Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!mips!wdl1.wdl.loral.com!wdl1!mikeb From: mikeb@wdl31.wdl.loral.com (Michael H Bender) Newsgroups: comp.ai Subject: Re: Mercury? Message-ID: Date: 7 Jan 91 20:59:38 GMT References: <1991Jan5.040204.4600@cbnewsl.att.com> <5194@trantor.harris-atd.com> Sender: root@wdl1.wdl.loral.com (SUPER USER) Organization: Ford Aerospace, Western Development Laboratories Lines: 15 In-Reply-To: jgronosk@x102c.harris-atd.com's message of 7 Jan 91 14:20:11 GMT Nntp-Posting-Host: wdl31 With regard to Mercury and Gensym -- I have used neither so obviously I am an expert and can thrown my opinion in (who knows -- it may even be worth 2 cents!). In the past I have looked at both systems (superficially) And here is the conclusion I arrived at at the time: They are designed for different classes of applications. Gensym seems to be oriented towards model-based reasoning whereas Mercury is a more typical expert-system shell. In other words, if I had to build a knowledge-based system based on a physical/mechanical system, I would definitely use Gensym. However if I had to build an inferencing system that used frames, forward chaining, and backward chaining, I would look towards Mercury first. Mike Bender