Newsgroups: comp.lang.lisp,comp.lang.scheme
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!nntp.sei.cmu.edu!news.psc.edu!wink.radian.com!gatech!news.mathworks.com!fu-berlin.de!cs.tu-berlin.de!news.uni-hamburg.de!news.Hanse.DE!wavehh.hanse.de!cracauer
From: cracauer@wavehh.hanse.de (Martin Cracauer)
Subject: Re: Why lisp failed in the marketplace
Message-ID: <1997Mar12.122514.12300@wavehh.hanse.de>
Reply-To: cracauer@wavehh.hanse.de
Organization: '(a (cons tructive organization))
References: <5edfn1$83b@Masala.CC.UH.EDU> <m3pvxxylnf.fsf@laphroig.mch.sni.de> <330B3744.10E3@acm.org> <5f0gn3$hd$1@goanna.cs.rmit.edu.au> <sdm7g-0303971459080001@bootp-17-17.bootp.virginia.edu> <332676CE.7E565B54@mrj.com>
Date: Wed, 12 Mar 97 12:25:14 GMT
Lines: 119
Xref: glinda.oz.cs.cmu.edu comp.lang.lisp:25998 comp.lang.scheme:19093

Randy Crawford <crawford@mrj.com> writes:

>sdm7g@Virginia.EDU wrote:
>> 
>> [1] Re: "Why lisp failed in the marketplace"

>I've come late to this old saw, but what the heck...

>> 
>> In reality, Lisp has been a phenominal success.

>Your definition of "phenomenal" must differ from mine.

>> It's one of the earliest computer languages that is still around 
>> and viable.  It's elders, Fortran and Cobol, have largely hung on 
>> due to the inertial of a large base of existing and installed code. 

>This is at least as true of Lisp as it is FORTRAN.  In fact, I think 
>it's probably accurate to say there's a GREAT deal more new development
>taking place in FORTRAN and COBOL than in Lisp.  The active evolution
>of both languages indicates their popularity.  Alas, by the same
>measure,
>Lisp is not faring well outside of academia (even inside, the academic 
>research market has moved onto better typed functional languages).

Mostly of those researches wouldn't call *differently* typed langugae
*better*. Static or string typing is a trade-off and a matter of
taste.

[...]

>> So how is it a "failure" ?
>>   Well -- it's been a failure at making the Lisp Machine vendors or the Lisp
>>   compiler & tools vendors rich.

>And those who believed to the contrary, like Lucid and Symbolics, are 
>dead, dead, dead.

Lucid is dead because their C++ projects failed and wasted so much
money that the Lisp business (which was profitable) couldn't save
them.

>>   Except as embedded in products like AutoCad, it's been a failure in being
>>   a large presence in the PC world, which is where the mass market is.

>This invites the question of whether Autocad scripts are really Lisp, or
>whether scheme, T, flavors, etc are really Lisp.  If they ALL are, then 
>it might be claimed that Lisp *had* a little success in the narrow
>market
>related to AI apps, but certainly less so than the Lotus 1-2-3 scripting 
>language, and certainly less than the venerable GW Basic.

>> Although it's been successful, considering all of it's advantages,
>> perhaps it ought to have been a much larger success.

>I think the reasons why Lisp never did well were obvious -- *large* 
>syntax, slow in most incarnations, large memory usage, poor interface
>to GUIs and external resources (like databases), and most of all, the
>inability if the Lisp afficianados to realize that any language asked
>to perform outside of its strengths (rapid prototyping and rich data 
>types) will not succeed when pitted against other tools better suited 
>to that 99% of programming that focuses on the mundane -- input and
>output.  Lisp was simply overkill for most of the tasks at hand.

I think you miss the point about Lisp. I can have as much data types
in C++ as I like. And I don't understand this rapid-prototyping
business anyway (if defined different from "efficient" programming). 

For me, Lisp is about code desity, the ability to concentrate program
code in a way that the needed information is so well-structured that I
don't have to take ballast around when working on that code. Lisp
syntax and macros, reflectivity, the compiler at runtime and other
things like this help a lot and Common Lisp is quite unique in that
area.

And for most readers of the group Lisp's one of Lisp's greates
strengths is its flexibilty, its ability to be used for *very*
different tasks. Operations on simple data types and arrays of those
can be implemented overhead-free in Common Lisp. The language alllows
that by its definitions (while Java, for example doesn't) and
compilers like CMUCL are pretty good is using this potential.

And Lisp's flexibilty and code abtraction ability clearly show off in
CLIM. While it may be true that Lisp has not many bindings to
other-language GUI toolskits, it has show that it is a decent GUI
language by itself. Look at CLIM and Garnet. Likewise is its abilty to
use relational and OO database. Look at the toolkits Harlequin and
Franz provide and at statice. Lisp can be used to work with external
database data in a very transparent way, it can mimic operations on
external data so that it looks not different from working with
language-native constructs like classes and structs. And it does
without precompilers, syntax extensions and such.

>> So it might be valid do say that Lisp failed in the mass market, 
>> but it's a distortion to say it failed in the marketplace.

>I can't agree.  If Lisp is to be considered a success, it was in 
>academia and the early explorative years of AI.  By any standard, 
>Lisp's successes did not extend to the general marketplace.

>If every Lisp-based business has eventually crashed and burned (or 
>else left Lisp behind), then a commercial "success" like that of 
>Lisp is one I would not wish on any friend.

Computer businesses mostly don't fail because of the technical
abilities. If your goal is to use software where technical excellence
domainates usage and survival, use free (not cost-nothing) software.

As long as Common Lisp doesn't loose its adaptabilty (Dylan looses
some flexibilty, code desity and syntaxtical variance, Scheme looses
the ability to go down to memory locations and data layout) and free
implementations like CMUCL are available, Lisp is hard to kill.

Martin
-- 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin_Cracauer@wavehh.hanse.de http://cracauer.cons.org  Fax.: +4940 5228536
"As far as I'm concerned,  if something is so complicated that you can't ex-
 plain it in 10 seconds, then it's probably not worth knowing anyway"- Calvin
