- From: Alan Ruttenberg <alanruttenberg@gmail.com>
- Date: Tue, 30 Oct 2007 19:59:54 -0400
- To: Jeremy Carroll <jjc@hpl.hp.com>
- Cc: Boris Motik <boris.motik@comlab.ox.ac.uk>, public-owl-wg@w3.org, Alan Rector <rector@cs.man.ac.uk>
On Oct 30, 2007, at 10:20 AM, Jeremy Carroll wrote: > Alan Ruttenberg wrote: > >> Proponents of punning would label this: >> "OWL DL, now with more, but not all, of OWL Full goodness" > > To some extent this is about labelling .... Mostly, I think. > I have been surprised how strongly some of my colleagues feel about > this issue - my original take, and still my personal prejudice (but > not the position I represent) is that punning is simply an area in > which the DL implementations are incomplete. I would be interested in hearing their views. Any way we can arrange that? > I suspect, but have not checked, that many of the WG would be > unhappy with such a position though - and since it would take me > some effort to convince HP of it - I haven't given either task much > thought. I don't know about the WG. You would know about HP. But I think it would be good to surface all the thinking about this matter so we could take it in to consideration. I found Evan's note about chaining data values quite helpful, for instance, and perhaps this leads to an easily implementable improvement. > i.e. > - semantically using the same name for say, a class and an > individual, has the consequences that one might expect from OWL Full > - the OWL DL profile is incomplete, in that, consequences resulting > from punning are not computed (but still legal OWL consequences). > i.e. an OWL DL reasoner may implement punning in the way that OWL > 1.1 reasoners currently are doing so, but the resulting (lack of) > entailments is incomplete. I presume that this is a restatement of the proposed situation. If so, I think it is accurate. -Alan
Received on Wednesday, 31 October 2007 00:00:08 UTC