- From: pat hayes <phayes@ai.uwf.edu>
- Date: Fri, 21 Feb 2003 14:43:32 -0600
- To: Graham Klyne <GK@NineByNine.org>
- Cc: w3c-rdfcore-wg@w3.org
>At 05:17 PM 2/18/03 +0000, Brian McBride wrote: >>At 11:16 18/02/2003 +0000, Graham Klyne wrote: >>>At 10:43 PM 2/13/03 +0000, Brian McBride wrote: >>> >>>>At 21:32 13/02/2003 +0100, Jeremy Carroll wrote: >>>> >>>>>I am posting this message to three lists, sorry for duplicate copies. >>>>> >>>>>There has been a significant discussion on the social meaning >>>>>parts of the RDF Concepts Last Call. >>>> >>>>Really! Where? >>> >>>There's been some discussion on WebOnt, starting here: >>> http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/0280.html >>> >>>Also, it doesn't count as discussion, but I indicated a position here: >>> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jan/0228.html >>> >>>But I'll agree with you that there's not yet been sufficient >>>discussion to see where this is going. >> >>Thanks Graham, the pointers were very useful. >> >>Jeremy is suggesting - lets see if we can find a form of words that >>satisfies everyone. I'm hoping that doesn't mean fudging the >>issue. I'm also very concerned that you won't be present if there >>is a f2f discussion at the tech plenary. Maybe we should be >>thinking about dialing you in? >> >>I'm also wondering about laying some of the groundwork. I'm seeing >>a lot of very unstructured discussion, and I fear there is great >>risk of confusion clouding the discussion. >> >>Do you and Jeremy have any ideas on how we might best >>prepare/clarify the question. > >I don't think we need to compromise anything here -- I think most of >the concerns expressed have been concerns of the WG, and the problem >has been in the explanation. In hindsight, I think the example we >used has not been helpful, and I'd like to drop it. As originator of the example, I agree. It was invented for use in a talk, where its shock value had some rhetorical purpose. As part of a spec document it probably raises way too many dangerous-seeming issues. We can make the essential point with a much blander (and simpler) example if necessary, eg A defines a class name and says something informal but significant about it (such as : Frobs; we can deliver these in one working day) , B uses it to assign a type, (#g000345 rdf:type A#Frob) then if the informal meaning applies to the class name (we assume A is telling the truth) then it has to be understood as applying to the instance name (A said he could deliver #g000345 in one working day). In other words, doing a formal inference doesn't break any social meaning. Kind of obvious, right? Pat > >Here are some thoughts I have: > >[[ >I think the issue of social meaning is poorly handled, and needs to >be improved. I think the main point we need to convey here is that >there may be social meaning associated with some RDF that is opaque >to automated reasoning processes. The secondary point is that such >meaning may be embodied in some collection of RDF statements, and >those statements may be obtained by application of a logically valid >reasoning process. But there is no intent that RDF agents somehow >need to be aware of the social meaning. >]] > >There is a separate issue of how a URI gains its meaning, which I >think should be handled separately. > >#g > > >------------------- >Graham Klyne ><GK@NineByNine.org> -- --------------------------------------------------------------------- IHMC (850)434 8903 or (650)494 3973 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32501 (850)291 0667 cell phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes s.pam@ai.uwf.edu for spam
Received on Friday, 21 February 2003 15:43:37 UTC