W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2002

Literals: language and xml (was: Comments on new datatyping document, part 1)

From: Graham Klyne <GK@NineByNine.org>
Date: Tue, 10 Sep 2002 12:07:23 +0100
Message-Id: <>
To: "Patrick Stickler" <patrick.stickler@nokia.com>
Cc: "RDF core WG" <w3c-rdfcore-wg@w3.org>

Well, I think this is fine if that's what the WG agrees on, though it seems 
a little counter-intuitive to me, for two reasons:

- it means that the internal structure of literal forms is not quite opaque 
to datatyping:  the "xml" flag has an effect.

- the language tag and xml flag are treated differently for the purposes of 

As I said, I don't have a deep problem with either of these even if I do 
have a mild dislike.  But I guess we should be clear about what we're deciding.


At 12:21 PM 9/10/02 +0300, Patrick Stickler wrote:
> > For example consider whether:
> >    < <xsd:integer>"25"   , 25 >
> >    < <xsd:integer>"25"-en, 25 >
> > are distinct members of a datatype mapping.  Similarly, are the following
> > distinct?
> >    < <xsd:integer>"25"   , 25 >
> >    < <xsd:integer>xml"25", 25 >
>The XML flag and xml:lang code do not participate in any way
>with datatyping semantics. They are invisible/ignored/discarded/whatever
>when considering the L2V mapping. Only the unicode string portion is
>relevant, and it is taken, alone, to represent a lexical form, a
>member of the lexical space of the datatype.
>Also, Part 1 does not define any participation of XML literals in
>datatyping, only non-XML literals.
>Thus, all of the following typed literal nodes denote the very
>same value (ten):
>    <http://...#integer>"10"
>    <http://...#integer>"10"-en
>    <http://...#integer>"10"-fi
>    <http://...#integer>"10"-sp
>    <http://...#integer>"10"-en_UK
>And the following are disallowed
>    <http://...#datatype>xml"LLL"
>    <http://...#datatype>xml"LLL"-xx
>    <http://...#datatype>xml"LLL"-xx_XX
>XML literals are not datatyped (at least as far as Part 1 is
>concerned. As an aside, I think they *could* be datatyped,
>with complex datatypes, but that remains in Part 2 and is not
>part of the recent concensus.

Graham Klyne
Received on Tuesday, 10 September 2002 06:51:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:15 UTC