From: Nick Walton)@isengard.demon.co.uk (Nick Walton
Subject: RE: Modern RISC's compilers (was Re: Commercial Lisp Work in UK??)
Date: 
Message-ID: <4@isengard.demon.co.uk>
In article <··········@info-server.bbn.com> ·····@bbn.com writes:

>
>it's really weird. you don't go over to CompUSA and ask what language
>Microsoft Word was written in...but we seem to have to spend a lot of
>time defending why we use Lisp. my project here at BBN is going to have
>to do it again this month. it caused trouble a year ago at my previous
>employer/project, for more or less the same reason. it was OK for 14
>months, then suddenly it wasn't any more.
>
>why does the question ever even come up?
>
> -- clint
>

The sad answer all too often is not to be found in the language per se, but in
the people who use LISP as a tool all too often as a blunt instrument. The
system I am working on/supporting at the moment is typical. Much of the code
was obviously put together quickly, without little thought to design or
maintenance. LISP Seems to breed this sought of sloppy code all too often.

This is not unusual in my estimation, the worst case was when I worked on a CAL
system using Symbolics (supposedly the largest such net in Europe at the time).
I remember hearing the project manager say one day "Oh that's what mapcar
does!". When I queried why the code was being written mainly by contractors
inially hired to write C I was told that this would make the code easier to
port :-(

Surprise, surprise they went bust!

If LISP, esp CLISP, is to be an effective tool for software development then it
is up to those of us who develop with it to do a good job for the CUSTOMER and
not just for our own enjoyment.
From: Ken Anderson
Subject: Re: Modern RISC's compilers (was Re: Commercial Lisp Work in UK??)
Date: 
Message-ID: <KANDERSO.94May13105718@wheaton.bbn.com>
In article <·@isengard.demon.co.uk> ·····@isengard.demon.co.uk (Nick Walton)@isengard.demon.co.uk (Nick Walton) writes:

   In article <··········@info-server.bbn.com> ·····@bbn.com writes:

   >
   >it's really weird. you don't go over to CompUSA and ask what language
   >Microsoft Word was written in...but we seem to have to spend a lot of
   >time defending why we use Lisp. my project here at BBN is going to have
   >to do it again this month. it caused trouble a year ago at my previous
   >employer/project, for more or less the same reason. it was OK for 14
   >months, then suddenly it wasn't any more.
   >
   >why does the question ever even come up?
   >
   > -- clint
   >

   The sad answer all too often is not to be found in the language per se, but in
   the people who use LISP as a tool all too often as a blunt instrument. The
   system I am working on/supporting at the moment is typical. Much of the code
   was obviously put together quickly, without little thought to design or
   maintenance. LISP Seems to breed this sought of sloppy code all too often.

What is it about the code that is sloppy?  Is it different styles of
programming, bad design, silly algorithms, silly programmers?

If it is easy to write code in Lisp, then it should be easy to write sloppy
code.  It is easy to write prototypes in Lisp, and prototypes "attract"
sloppy code (its always the night before the big demo, i'll rewrite it
tomorrow, ...) 

If the code is worth writing, it is worth writing sloppy.  

Perhaps you are working on a prototype that had no intention of being
"maintained".  Prototypes often start off as a demo but then evolve into a
real system, by rewriting large hunks of it (1/3 say) as often as possible.
But, time and money constraints usually limit that, so you get stuck
maintaining sloppy code.  Sometimes sloppy code does get improved,
but not completely so it's no longer sloppy, but only "twisted".  Sometimes
its hard to tell the difference.  Sometimes the sloppiness spreads like a
cancer due to deadline pressure.

So, this suggests that Lisp is a big magnet for sloppy code, and a perfect
medium for sloppy code to grow and thrive.  Perhaps this is because Lisp is
so "habitable" (Gabriel's term, see Critic at Large column in JOOP).
Imagine what it would be like to program in a language that required
perfect elegance, like ML or Scheme, or attention to every detail for
efficiency, likc C++.

Maybe we write sloppy code because we've only seen sloppy examples.  Can
anyone suggest some good code to read?  I recommened PCL and symbolics
source code (with reservation).

   This is not unusual in my estimation, the worst case was when I worked on a CAL
   system using Symbolics (supposedly the largest such net in Europe at the time).
   I remember hearing the project manager say one day "Oh that's what mapcar
   does!". When I queried why the code was being written mainly by contractors
   inially hired to write C I was told that this would make the code easier to
   port :-(

   Surprise, surprise they went bust!

   If LISP, esp CLISP, is to be an effective tool for software development then it
   is up to those of us who develop with it to do a good job for the CUSTOMER and
   not just for our own enjoyment.

This sounds like an example of a management disaster for which Lisp becomes
the scape goat.
--
Ken Anderson 
Internet: ·········@bbn.com
BBN STC              Work Phone: 617-873-3160
10 Moulton St.       Home Phone: 617-643-0157
Mail Stop 6/4c              FAX: 617-873-3776
Cambridge MA 02138
USA