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 Brought to you by Super Global Mega Corp .com