- From: Andy Seaborne <andy.seaborne@epimorphics.com>
- Date: Tue, 11 Jan 2011 16:51:08 +0000
- To: Axel Polleres <axel.polleres@deri.org>
- CC: SPARQL Working Group <public-rdf-dawg@w3.org>
On 11/01/11 16:45, Axel Polleres wrote: > > On 11 Jan 2011, at 16:37, Andy Seaborne wrote: > >> (this is out of date - the current draft has more work on this area) >> >> On 11/01/11 14:45, Axel Polleres wrote: >>> Looked briefly over the numeric, date, etc functions, looks great! >>> >>> Just a small question for the moment regarding: >>> >>> " >>> 17.4.5.9 tz >>> >>> simple literal timezone (xsd:dateTime arg) >>> >>> Returns the timezone part of arg as a simple literal. >>> " >>> >>> we should mention what happens when the timezone part is empty. in fn:timezone-from-dateTime >>> this returns an ewmpty sequence, what is our equivalent of returning an empty sequence? an empty sting literal? >>> or shall we return an error? that question actually may also apply to 17.4.5.8 timezone. >> >> timezone(no TZ) raises an error because we have nothing for empty >> sequence AKA nil (rdf:nil would be confusing IMHO). > > +1 to nil being a bit awkward. > >> >> tz(no TZ) returns "" > > ok wirth me in principle, but why differnet behavior from timezone? because: 1/ tz() is an accessor into the literal lexical form (it's substring, give or take the variable length seconds part), timezone is concerned with values. There is no "no value" choice. 2/ There is a sensible choice for tz(), there isn't for timezone() use case for TZ: displaying the timezone. Andy > > thanks, > Axel > >> >> >> Andy >> >
Received on Tuesday, 11 January 2011 16:52:01 UTC