RE: A Last Call comment about the seamtnics of xsd:dateTimeStamp in OWL 2 (caused by the yesterday's decision about numerics)

Hello,

I'm sorry for the religious metaphor in my earlier e-mail :-)

I believe that I've found further arguments for changing our treatment of
xsd:dateTime: our current spec is incompatible with the date-time functions from
XQuery. In particular, the document

http://www.w3.org/TR/xpath-functions/

defines in Section 10 several functions on the xsd:dateTime datatype (this
includes then xsd:dateTimeStamp). For example, the fn:timezone-from-dateTime
function takes a date-time value and returns its time zone.

For this function to be well-defined, the time zone information needs to be
present in the actual value in the value space. Consider the following example
of a literal "1st of March 2009, 3am GMT". In XML Schema, this literal is mapped
to the following 7-tuple in the value space:

year=2009
month=3
day=1
hour=3
minute=0
second=0
timezone=0

The fn:timezone-from-dateTime function now simply returns the 'timezone' element
of this structure.

In our version of the value space of xsd:dateTime, however, the above literal is
mapped to a number representing the number of seconds from a well-defined time
point. This effectively removes any time zone information; in fact, one can
reinterpret the same value in any time zone! Clearly, it is not possibleany more
to implement the fn:timezone-from-dateTime function because the necessary
information is missing in the value space.

Effectively, most of the functions from QXuery become unimplementable if we are
to stick with our current interpretation of xsd:dateTime.


Finally, I just wanted to point out that, as we've decided at the F2F, we now
have a statement in the Syntax document that an implementation might support
other XSD datatypes, but it must do so according to the XML Schema
specification. Currently, this is a contradictory statement: it is impossible to
support xsd:dateTime if we stick to our interpretation of xsd:dateTimeStamp.

Regards,

	Boris 

> -----Original Message-----
> From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org] On
> Behalf Of Alan Ruttenberg
> Sent: 17 March 2009 04:53
> To: Antoine Zimmermann
> Cc: Boris Motik; W3C OWL Working Group
> Subject: Re: A Last Call comment about the seamtnics of xsd:dateTimeStamp in
> OWL 2 (caused by the yesterday's decision about numerics)
> 
> On Fri, Mar 13, 2009 at 8:13 AM, Antoine Zimmermann
> <antoine.zimmermann@deri.org> wrote:
> > Boris Motik a écrit :
> >>
> >>  [...]
> >> Note that this is *exactly* the same problem as the one we have with
> >> xsd:decimal
> >> and xsd:double; hence, I consider it really strange to use one solution
> >> for
> >> numerics but a completely different one for dates.
> >
> > I agree. And for consistency, it would be reasonable to adopt this change,
> > IMHO.
> >
> >> [...]
> >> - Nobody (such as RIF) can scorn us for going our way: we can always point
> >> to
> >> XML Schema and say "Here is the holy bible!"
> >
> > The Bible is all about interpretation ;-)
> 
> Hello Antoine.
> I'd consider something of a failure if anything in our normative
> specification is subject to interpretation. Would you not agree that
> the bible is a rather bad example to follow if one is writing a
> specification?
> -Alan
> 
> >
> >
> > Regards,
> > --
> > Antoine Zimmermann
> > Postdoctoral researcher at:
> > Digital Enterprise Research Institute
> > National University of Ireland, Galway
> > IDA Business Park
> > Lower Dangan
> > Galway, Ireland
> > antoine.zimmermann@deri.org
> > http://zimmer.aprilfoolsreview.com/
> >
> >

Received on Tuesday, 17 March 2009 14:59:39 UTC