> dateTime allows arbitrary sub-second precision and arbitrary
> timezones
> expressed as hh:mm offsets from UTC.
>
>
> I suggest subclassing xsd:dateTime to: disallow fractional
> seconds and
> timezones other than UTC.
>
Rich, I think we just cross-posted on this, so let me ask you this:
Since dateTime does uniquely identify a time what is the value of
requiring dateTime to be expressed in strict UTC?
It seems to me that the only benefit here is that SAML-processors need a
little less code (i.e. the code that transforms a local+UTC offset into
UTC).
That argument strikes me as pretty soft, since it seems to me that the
majority of systems won't be coding from a scratch--they'll be built on
top of some level of Schema-aware libraries that should handle this.
I do agree with Stephen that we need to make some explicit statement
about fractional seconds, but I don't think we need to disallow them.
Adding a note to the effect that all unreported sub-second precision
digits in a given time are zero, with the standard time comparison rules
in effect, should handle any ambiguities without removing any potential
utility in fractional second precision.
C.
