Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!gatech!ukma!david From: david@ms.uky.edu (David Herron -- One of the vertebrae) Newsgroups: news.software.b Subject: Re: Review of NN, a Usenet news reader Keywords: rn, .newsrc, b&d Message-ID: <12076@s.ms.uky.edu> Date: 6 Jul 89 14:26:05 GMT References: <1836@papaya.bbn.com> <1150@sequent.cs.qmc.ac.uk> <1989Jul6.024247.21653@DSI.COM> <89Jul6.071651edt.10388@neat.ai.toronto.edu> Reply-To: david@ms.uky.edu (David Herron -- One of the vertebrae) Organization: U of Kentucky, Mathematical Sciences Lines: 31 How can using subject lines to select articles be good when articles often have little to do with their subject line. No, the correct way to select articles is with keywords, but the rub is that it's hard to pick out good keywords. People on the net have made references to studies which've proven that it's hard for even trained keyword-ists to pick out good keywords (The referenced paper concentrated on law journals, so is there something uniquely different between law journals and Usenet?) it's hard for computer software to pick out good keywords. Indeed! Picking out good keywords probably requires understanding of the semantics present in the text. This certainly hasn't, yet anyway, been encoded into a computer program. Anybody from the NSA care to comment? ;-) oh, this article is supposed to be about NN since that's what the Subject: line says. But I can't think anything about it since I haven't used it yet. However, it was NN which started off this train of thought, and it's kind of globally about NN since I'm concerned that the only article selection method, as I understand it, in NN is the subject line. But then I kill articles based on subject lines all the time ... sigh -- <- David Herron; an MMDF guy <- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET <- <- New word for the day: Obnoxity -- an act of obnoxiousness