- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Thu, 17 Nov 2005 15:41:41 +0000
- To: Feng Pan <pan@ISI.EDU>
- CC: public-swbp-wg@w3.org, Jerry Hobbs <hobbs@ISI.EDU>
Feng Pan wrote: > > Hi Jeremy, > > Thank you very much for your comments on the time ontology note! Please > see my reply below. > >> 1) duration >> >> In the migration from DAML to OWL, the WG identified that the semantics >> of xsd:duration was unusable, and say that it *SHOULD NOT* be used. >> >> See >> http://www.w3.org/TR/rdf-mt/ >> >> (note this text originated in WebOnt) >> >> The SWBPD's datatype note addresses this issue >> >> http://www.w3.org/TR/2005/WD-swbp-xsch-datatypes-20050427/#section-duration >> >> >> The time ontology should be updated in line with the advice there. > > Thanks for pointing out this issue. We will add a warning below our > "durationDescriptionDataType" property, and say that: > > The [RDF Semantics] Recommendation argues that xsd:duration does not > have a well-defined value space, and should not be used. For example, it > can not provide with an unequivocal answer to the question "How many > days in a month?". > I am not sure here. It is not sufficient to merely add a warning. The use of duration is a bad practice, and should not occur in any of our documents. I would strongly prefer that the time ontology simply does not use the duration datatype. Of course, using the two replacement types needs some explanation, but they should not be presented as an alternative to duration, on some sort of equal footing, but more as the preferred practice; with the use of duration being a non-practice. (the SHOULD NOT is a strong statement in RDF Semantics, I have not seen any argument as to why this note should in anyway overall, and choose to use duration) Jeremy
Received on Thursday, 17 November 2005 15:44:11 UTC