Saašha pointed out (several messages ago in mid August) that 'string' is not yet defined.
I think it should be defined simply in terms of a character set. I don't think that we want to define it in terms of qname. I understand that the capability to namespace will be important. We can simply include the colon (":") in the character set that defines 'string' and people can develop vocabularies based on qnames or not as they see fit.
The question is, what should be that character set? (Suggestions?)
> -----Original Message-----
> From: Discussion of the Developing Date/Time Standards
> [mailto:[log in to unmask]] On Behalf Of Saašha Metsärantala
> Sent: Thursday, August 18, 2011 1:15 PM
> To: [log in to unmask]
> Subject: Re: [DATETIME] Season
> > QNames are highly controversial.
> I think we agree that we won't use xs:QName for qualifiers as of phase
> one. Maybe we could move this interesting topic to phase two and try to
> focus on what can be useful for phase one.
> My main suggestion was to exclude from EDTF qualifiers (phase one), all
> characters excluded by xs:NCName (such as space and colon).
> In other words: Could we agree on xs:NCName or a subset thereof?
> One possible, but very limited subset of xs:NCName would be qualifiers
> (or similar depending on which regex flavor you use), but we would
> probably prefer to be able to use other alphabets and characters, too.
> Maybe xs:NCName is too wide for EDTF qualifiers. In that case, the
> question is: Which of the characters allowed within xs:NCName should be
> excluded to create the datatype for EDTF qualifiers?