- From: Jeff Z. Pan <jpan@csd.abdn.ac.uk>
- Date: Thu, 15 Nov 2007 21:14:03 -0000 (GMT)
- To: "Jeremy Carroll" <jjc@hpl.hp.com>
- Cc: "Bijan Parsia" <bparsia@cs.man.ac.uk>, "Carsten Lutz" <clu@tcs.inf.tu-dresden.de>, "Web Ontology Language Working Group WG" <public-owl-wg@w3.org>
Hi all, Interesting discussions. Some comments: 1) Mathematical datatypes and XML Schema datatypes are both needed. 2) There exists a common subset of the two kinds of datatypes, including xsd:Integer. 3) It might be an idea to consider additional mathematical datatypes such as owl:real and owl:rational. The lexical forms of these datatypes might need some work and discussions. 4) For rounded datatypes, users don't actually consider precise equivalence, but approximate equivalence [1] instead, such as type promotions used in XPath 2.0 [2]. Before using values of rounded datatypes in reasoning, one might want to transform them back to unrounded datatypes first. Greetings, Jeff [1] http://www.w3.org/TR/swbp-xsch-datatypes/#sec-use-sparql [2] http://www.w3.org/TR/xpath20/#promotion On Thu, November 15, 2007 12:36 pm, Jeremy Carroll wrote: > > Skimming this thread somewhat - I believe the RDF Datatypes design is > meant to be open to non-XML Schema datatypes defined by fiat, which may > allow a WG to do what Bijan appears to be discussing - e.g. define an > owl:integer an owl:real, owl:rational, owl:complex datatypes .... > > > It may even allow for an owl:real datatype whose value space includes > all the reals, even those that can't be written down. > > It would be another little project though, probably out-of-charter > (certainly, if the timeline is taken into account). > > > Jeremy > > > > > >
Received on Thursday, 15 November 2007 21:30:19 UTC