- From: <ruediger.klein@daimlerchrysler.com>
- Date: Thu, 6 Jun 2002 14:52:14 +0200
- To: <horrocks@cs.man.ac.uk>
- Cc: <www-webont-wg@w3.org>
horrocks@cs.man.ac.uk 06.06.02 14:43 Bitte antworten an horrocks An: Ruediger Klein/FT/DCAG/DCX@WK-EMEA2 Kopie: www-webont-wg@w3.org Thema: Re: Antwort: Re: WOWG ADMIN IMPORTANT: Issue list: cleanup of ISSUES 2.1-3.4 On June 6, ruediger.klein@daimlerchrysler.com writes: > Hallo Ian: > > from a applicational point of view both enumerated classes (one-of) AND inverse > are very necessary (not to say indispensible!) Hi Rudiger, First off, I would like to make it clear that my remarks weren't intended to indicate a position as to whether inverse and/or oneOf should/should not be included in the language - I was only trying to inform the discussion. > > If both features together makes reasoning much more complicated - I don't know > what to do. > > Can we find a way which allows the user to use both features which will (in a > normal case or so) NOT result in reasoning complications? > For instance, in many cases inverse relations will be used only as a kind of > syntactic sugar (no additional semantic provided in comparison to the original > relation). But can users not familiar with DL understand such aspects of > modeling? I'm not sure what you mean by syntactic sugar in this context. Perhaps you mean that asserting either [x R y] or [x inv-R y] is taken to be equivalent to asserting both, and that otherwise there is no semantic relationship between R and inv-R? I believe that this should be possible from a technical perspective (Grail uses a similar device), but might not meet user requirements. what I mean is that a user may want to assert [y inv-R x] what is logically equivalent to [x R y] but s/he has a kind of "object oriented view" and wants to say something about y and its relation to x. Regards, Ian > > Regards > > Ruediger
Received on Thursday, 6 June 2002 08:52:51 UTC