davidsen@ariel.crd.GE.COM (william E Davidsen) writes:
quoted text
>  Let me say a word about documentation. I ahve written a number of
>things which have gone out through source groups, all the way back to
>mod.sources. During that time I wrote a user's guide to zoo, the
>archiver. The guide is about 35 pages long, and took almost as long to
>write a a major program. I have had a total of two people drop me a line
>to tell me they liked it, and few if any archive sites carry the manual
>even though they have the software.
>
>  Given that documentation is less fun to write than software, brings no
>recognition whatsoever, and takes longer page for page than code, you
>need to find someone who really wants to write documentation, or it's
>not going to happen. If there's someone who is a good writer and a
>complete loss as a programmer, and who wants to really contribute to the
>effort, that's the person who should do it. And all the postings in the
>world about what we need won't help unless you can find someone to do
>it.

Why not try to use documentation that already exists (where possible, of
course)? Why keep reinventing the wheel just for the sake of reinventing the
wheel???

For example, I can buy a whole suite of manuals on SVR4/386 from Prentice Hall
for about $300. Yes yes, that's pretty steep, but what are the realistic
alternatives?

Suppose Linux was heading to where SVR4/386 already is: most of our manuals
would already be written for us. And we would rarely have to cope with Linux
quirks when porting the various applications. And we would have a roadmap of
where we were going.

I know we don't want to do this, but as a hacker, I often (though not always)
like hacking on new things, not redoing existing things just so I can say
"I did it my way".
-- 
Doug Evans               | "You're just supposed to sit here?"
dje@sspiff.ampr.ab.ca    |               - Worf in a mud bath.