Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!mcvax!cernvax!ethz!ethz-inf!wyle From: wyle@inf.ethz.ch (Mitchell Wyle) Newsgroups: news.misc Subject: In Moderation: A Developer's Response Message-ID: <273@ethz-inf.UUCP> Date: 14 Jun 89 17:39:11 GMT References: <41197@bbn.COM> <12113@well.UUCP> <41207@bbn.COM> <632@biar.UUCP> <3305@epimass.EPI.COM> <2727@nmtsun.nmt.edu> Reply-To: wyle@ethz.UUCP (Mitchell Wyle) Organization: Departement fuer Informatik, ETH Zuerich Lines: 74 Hey people, Like, the way it was... In 1945 W. Bush wrote an article which mentioned a fantasy machine called "memex." That article is the basis for the fields of databases, hypertext, and information retrieval (another good read). ...the way it is... >>The reason we're seeing a lot of proposals like Geoff's and Brad's at >>this time is because there's a demand for it. The net has simply >>become too large. A better response to these proposals is to come up >>with competing proposals that would be free. >>-- Joe Buck jbuck@epimass.epi.com, uunet!epimass.epi.com!jbuck > I have been working on a system like this for almost 6 months me too. Mine is a selective dissemination of information (SDI [1]) service based on WAN information, not just usenet. >The basic idea behind my approach is that people can "locally" >(defined by cheap communications) communicate recommendations of good >articles, and readers can use these recommendations to select >articles. Added to this would be a number of national-level >moderators who would select articles from local areas for a national >recommendation, and add keywords to the article. Moderation would not >be a paying activity, but many people would still moderate for ego >reasons. You, like Brad and Geoff want to use SEMANTIC (human-based) indexing of usenet articles with the usenet "here is everything" model of information flow to the user. I want to use statistical text analysis, user profiles, and automatic indexing to pick out the relevant information for each user based on his information needs and desires. There is a general ignorance of the science of information retrieval (IR) among computer people. I want to keep this article short, so if you want more references, more info, or whatever, send mail. A good start for info about IR is: %A G Salton %A M McGill %T Introduction to Modern Information Retrieval %I McGraw Hill %C New York %D 1983 Information Retrieval is sometimes refered to as "full text retrieval" and sold as "commercial database" systems with boolean key-word queries. There is an ACM Special Interest Group (SIGIR) with an annual conference and lots of books, papers, and literature has already been published. I am doing IR research using usenet articles as a test collection. A paper about the system I am developing will appear in the proceedings of ACM SIGIR89 in Boston. The system is live in a small alpha test. I hope to go beta this year. Let me know if you want to subscribe or have ideas to make the system better, etc. >I am interested in hearing from anyone who is doing similar, >or related work, or anyone interested in programming this beast. I am >also interested in any opinions about this. me too. ---------------------- [1] SDI as an acronym for selective dissemination of information was coined in 1958 by Luhn. Star Wars came much later. Lay back, relax, stay cool, hack code; like, the way it shall be... -- -Mitchell F. Wyle Institut fuer Informationssysteme wyle@inf.ethz.ch ETH Zentrum / 8092 Zurich, Switzerland +41 1 256 5237