Subject: Re: source access vs dynamism
From: Erik Naggum <erik@naggum.no>
Date: 1999/08/27
Newsgroups: comp.lang.lisp
Message-ID: <3144733008312622@naggum.no>

* Christopher Browne
| In short, the situation where source access causes problems appears to
| represent a rather peculiar scenario that is not representative of any
| widespread phenomenon.

  someone here accused me of confusing fact and right, but I think what you
  have posted is just that.  I have no interest in discussing numbers of
  people or magnitude of practical problems.  my interest is on an entirely
  different axis: dynamism in software.  it's right up there in the subject
  line, too.  I'm arguing that those who want dynamism and think they need
  source code will get less dynamism when they get source code than a they
  would if they (1) chose a dynamic programming language and (2) could do a
  lot of interesting things without source code.  we are obviously not
  talking about people who do not benefit from source code because they
  don't own computers, either, and frankly, I don't understand the point of
  arguing about "source access vs dynamism" in such terms.

  my purpose was to show that people _need_ dynamism in their software,
  which you might of course argue against by saying that only N people do
  it and the rest are happy without it, but I only care about those N
  people in my argument.  I argue that those N people will not learn to
  write dynamic software that can adapt without source access or even at
  runtime because they have source access and think that's great, when it's
  only great compared to _not_ having source access -- it is not great
  compared to having fully dynamic behavior in the software, and since they
  are used to 10% dynamism-via-source-code-in-static-language and don't
  even see what 40% dynamism-via-dynamic-languages-without-source would
  mean for them or even 90% dynamism-via-dynamic-languages-with-source in
  the case where you can experiment with a change to a function in a
  running system.

  I honestly wonder why so many people don't see the dynamism argument and
  only latch onto the source access argument.  is it because people don't
  really know why they want source access?  Christopher's argument appears
  to be that it isn't needed.  my argument is that people who don't need
  source code still need dynamic behavior.  take GNU grep.  wouldn't it be
  great if you could make GNU grep always print the filename with an option
  instead of tacking on /dev/null at the end like Emacs does?  wouldn't it
  be great if you could instruct GNU grep to default to case-insensitive
  searches?  how about enclosing the filename in double quotes so Emacs can
  find matches in files that happen to contain colons in their names?
  these would be simple local patches in a dynamic system, but it may just
  be too much work to fix the source, submit a patch, and argue for the new
  features.  as you correctly observe, local fixes die with source access,
  but if the fixes are in the form of manageable advise code, they would
  survive an upgrade.  again, dynamic languages win on all points, but
  since people have source access, they won't think they need it, even with
  the many problems caused by source access, such as the ones you bring up.

#:Erik
-- 
  save the children: just say NO to sex with pro-lifers