erik Naggum <····@naggum.net> writes:
> Software that can't fail. For software that can fail, we have any number
> of other options, naturally. Once the requirement that it cannot fail is
> dropped, Lisp would no longer have its major edge, and we could just find
> ways to painlessly restart everything. Before my Common Lisp system was
> deployed, that was indeed what people did: Reboot the server, call each
> of the contacts and ensure that things got back up gracefully. This used
> to be a daily routine. My system has been running for two and a half
> years with less than two hours of downtime during operating hours.
I think this is really well put, and the bit that I elided is pretty
significant too. It's obvious that you can make brutally reliable
systems in C -- our Unix machines for instance have *never* had
software-induced crashes (yet!) and have uptimes limited by occasional
upgrades and hardware failures (what I'm trying to say is that we
don't avoid SW failures by rebooting once a week or something). But
the way they've achieved this reliability is 30 years of development.
With a lisp system, if you're competent, you stand some kind of chance
of being able to produce something which is very reliable a lot faster
than this. My belief is that you just can't do this in many languages
-- of course *some* people can, and have, but those people are
incredibly smart, incredibly rare, and incredibly expensive, whereas
good Lisp people are merely quite smart, quite rare and quite
expensive.
(Yes, I am arguing against my article of yesterday. I contain
multitudes.)
--tim
From: ··········@my-deja.com
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <96u3nu$j6i$1@news.netmar.com>
In article <··············@optushome.com.au>, Alain Picard
<·······@optushome.com.au> writes:
>
>Dear Lispers,
>
>
>The company at which I work is currently speaking with
>potential investors. These investors would like to know
>why we are using Lisp to implement our product, and,
>who else uses it, and for what applications.
>
>We have explained to them the technical advantages of Lisp
>(stability, programmer productivity, cross-platform development, etc)
>and while they appreciate these reasons, they wonder why they
>have not heard of anyone else using it.(*)
Micromanagement at this level from your investors is bound to be
counter-productive for them.
Having said that, i think your best defense is to show that Lisp is an
established international standard which is stable supported by multiple
vendors, so your company is not at the mercy of any individual company. You
won't be dead if one or more Lisp vendors go out of business, and you won't
be hurt by a vendor changing the specification with each version they
release.
Good luck!
g
----- Posted via NewsOne.Net: Free (anonymous) Usenet News via the Web -----
http://newsone.net/ -- Free reading and anonymous posting to 60,000+ groups
NewsOne.Net prohibits users from posting spam. If this or other posts
made through NewsOne.Net violate posting guidelines, email ·····@newsone.net
Dear Lispers,
Thank you very much to all those of you who sent in your
stories, both public and private. I trust they will have
the intended effect!
-- Alain Picard
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
From: Kent M Pitman
Subject: Re: THANKS! [was : Help me convince skeptics!]
Date:
Message-ID: <sfwbsrug6qr.fsf@world.std.com>
Alain Picard <·······@optushome.com.au> writes:
> Dear Lispers,
>
> Thank you very much to all those of you who sent in your
> stories, both public and private. I trust they will have
> the intended effect!
If you create a summary and can share it, I'm sure someone can find a
place to display it on the ALU web site (if you don't have a web
site of your own to use...) We need more materials like this.
>>>>> Kent M Pitman writes:
Kent> If you create a summary and can share it, I'm sure someone can find a
Kent> place to display it on the ALU web site (if you don't have a web
Kent> site of your own to use...) We need more materials like this.
Good idea. The most interesting replies, however, came
via e-mail and must be treated as confidential.
Still, I'll see what I can do.
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
From: Marc Battyani
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <975hf4$69$1@reader1.fr.uu.net>
"Alain Picard" <·······@optushome.com.au> wrote
> The company at which I work is currently speaking with
> potential investors. These investors would like to know
> why we are using Lisp to implement our product, and,
> who else uses it, and for what applications.
>
> We have explained to them the technical advantages of Lisp
> (stability, programmer productivity, cross-platform development, etc)
> and while they appreciate these reasons, they wonder why they
> have not heard of anyone else using it.(*)
>
> So, what do you use Lisp for? Real life stories of
> successful systems built in Lisp would be greatly appreciated.
We make real time industrial applications in Lisp (some work 7x24).
The really real time part (read 10s of ns) is in VHDL the rest in Lisp.
(you can email me if you have questions)
> Oh, I know about the lists maintained on www.franz.com,
> I am asking for your personal testimonials and estimation
> of how successful lisp has been as a choice of technology
> in your commercial endeavour.
The only drawback is that I had to give the name of several
companies/universities/people that stated they know lisp and could work on
the systems if we were gone.
May be we could reference companies/people knowing lisp and able to do
contract work in some place (ALU?)
Otherwise the "no-fail" and "fix pbs without stopping the system" properties
of lisp made us get some contracts against the C++ crowd.
Lisp developpers cost more but it's nothing compared to stopping a
production factory because a C++ program crashed. And writting safe programs
in C++ will take longer and cost more than in lisp.
Marc
"Alain Picard" <·······@optushome.com.au> wrote in message
···················@optushome.com.au...
>
> Dear Lispers,
>
>
> We have explained to them the technical advantages of Lisp
> (stability, programmer productivity, cross-platform development, etc)
> and while they appreciate these reasons, they wonder why they
> have not heard of anyone else using it.(*)
>
> So, what do you use Lisp for? Real life stories of
> successful systems built in Lisp would be greatly appreciated.
The company i recently left uses lisp for the major server component of its
system that provides "day of operations" crew control software for airline
crew controllers. It is a (pseudo) "shrink-wrapped" (with varying amounts
of customisation) product that has been installed at c.5 airlines since
leaving its R and D phase about 2 years ago. The software runs 24/7 for one
month periods. It is then rebooted for reasons that have to do with airline
procedure and system design.
Lisp gave us the advantage of being able to modify the program without
having to shut down (eg to implement new government or union regulations or
(ahem) fix bugs..i mean apply enhancments) plus all the usual advantages
lisp comes with "out of the box"
This is generally a success story marred only by horrible mis-management.
Coby
(sorry for such a late reply, i was simultaneously without home internet
access *and* cut off at the knees by Deja/Google)
....and how else could i have known you were asking? ;-)
From: Andrew K. Wolven
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <3A9A37BB.4AD0AAEF@redfernlane.org>
Coby Beck wrote:
> "Alain Picard" <·······@optushome.com.au> wrote in message
> ···················@optushome.com.au...
> >
> > Dear Lispers,
> >
> >
> > We have explained to them the technical advantages of Lisp
> > (stability, programmer productivity, cross-platform development, etc)
> > and while they appreciate these reasons, they wonder why they
> > have not heard of anyone else using it.(*)
> >
> > So, what do you use Lisp for? Real life stories of
> > successful systems built in Lisp would be greatly appreciated.
>
> The company i recently left uses lisp for the major server component of its
> system that provides "day of operations" crew control software for airline
> crew controllers. It is a (pseudo) "shrink-wrapped" (with varying amounts
> of customisation) product that has been installed at c.5 airlines since
> leaving its R and D phase about 2 years ago. The software runs 24/7 for one
> month periods. It is then rebooted for reasons that have to do with airline
> procedure and system design.
Neat.
Seems like scheduling/planning applications could have their own lisp sub-cult
like bioinformatics.
It's popular. Doesn't Swissair do scheduling with lisp?
I would think that it is because of the runtime updateability of clos objects
and metaobjects.
Would anyone say that is because of the dynamicity (if that is a word) of clos?
Thinking just of base lisp, is the dynamicity of the language more or less due
to the fact that alpha and beta conversions happen at runtime? (except for
normal compilation, which isn't really supposed to make the s-expression act
different)
How does C do it? How does this affect what the end user gets out of their time
on a machine?
I am a little bit at a loss for terms. I started with BASIC, brushed very
quickly over pascal and Fortan, got into scheme and lisp, touched java, but more
or less never looked back [at C++ or whatever]. Am I at a loss for only
sticking to CL for the most part? CBind is a little painful. COM a little
quirky and new. Variant?
Is there a technical reason for not using the funcall/apply interface for O.S.
level stuff?
I am a dumbass. Sorry.
>
> Lisp gave us the advantage of being able to modify the program without
> having to shut down (eg to implement new government or union regulations or
> (ahem) fix bugs..i mean apply enhancments) plus all the usual advantages
> lisp comes with "out of the box"
>
> This is generally a success story marred only by horrible mis-management.
>
> Coby
> (sorry for such a late reply, i was simultaneously without home internet
> access *and* cut off at the knees by Deja/Google)
>
> ....and how else could i have known you were asking? ;-)
From: Marco Antoniotti
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <y6chf1hzex7.fsf@octagon.mrl.nyu.edu>
"Andrew K. Wolven" <·······@redfernlane.org> writes:
> Neat. Seems like scheduling/planning applications could have their
> own lisp sub-cult like bioinformatics.
Not really. Perl and Python rule supreme in the field of the string
manipulators. And C/C++ is used to do string matching "efficiently".
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
Like DNA, such a language [Lisp] does not go out of
style.
Paul Graham, ANSI Common Lisp
From: Marco Antoniotti
Subject: Re: Help me convince skeptics! Lisp success stories wanted...
Date:
Message-ID: <y6cd7c5zeks.fsf@octagon.mrl.nyu.edu>
Marco Antoniotti <·······@cs.nyu.edu> writes:
> "Andrew K. Wolven" <·······@redfernlane.org> writes:
>
> > Neat. Seems like scheduling/planning applications could have their
> > own lisp sub-cult like bioinformatics.
>
> Not really. Perl and Python rule supreme in the field of the string
> manipulators. And C/C++ is used to do string matching "efficiently".
>
Well, I should qualify that.
www.biolisp.org
is online thanks to Larry Hunter. I write in CL (under great pressure
to switch to something "other people use"), but the above still mostly
holds.
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
Like DNA, such a language [Lisp] does not go out of style.
Paul Graham, ANSI Common Lisp