Why is xml:lang not allowed on typed literals?

[Bernard Vatant]

Hi all

This list does not seem very active, but hopefully someone is still monitoring it and will be able to answer

In http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Graph-Literal I read

Plain literals have a lexical form and optionally a language tag as defined by [RFC-3066], normalized to lowercase.

Typed literals have a lexical form and a datatype URI being an RDF URI reference.

Between the lines I read that the language tag xml:lang is not allowed on typed literals. Actually I just tried to do this. The rationale is to define a datatype "One Sentence" which must contain a single sentence, starting with a upper-case, ending with a dot etc ... and using this datatype for a "tagLine" property - which of course has also a language.

So I tried the syntax below and proposed it to various tools

- W3C validator validates it, seems to ignore the xml:lang tag

- Protégé does the same, imports the file and ignores the xml:lang tag when saving

- SWOOP does the other way round, ignores the rdf:datatype but keeps the language tag.

My question is, just out of curiosity, what is the rationale behind not allowing xml:lang on typed literals?

Thanks for any clue

Bernard

<?xml version="1.0" encoding="UTF-8"?>
<rdf:RDF
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
    xmlns:voc="http://example.org/voc#">
   
            <rdf:Description rdf:about="http://products.example.org/widgets/EBSW">
                <rdfs:label xml:lang="en">Example Best Super Widget</rdfs:label>
                <voc:tagLine xml:lang="fr" rdf:datatype="http://example.org/datatype#OneSentence">Notre meilleur widget est le plus beau et le moins cher du monde.</voc:tagLine>
            </rdf:Description>
   
</rdf:RDF>

--
Bernard Vatant
Senior Consultant
Vocabulary & Data Engineering
Tel:       [phone]
Mail:     bernard.vatant@...
----------------------------------------------------
Mondeca
3, cité Nollez 75018 Paris France
Web:    http://www.mondeca.com
Blog:    http://mondeca.wordpress.com
----------------------------------------------------

[jan.grant]

On Thu, 15 Oct 2009, Bernard Vatant wrote:

> Hi all
>
> This list does not seem very active, but hopefully someone is still
> monitoring it and will be able to answer
>
> In
> http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Graph-Literal I
> read
>
> Plain literals have a lexical
> form<http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#dfn-lexical-form>and
> optionally a language
> tag as defined by
> [RFC-3066<http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#ref-rfc-3066>],
> normalized to lowercase.
>
> Typed literals have a lexical
> form<http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#dfn-lexical-form>and
> a datatype
> URI being an RDF URI
> reference<http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#dfn-URI-reference>
> .
>
> Between the lines I read that the language tag xml:lang is not allowed on
> typed literals. Actually I just tried to do this. The rationale is to define
> a datatype "One Sentence" which must contain a single sentence, starting
> with a upper-case, ending with a dot etc ... and using this datatype for a
> "tagLine" property - which of course has also a language.
>
> So I tried the syntax below and proposed it to various tools
>
> - W3C validator validates it, seems to ignore the xml:lang tag
>
> - Prot?g? does the same, imports the file and ignores the xml:lang tag when
> saving
>
> - SWOOP does the other way round, ignores the rdf:datatype but keeps the
> language tag.
>
> My question is, just out of curiosity, what is the rationale behind not
> allowing xml:lang on typed literals?
>
> Thanks for any clue

I believe the rationale was along the lines that if the value of a typed
literal was represented by an XML construction, the xml:lang belonged _in_
the representation, not _on_ it. That is, that if a literal's values are
represented in infoset terms, the xlm:lang belongs in the representation.


--
jan grant, ISYS, University of Bristol. http://www.bris.ac.uk/
Tel [phone]   http://ioctl.org/jan/
Usenet: The separation of content AND presentation - simultaneously.