Path: utzoo!attcan!uunet!tektronix!percival!qiclab!sopwith!snoopy
From: snoopy@sopwith.UUCP (Snoopy)
Newsgroups: comp.society.futures
Subject: Re: The public's network of the future
Message-ID: <75@sopwith.UUCP>
Date: 21 Dec 88 18:41:11 GMT
References: <255@ultb.UUCP> <KARL.88Dec13170438@triceratops.cis.ohio-state.edu>
Reply-To: snoopy@sopwith.UUCP (Snoopy)
Organization: The Daisy Hill Puppy Farm
Lines: 39

In article <KARL.88Dec13170438@triceratops.cis.ohio-state.edu> karl@triceratops.cis.ohio-state.edu (Karl Kleinpaste) writes:
|jdb9608@ultb.UUCP (J.D. Beutel ) writes:
|   The volume of data carried on Usenet doubled, from 2 mb/day to 4 mb/day,
|   last year.  I wonder where it's all heading.  I'd like to hear
|   some informed speculation.  Wha'd ya think?
|
|There's a 3B2/400 in my basement that has 140Mbytes of disc and a
|modem on it.  I pay my own electric and phone bills.  As long as that
|situation continues, that machine can get a full newsfeed.  The volume
|would have to more than octuple before I would be concerned.  That
|gives me a minimum of 3 years before I even get nervous - and by then
|I'll have some higher-capacity hardware to accommodate it.
|
|--Karl

Whether the machine can handle the load is only one part of the problem.
The other part is the human interface.  There is a lot of useful information
in usenet, but it is often buried in garbage.  How much time do you want
to spend skimming through articles hitting 'k' or 'n'?  KILL files help,
but not enough.  Moderation sometimes helps, but often the moderator bends
a group far too strongly in the direction of their own interests.

What do you do with articles that you might want to refer to later?  Print
them out?  Ugh!  Save them on disk?  Uses up lots of disk space fast, leading
to putting them on tape where they aren't very handy.  After awhile, even
saving directories of the tapes gets to be huge.  And when you want to
find that article that mentioned 'foo'?  "egrep foo *" takes awhile.

We need some method of automatically selecting articles that is an order of
magnitude or two better than KILL files.  Any ideas?  Could AI help here?
We need some method of efficiently storing useful information for later
use.  Something better than throwing an article on disk with a couple of
keywords in the filename.

    _____     
   /_____\    Snoopy
  /_______\   
    |___|     tektronix!tekecs!sopwith!snoopy
    |___|     sun!nosun!illian!sopwith!snoopy