Subject: Re: Pitching a "Big Tent" for Free Software
Date: Wed, 15 Dec 1999 09:26:41 -0500

Rich Morin wrote:
> SJT> I think this discussion is interesting ...
> I do, too, but it has essentially nothing to do with the original topic.
> Could you folks please generate a new Subject line?
Damned if you do, damned if you don't.  IME more people complain about
wandering subject lines than off-topic messages so please don't

> Meanwhile, I'd like to know what folks think can and should be done to
> build infrastructure for the Open Source community...

Well project number (begins counting down queue - gives up), somewhere down
there that I am kicking around...

One basic problem is organizing information that is placed in haphazard
places, organized in a haphazard way, which people tend to want to grab
semi-random stuff from.  Your approach is to put it all in one organized
place, and then teach people how to search for what they want.  Another way
to slice the problem is to build a better search facility.  Lest you think
that is a ridiculous idea, why is deja (used to be dejanews) so commonly
listed as a technical resource in the Linux community?

Let me toss out the one on the bottom of my queue.  One type of a searching
facility that users find transparent are infobots.  These are automated
programs that sit on IRC channels, get fed tidbits, and then spit out
likely-looking tidbits as answers to likely-looking questions.  They work
surprisingly well - well enough that many a newbie has come into a channel,
asked the FAQ question, received an answer, thanked the answerer and left
without being aware that the respondant was a computer program.

Well the idea that I am kicking around but definitely won't get to before
next summer is to modify an infobot to be able to load the data backing
several infobots, and then turn it into an interactive Unix utility.  The
idea being that you locally would have backed your "tip" utility with data
from channels appropriate to what you have locally installed, and when in
need of something can often get a useful tidbit (eg a pointer to a
tutorial, utility, or CVS repository) from it.

Some thought into building tools like this could help reduce users learning
curves, and speed the acceptance of better mouse-traps.  (Why is that small
thing with whiskers glaring at me?:-)

Other examples of useful tools would be search engines directed to mailing
lists, and collections of mailing lists, search engines for specific types
of documents (eg programming tutorials), and other such tools that attempt
to bridge the gap between useful information being somewhere out there and
it being in your hands.