Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!zaphod.mps.ohio-state.edu!uwm.edu!psuvax1!rutgers!aramis.rutgers.edu!paul.rutgers.edu!moskowit From: moskowit@paul.rutgers.edu (Len Moskowitz) Newsgroups: comp.ai Subject: Re: OPS5 on PC's. Message-ID: Date: 21 May 91 13:27:14 GMT References: <12850@rnd.GBA.NYU.EDU> Organization: Rutgers Univ., New Brunswick, N.J. Lines: 34 A few possibilities: 1. There used to be a version called OPS5+ from Computer*Thought Corporation (1721 W. Plano Parkway, Suite 125, Plano Texas 75075 (214-424-3511). It was reasonably complete except that it didn't use Rete so that on large rule/data sets it slowed down horribly. I last contacted them ijn 1986 or '87 so I can't say if they're still around. 2. There are at least three C versions of OPS5 out these days. CMU (Anurag Acharya, Milind Tambe, Dirk Kalp), U of Texas at Austin (Dan Miranker and company), and Columbia U (Sal Stolfo and company) all have different C-based OPS5s. (The CMU people have a version of theirs for a shared-memory parallel processor.) My group at Bendix TSD ported the CMU uniprocessor version to the Suns in very short order. Porting any of these to a PC should be relatively straightforward. 3. You might consider using NASA's CLIPS rather than OPS5. The CLIPS distribution system includes a version for PCs. CLIPS is much like OPS5 but has a much more versatile pattern matcher and some other useful features. It comes with full source code so you can hack to your heart's content. It's available from COSMIC down at U of Georgia (Athens, GA). 4. Charles L. Forgy, the original author of OPS5, has a product called OPS83 that has a PC port. His company is called Production Systems Technologies in Pittsburgh. Len Moskowitz Allied-Signal Aerospace Co. Bendix Test Systems Division Teterboro, NJ