Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.kei.com!newsfeed.internetmci.com!in1.uu.net!siemens!darken
From: darken@scr.siemens.com (Christian Darken)
Subject: Re: Look Who's Talking
Message-ID: <DIvD1H.ELA@scr.siemens.com>
Sender: news@scr.siemens.com (NeTnEwS)
Nntp-Posting-Host: avocet.scr.siemens.com
Organization: Siemens Corporate Research, Princeton NJ
Date: Thu, 30 Nov 1995 18:36:52 GMT
Lines: 20

>I want to program a character that the users can talk with.  I have 
>studied some old eliza programs and made a similar one on a mud, but 
>the character is just not good enough to fool anyone.
>The principle I am using is simple word or phrase detecting and giving
>a response that hopefully fits out of an array of possible responses.

If you haven't already, you might want to check out Michael "Fuzzy"
Mauldin's (yes, the author of the venerable rogue-playing program
Rog-O-Matic) work along these lines.  His mud robot Julia (which also
went by other names) has had good success fooling users and has been
entered in Loebner's Turing Test contest.  And some code (perhaps
dated) has recently been available in mud archives (see the
rec.games.mud FAQ). He has published at least one paper on the robot
and the results of the contest.  Mauldin's home page is:

http://fuzine.mt.cs.cmu.edu/mlm/

Look for the "Chatterbots" paper.


