- From: tim finin <finin@cs.umbc.edu>
- Date: Thu, 19 Sep 2002 22:52:22 -0400
- To: www-webont-wg@w3.org
I think this issue is very important but that it is also one where we don't have clear ideas of what's needed, desired, useful, practical, etc. The problem of keeping track where rdf triples came from can quickly evolve into much more complicated problems of provenance and justifications. These are compounded by the promiscuous environment of the web, where links are cheap. So, my gut tells me to keep the issues out of the language for now and let people play with ideas and applications that make use of them. If we offer a simple model of what import means (e.g., syntactic inclusion) it is just a model. My program might do something much more elaborate, keeping track of the (perhaps multiple) justifications for why the triple is (or is not) in the KB and possibly deriving interesting properties (e.g., trust) from these. tim
Received on Thursday, 19 September 2002 22:53:10 UTC