🚀 go-pugleaf

  Kent, there is an inherent difference between the problems that were
found in your townmaze program after you posted it, and the problem
with Zsh that Gregory Woodbury was pointing out.

  Your program was pretty much machine-independent.  The bugs that you
found were bugs on *any* system, not just on the ones that stumbled
over it.  Granted, the bugs didn't *turn up* on all systems on which
you tested the program (else, as you pointed out, they would have
turned up during your testing and would not have been in the version
you released), but they were bugs nevertheless, and they might have
turned up if you have used a debugging malloc or Saber-C or something
like that.

  On the other hand, Zsh has in it code that is inherently
non-portable, and that requires certain environments in order to
compile it.  Now, my suspicion is that anyone who is "sophisticated"
(your word; I'm not sure it's the one I would have chosen, but it
conveys what I'm trying to say in parallel with what you said) enough
to write a shell, also knows enough to know when he's writing code
that will only compile on certain types of system.

  If he doesn't, then I cannot fault him for not pointing out the OS
dependencies, and I doubt anyone else would either.  But I think it is
reasonable to expect him to say, *in an introductory posting separate
from the source code he's posting*, which environments he has compiled
for.  Furthermore, if he *knows* what OS dependencies there are, it is
reasonable to expect him to point them out in that introductory
posting.  That's what Greg is asking for.

  You couldn't have known about the bugs in your code, so you couldn't
have pointed them out.  But the author of Zsh did know about the
dependencies, so he could have pointed them out, and probably should
have, in an introductory posting.

Jonathan Kamens			              USnail:
MIT Project Athena				11 Ashford Terrace
jik@Athena.MIT.EDU				Allston, MA  02134
Office: 617-253-8085			      Home: 617-782-0710