- From: Libby Miller <libby@asemantics.com>
- Date: Fri, 14 Oct 2005 09:18:36 -0700 (PDT)
- To: public-swbp-wg@w3.org
- cc: hobbs@isi.edu, pan@isi.edu
This message provides my comments on the editor's draft of the Note "Time Ontology in OWL"[1] and "Time Zone Resource in OWL"[2] I'm sorry this is rather later than I'd hoped - I hope it's still useful. I am not an expert on time ontologies but I have worked for a long time on the Semantic Web Interest group's taskforce on Calendaring, which has been converting RFC 2445(iCalendar) to RDF/OWL.[3],[4] In my opinion it would be good to try and get Pat Hayes to review these documents if he has time - I know he put valuable input into the DAML Time ontology[5]. Dan Connolly would also be a very useful reviewer if he has time - he has worked extensively on the SWIG calendaring taskforce and has implementation experience there. -- "Time Ontology in OWL" Overall This is a huge and difficult subject and I appreciate the authors' work in scoping the issues to a manageable subset. I think it's very important to provide usecases as the authors have, although I think the document requires several more to be compelling. It would help in particular to emphasise the benefits of an ontology over, for example, an xml schema; or at least clarify the circumstances where advantages would occur. It would also be useful to show some examples which illustrate this ontology mapped to iCalendar, if only for the reason that iCalendar is very commonly used in PIM applications. I like the distinction made between Topological Temporal Relations, and Duration Description and Calendar-Clock Description, distinguishing (I think) between the ontological constructs and the descriptions we give to them (e.g. in calendars). Minor nits: It would be better for sections to be numbered to allow easier referencing. Typos: "meeing" References need filling in. Details Topological Temporal Relations "Allen's temporal interval calculus" - I'm not familiar with this - I think a bit more explanation would be useful here, especially as the reference is not online. Time Zones Language is a bit vague: "Probably seconds are not relative to the time zone." Use Cases for Time in OWL-S More would be good (and more general, non-OWL-S usecases, unless the ontology is specific to that area). [[ In this example Instant, a subclass of TemporalEntity, would be a better class to use than TemporalEntity to describe CreditCardExpirationDate, because the expiration date is actually an instant -- the midnight, of the day the credit card expires. (This was already changed in the later release of OWL-S.) ]] The example should be updated to reflect this. ------- "Time Zone Resource in OWL" Overall It would be good for the authors to take a look at "Working with Time Zones"[6] which is a very clear description of the issues developers commonly run into here. "GMT" should probably be replaced with "UTC" throughout: [[ Greenwich Mean Time is a widely used historical term, but one that has been used in several ways. Because of the ambiguity, its use is no longer recommended in technical contexts. ]][7] The authors might want to consider having a last updated property or similar to mitigate cases where daylight savings policies change. Minor nits: It would be better for sections to be numbered to allow easier referencing. Libby [1]http://www.isi.edu/~pan/SWBP/time-ontology-note/time-ontology-note.html [2]http://www.isi.edu/~pan/SWBP/time-zone-note/time-zone-note.html [3]http://www.w3.org/2002/12/cal/ [4]http://www.w3.org/TR/rdfcal/ [5]http://www.daml.org/listarchive/daml-time/0031.html [6]http://www.w3.org/TR/2005/NOTE-timezone-20051013/ [7]http://aa.usno.navy.mil/faq/docs/UT.html
Received on Friday, 14 October 2005 16:19:41 UTC