RE: Timezone concept broken in XPath 2.0?

> would the system still be conformant 
> if it returned an arbitrary, but formally correct datetime 
> "current at some time during the evaluation of a query or 
> transformation in which fn:current-dateTime() is executed", 
> which the system could claim is the correct system time, and 
> on which date arithmetic could be performed correctly?

Yes. The current date/time (and the implicit timezone) are part of the
evaluation context. The spec doesn't say how values in the evaluation
context are initialized or whether they are under user control. A system
that always initialized the current date/time to 12:00:00 on 29 Feb 1848
would be perfectly conformant - just not very saleable. The same is true of
other aspects of the system context, for example an implementation could
return <gotcha/> for every call on the doc() function, regardless of the
URI.

Michael Kay