Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!snorkelwacker!tut.cis.ohio-state.edu!pt.cs.cmu.edu!rochester!rit!ultb!jdb9608
From: jdb9608@ultb.isc.rit.edu (J.D. Beutel)
Newsgroups: comp.sys.atari.st
Subject: Re: Programming self-help
Summary: keywords are the key
Keywords: programming sources utilities archive
Message-ID: <3070@ultb.isc.rit.edu>
Date: 6 May 90 16:49:07 GMT
References: <1990May4.163735.9794@chinet.chi.il.us> <1990May5.161812.440@murdoch.acc.Virginia.EDU>
Reply-To: jdb9608@ultb.isc.rit.edu (J.D. Beutel)
Organization: Information Systems and Computing @ RIT, Rochester, New York
Lines: 40

>saj@chinet.chi.il.us (Stephen Jacobs) writes:
>>
>>There are a lot of questions of the form "How do I get the ST to do job X
>>in language Y".  These are well answered with small example programs.

I agree this would be a wonderful thing.  No, I can't do it.
I don't own a site, nor have the storage, and I keep moving.
I just wanted to say that I think that good indexing will be
crucial to its usefulness.  I can imagine stuff going in but
never coming back out because users can't find what they're
looking for--then it would be no better than the net now.

I suggest that the manager of the archive will have to read
each submission (I doubt there'd be more than one or two per day),
anticipate its usefulness, and assign appropriate
keywords.  (This would be important and take some skill.)
The new article number (or whatever naming scheme is used) and
the keywords for that article can then be added together on one line
to a global index file.  Users can `ftp` the index file and then do a
`grep '(key1)(key2)(key3)' index` kinda command to get just the
files they want.  E.g., `grep '(hirez)(vdi)' index | grep -n '(window)'`
would list all the files related to vdi calls on monochrome systems
which are not related to using windows.  Okay, so it's not a very
practical example, but you get the idea of the flexability and power
once the indexing is done and the keywords are made available.
Regularity of the index would be crucial, so using the supplied
keywords would be unreliable.

It doesn't have to be just source code either.  I've seen some
awfully useful little chunks of text floating by here that I was
sure I'd want later, but knew I'd never be able to find again
even if I tried to archive them.

Of course, a server should be available to do the `grep` locally
and return the results to those who only have mail capability.

-- 
--
J. David Beutel  11011011  jdb9608@cs.rit.edu      "I am, therefore I am."
Looking for co-op Sept..Feb in NN, OS, VR, or AI.