Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/17/84; site hao.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!hao!woods
From: woods@hao.UUCP (Greg Woods)
Newsgroups: net.news
Subject: Re: Keyword based news
Message-ID: <1787@hao.UUCP>
Date: Thu, 3-Oct-85 12:17:20 EDT
Article-I.D.: hao.1787
Posted: Thu Oct  3 12:17:20 1985
Date-Received: Sat, 5-Oct-85 07:21:29 EDT
References: <16460@watmath.UUCP> <1419@utcsri.UUCP> <132@mck-csc.UUCP> <3189@nsc.UUCP> <265@ukecc.UUCP>
Organization: High Altitude Obs./NCAR, Boulder CO
Lines: 33

> 	But in order for all this to work, you must still rely on posters
> to label thier postings with the proper keywords. What's to prevent some
> malicious type person from labeling a particularly offensive posting
> with "Keywords: sex UNIX pontiac"?

  Nothing, of course. This is only ONE problem with keyword-based news:
deliberate bad keywords. We see this type of thing with newsgroups, too,
where articles get posted in inappropriate places (such as anti-gay flames
in net.motss). I think it would be much worse when the number of "groups" 
(i.e. keywords) increases virtually without bound. Let's put it this way:
suppose I think a particular issue is VERY important, and I post my $0.02
worth and label it "information". Someone else who disagrees with me might
think it should be laballed "opinion", or even "flame". I could do this
deliberately, to try and get more people to read my article (how many people
would actually READ an article that admits right off to being a flame?).
This could also occur accidentally; i.e. I could totally non-maliciously
label my article as information when someone else perceives it as a flame.
Just look at the furor that started from a totally innocent (at the time)
remark about emotions and choice in net.singles. I bet poor Gypsy never
realized she was posting something that would offend so many people. That
is just ONE example. That's the second problem with keyword-based news: 
accidental, or non-malicious, incorrect keywords. Experience has shown that
users CANNOT BE COUNTED ON to choose correct keywords. They can't even
choose correct newsgroups. What happens when we have thousands of keywords
instead of a couple hundred newsgroups? The third problem was conceptual
problems with keyword-based systems, i.e. you can miss articles you wanted
to see and/or be shown articles you didn't want even when "correct" keywords
are chosen, because of problems inherent in keyword-based systems (e.g.
"phone" vs. "phones" vs. "telephone", etc.) Lauren already addressed this
issue in a previous article. Keyword-based news is a BAD idea. It only 
INCREASES the problems associated with newsgroups.

--Greg