Date: Mon, 12 Jun 1995 08:15:29 +0300
From: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
To: jhunter AT kendaco DOT telebyte DOT com
Subject: Re: VOTE: CODE STANDARDS
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu

> >the pace we are working.  If everyone getting this message (about 700
> >people, I'm guessing) wrote a page of documentation, or documented one
> >source file, we'd be done in a day.
>
>   Is this a request?

Yes, it is.  As we all know, docs usually lags behind because programmers
don't like writing (or even reviewing) it, and professional docs writers
usually don't understand the subject matter good enough.  So yes, if you
can sit and review the library docs, and post improvements, you won't be
one of a thousand.

>   How about the guy who started this thread try to
> organize such an ordeal.

This won't help.  You don't get the job done by telling other people to do
it, not in a ``voluntary monarchy'' (DJ's words) like DJGPP is.

>                               Even if I didn't write the libraries, I'll
> bet that at least 50% (probably more like 80%) are clear enough that
> your average DJGPP programmer could document them without much
> difficulty.  Heck, I feel it's the least I could do.  I mean honestly,
> how long does it REALLY take to figure out what __tb is???

For somebody who has perused the library in every direction several times--
probably a millisecond, even if it will be called __my_secret_var_you_dont_want_to_know;
it's enough to see it assigned to two registers before calling __dpmi_int().
For the rest of us--I don't know.  And why should a user download and look
into the source to understand what a library function does?