pjc@cs.man.ac.uk (Peter Crowther (CAG ra)) (10/29/90)
[First posting in over six months - apologies for stylistic cock-ups.] The research project of which I am a member has used Prolog for a pilot database allowing content retrieval of images. Unfortunately, the system we have will not scale well to large systems involving thousands of images - all Prologs we know have VM-resident databases, and Prolog (NOTE: NOT Parlog or anything else) is inherently tuple-based rather than set-based. We could give up some of the flexibility of Prolog (cuts, infinite set generators such as father(god, _), anything that makes it tuple-based) in order to get set-based processing such as that found in relational databases (RDBs) or tuple stores. We would like to keep a Prolog-like syntax. Datalog appears to be a (fairly) well-established Prolog-like language that can be implemented on top of RDBs. We do not have the manpower to implement Datalog. Therefore, does anyone know of any existing Datalog implementations, or implementations of other systems that might meet the requirements outlined above? Cost is, at present, not an issue; we are looking for any system. Reliability is not paramount either - the odd core dump (yes, we're on UN*X) isn't a disaster as long as we have enough information to fix it. Not having to build it ourself is the biggest constraint. My thanks for any information you may have. - Peter P.S. Known projects are MCC's LDL, Stanford's NAIL!, Birkbeck's Exegesis and ALGRES, which is a part of the METEOR Esprit project (hence the eunet.esprit crosspost). -- Peter Crowther, Multimedia Research Group, Dept. of Electrical Engineering, University of Manchester, Manchester M13 9PL, England. Internet: pcrowther@cs.man.ac.uk Janet: pcrowther@uk.ac.man.cs If those fail, try: pcrowther%cs.man.ac.uk@nsfnet-relay.ac.uk