Socioeconomic equivalence of Lisp and Ada

Socioeconomic equivalence of Lisp and Ada

[haven.umd.edu!darwin.sura.net!wupost!sdd.hp.com!caen!sol.ctr.columbia.edu]

    What follows is an editorial that appeared in the latest edition of the
journal "Computer Language".  What struck me is that you could substitute
Ada for Lisp throughout his comments, and it would be an equally valid view
of the programming world.
    In both cases, Ada and Lisp, I think that they are excellent languages
for certain applications, approaching that of anlegance for their fit.
However, they both are similar socioeconomically as being viewed as niche
languages - in the free marketplace, they languages had their chance and
failed in terms of broad free market acceptance (though for different reasons).

     My postings have nothing to do with the technical merits of Ada, of which
there are many.  But rather the pathetic way in which Ada policies have been
prepared and the language promoted with my tax dollars.

Greg Aharonian
Source Translation & Optimization

==============================================================================

EDITOR'S NOTES  -  LARRY O'BRIEN  -  NOVEMBER 1992 -  COMPUTER LANGUAGE

"I was recently given an opportunity to speak at a LISP vendor and user's
conference as part of a roundtable on "The Common Lisp Object System vs. C++".
I accepted the role of C++ advocate primarily because the conference was in 
San Diego in August, and I figured being the subject of a flame session was
a small price to pay for even a few hours at La Jolla Shores.
    I knew the likely points of attack in the debate:  the advantages of an
interpreted environment, the power of LISP macros, and built-in garbage
collection.  There's no good way to argue against any of those points.  Well,
maybe garbage collection, but the average Lisp programmer can't understand
the C viewpoint that a memory leak is a defect that should be caught during
development, not an inevitable part of programming that should be handled at
run time.  Luckily I had a secret weapon in the form of the HELP WANTED pages
of the Sunday paper.
     "Those jobs circled in blue", I said, indicating the 60 or so such ads,
"are jobs for C or C++ programmers.  This one circled in red is for a LISP
programmer, with some C programming experience preferred.  The greatest
argument for C++ over LISP is embodied in Stalin's phrase 'Quantity has a
certain quality all it's own'."
    Well, the roundtable was great fun, and I, at least, learned some things.
One of things I discovered was that the LISP community has an explanation for
the success of C.  It's called "Worse is Better" and is the brainchild of
Richard Gabreil of Lucid. This theory holds that the less capable a language,
the more a third-party industry is required and the more specialized
information, such as that found in books and magazines, is needed.  These
factors lead to increased viability and popularity of the language.
    To some extent, this is sour grapes on the part of the LISP community,
which is facing hard times only a few years after giddy venture-capital
driven days during the mid-80s.  One of the problems of LISP, which was
actually touted as a feature by Jon White, my opposite number on the
roundtable, is that LISP has always been "a step in front of the hardware"
while C has always been "a step behind".
    In other words, LISP has traditionally required an expensive workstation
while inexpensive C and C++ compilers can work on the sorts of PCs programmers
have at home. The LISP community will be in trouble as long as they view that
discrepancy as a selling point. 
    The "Worse is Better" concept does have a grain of truth, however. To be
highly productive in C, you're certainly using either home-grown or commercial
extensions to the base language product: development environments, interface
libraries, database libraries, and a tool to help you with memory leaks and
range violations.  A healthy third party industry has grown around providing
those tools.
    The competition has created great tools that continually improve.  A C
programmer can buy a compiler, confident that virtually any problem in the
package short of incorrect code generation will be addressed, rapidly, by
a third-party product.  LISP programmers have to hope that their gripe is
high on the vendors'swish list for the next release.
     Although "Worse is Better" is a catchy phrase, it's not really a fair
one. "Less is Better" is more accurate.  Someone at the conference asked me
what LISP vendors should learn from the success of C and C++.  Maybe it's
the model of a language that serves as a skeleton - intentionally
incomplete.  Also, to be totally honest, I think you have to accept the fact
that that for whatever reason, programmers find the burden of typing words
to mark blocks unacceptable. At least LISP has that based covered.  You just
have to match parantheses".

==============================================================================

    Many of the socioeconomic thoughts in the above editorial apply to Ada
as well as Lisp.  The DoD concern for free market forces, third party suport
for a language, competition, and other social and economic aspects of software
has been absymal.  Rarely do any of these concerns appear in discussion and
reports, such as the highly touted but quickly ignored Mosemann studies, whose
authors, having lived off of non-free market money, had no idea how to include
free market concepts into an study on Ada.

    The continual lack of attention to the socioeconomics of Ada will, despite
its being a good language, lead over time to higher and higher software
development costs for any Ada projects, as the majority of software resources
(programmer base, tool base, components base, educational base, university
base, advanced research base, etc) in this country are directed towards other
languages.



-- 
**************************************************************************
Greg Aharonian
Source Translation & Optimiztion
P.O. Box 404, Belmont, MA 02178