Xref: utzoo comp.ai:7916 news.misc:5606 Path: utzoo!attcan!uunet!tut.cis.ohio-state.edu!att!emory!gatech!purdue!tippy!sawmill!mdbs!zed From: zed@mdbs.uucp (Bill Smith) Newsgroups: comp.ai,news.misc Subject: Re: natural language recognition suitable for netnews? Message-ID: <1990Oct27.203714.1468@mdbs.uucp> Date: 27 Oct 90 20:37:14 GMT References: <45512@apple.Apple.COM> <1990Oct12.105710.29940@chinet.chi.il.us> <1990Oct19.112936.6561@chinet.chi.il.us> Reply-To: zed@mdbs.UUCP (Bill Smith) Organization: mdbs, Inc. Lines: 51 In article <1990Oct19.112936.6561@chinet.chi.il.us> laird@chinet.chi.il.us (Laird J. Heal) writes: >In article emv@math.lsa.umich.edu (Edward Vielmetti) writes: >>Cross-posted Followups to comp.ai. [and news.misc where it started] > >>In article <1990Oct12.105710.29940@chinet.chi.il.us> laird@chinet.chi.il.us (Laird J. Heal) writes: > >> What really needs to be done is to redirect inappropriate articles into >> more appropriate newsgroups, as with the reposted articles comprising >> comp.archives. > >>Not exactly. comp.archives does [about what I said, but I emphasize reposting] > >>I think it's possible to build an expert system (or [...]) [...] for >>comp.archives, and by extension any other group with a similar narrow >>focus. > >I was thinking _prospectively_; that is, have inews categorize the article or >carry out the recategorization done by upstream sites. It is clearly untenable >to flip control messages seven ways from Sunday reclassifying articles; >canceling them is hard enough. > >However, in Eric Fair's article starting this thread, he gave a very clear >opinion that better news presentation tools are required. Another of my ideas >would be to have the news readers add keywords. Then, an expert system proper >could base its censorship on what the current reader has found interesting in >the past correlated to the choice of keywords. > >I would prefer to be able to read all of the postings on topics of my interest >by reading particular newsgroups. Rhat means requesting that those posting the >article post to whichever newsgroup is appropriate, and that unfortunately is >not always done. What I generally do is find the time to read news and strip >out the interesting articles, which I later go through and make use of. >-- >My .signature is on vacation ------------- like me! Why don't you people call TRW at Redondo Beach. They already have build hardware and software that will make categorization unnecessary. It is cool and it works. The group that built that project deserves more public recognition. Basically it is called a Fast Data Finder. It searches through data with (arbitrary) regular expressions as fast as the disk array that feeds it can ship it through. It can search through an entire months worth of the net in a second or so. I haven't heard what the current status is, so by now it probably has perfectly friendly user interface. If it doesn't I'm sure someone on the net will volunteer to adapt their (already written) pet project to the task. Bill Smith pur-ee!mdbs!zed [Keep on smiling. Otherwise, people might think you are an axe murderer.]