Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!emory!gatech!prism!gt4084c From: gt4084c@prism.gatech.EDU (SRINIVASAN,K) Newsgroups: comp.object Subject: Re: Has anyone taken an approach like this? Message-ID: <24705@hydra.gatech.EDU> Date: 20 Mar 91 00:45:07 GMT References: <1991Mar19.042444.20834@netcom.COM> Organization: Georgia Institute of Technology Lines: 40 In article <1991Mar19.042444.20834@netcom.COM> rodent@netcom.COM (Ben Discoe) writes: > >I'd be interested to know: > 1. Has anyone ever seriously taken this approach? Yes, Wilensky proposed KODIAK representation scheme which does exactly what you had described - you only have to replace 'objects' in your posting with 'frames.' > 2. Are there any obvious drawbacks? The major objections I see are: 1. It does not fit the classical definition of objects as nouns from domain description (of course Booch, Rumbaugh, etc., do not seem to insist on this requirement any more). 2. Lack of cognitive correspondence - we think in terms of entities and not in terms of relationships (Wilensky refutes this argument at length). > 3. This method seems appropriate for modeling anything from an encyclopedia >to a rainforest to a coin collection. Why can't I find anyone else pursuing ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ >it? I've scanned all the papers and conference proceedings I can get my ^^ >hands on. You sure do have some enviable company! > 4. Am I being foolishly naive and grossly underestimating the potential > difficulties? With the current representation, it is indeed a choice between redundancy and brute force as you had said. The Cyc people (the only large implementation attempt well documented and pubilicized) seem to have chosen redundancy. I am really curious why they settled for it rather than trying something like what you, Wilensky, etc., have described. >Ben in San Jose, idealistic visionary, computer scientist, ecologist. -- SRINIVASAN,K School of Textile Engineering Georgia Tech. uucp: ...!{allegra,amd,hplabs,seismo,ut-ngp}!gatech!prism!gt4084c ARPA: gt4084c@prism.gatech.edu