- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Fri, 6 Mar 2009 14:07:31 +0000
- To: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
- Cc: W3C OWL Working Group <public-owl-wg@w3.org>
On 6 Mar 2009, at 13:51, Ian Horrocks wrote: > This seems to me to be a rather similar situation to the one with > Marijke, and we could make similar replies in the two cases: > > - even if functional datatype/key properties are "the norm", one > can easily imagine situations where modellers may want to use non- > functional properties (e.g., Michael's example of multiple email > addresses, where these are represented as a string); > - it is trivial for modellers to assert functionality if desired; > - we do not consider it to be appropriate for the WG to use the > design of the language to impose subjective decisions about > modelling style. > > The only coherent argument that I can see on Ralf's side is an > implementation one -- functional datatype properties would make > implementation easier. This seems to be a rather week argument, > particularly as several implementations already support general > datatype properties. And, frankly, no other implementor is calling for this, afaik. Given a choice of changing 3 or more implementations and changing 1...well.... Cheers, Bijan.
Received on Friday, 6 March 2009 14:03:59 UTC