long names in 'C' programs

[Guy Finney]

I have to agree.  The case in point here is the highly acclaimed GNU Emacs.
We got it shortly after it was announced over the net that it was
available.  We had been warned it would be somewhat 4.2-specific, but
figured it wouldn't be too long a job to convert to Sys V (has anyone
got it on Sys V - we'd like to hear from you).  Surprise!  #defines,
variable names, etc not unique within even 20 characters.  We're one of
the poor binary licensees alluded to by the author, and called the person
who gave us GNU Emacs for help.  He said something like "Oops, well, we think
self-documenting code is such a good thing that we traded portability
for it.  Can I interest you in a nice convert-o-matic program?".  Great.
This guy may be a whiz-bang software architect, but he doesn't seem to
understand maintainability at all.  We got the convert-o-matic program
(I think it's called clash), but then updates began coming out to GNU
Emacs, and the amount of time it took to port the original plus the
updates...well, we gave up trying.  We just threw away GNU Emacs
entirely.  In contrast, the folks who do Kermit seem to have
portability as their primary consideration.  C-Kermit took something
like an hour to get up and running without portability bugs,
and that includes compile time.  They've got the right attitude.
-- 
Guy Finney
{decvax|ihnp4|hao}!noao!terak!anasazi!guy