- From: Danny Ayers <danny@panlanka.net>
- Date: Tue, 20 Feb 2001 20:38:33 +0600
- To: "Aaron Swartz" <aswartz@swartzfam.com>, "RDF Interest" <www-rdf-interest@w3.org>
<- >> For our simplified testbed, we can just assume that an <- <- application will <- >> not be doing inserts based on reasoning. <- > Couple of questions, I guess I'm missing something - what <- kind of inserts <- > will be possible? What will it be testing? (I would have thought doing <- > inserts based on reasoning would be the whole point) <- <- I mean based on reasoning from the triples themselves -- this is <- to prevent <- infinite loops. I'm thinking of a sort of decentralized storage <- system where <- the information comes from some outside source (say, a user) and <- is stored <- in the decentralized network. <- <- -- <- [ Aaron Swartz | me@aaronsw.com | http://www.aaronsw.com ] <- So the inference mechanism would be centralised? If so then I suppose the problem here boils down to propagating the updates in a way that maintains the transaction idea. The effects that (potentially) lead to infinite loops could be a devil to calm (a distributed inference system could cause the same problems) - though as you seem to be swuggesting, one step at a time... BTW, were you thinking of hacking the inference mechanism together or using an existing tool/shell (e.g. Jess)? Cheers, Danny. --- Danny Ayers http://www.isacat.net
Received on Tuesday, 20 February 2001 09:40:36 UTC