RE: xmlsch-02 whitespace facet discussion

> >
> > No it doesn't entail at all.
> >
> > " 3 " is not an integer but "3" is. So the second graph is
> > not a consequence
> > of the first, because the whitespace makes a difference.
> >
> > (I frankly haven't a clue what the second graph means - I
> > promised Brian I
> > would read the semantics editors draft before the telecon, so
> > I might know
> > by then).
>
> Hmmm....
>
> This is an interesting entailment.
>
> I think the key question is whether XML Schema whitespace processing
> is part of the XML Schema specific interpretation of typed literals
> having XSD datatypes.
>
> You are saying that it is not. Is there anything from the XML Schema
> specs or the XML Schema WG that support that?
>
> IMO, it's really for the XML Schema folks to say whether "  3  "
> is or is not an acceptable lexical representation when interpreted
> in terms of the datatype xsd:int.
>
> It might be.
>
> My presumption was that it would be.
>
> If there is some clear evidence showing that my presumption is
> incorrect, I'd like to know it.


We might not be ready to close this one.

The XML Schema rec is clear (sort of like the RDF working drafts :) that
whitespace normalization is performed in part 1, but precisely which
processing is specified in part 2.

i.e. the processing chain is:

input-text => entity substitution => whitespace normalization => [[Lexical
Form]] => lexical-to-value mapping => value


If you look at
http://www.w3.org/TR/xmlschema-2/#decimal

for example, the defn of lexical space does not include the leading and
trailing blanks.

Whitespace normalization is specified in
http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#section-White-Space-Norm
alization-during-Validation

Jeremy

Received on Friday, 9 May 2003 09:22:14 UTC