Re: rdfms-xmllang: a proposal

Dan Connolly wrote:
> 
> Brian McBride wrote:
> >
> > The story so far:
> >
> > We have established that M&S clearly states that it consider langauge to be
> > 'part of' a literal.
> >
> > We identified 3 use cases requiring language support:
> >
> >   Martyn's
> >   Jan's
> >   OCLC's
> >
> > Of these Martyn's did not consider representation of language in a literal
> > to be important.  The other two found the M&S specification of language as
> > part of a literal to be useful and adequate for their needs.
> 
> Really? they found it useful?
> I thought the OCLC folks got their job done
> without using xml:lang.

Eric, please did I understood you correctly at the last telecon?  I 
understood you to say that the DC use case you provided did use
xml:lang as specified in M&S and that was adequate to meet the
requirements of that use case.


> 
> I guess I'll pore over the records, but I'd appreciate a pointer.

  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0013.html
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0111.html

> 
> > I propose therefore that:
> >
> >   o a literal be regarded as a pair (s,l) where s is a string of ISO10646
> >     characters and l is either null or a language identifier as
> >     defined in RFC 1766 or its successors.
> 
> That doesn't seem to cover the case of rdf:parsetype="Literal".

True - we've got that as a separate issue but really these should be
considered together.

> 
> >   o that an item be included in the errata for M&S:
> >
> >  http://lists.w3.org/Archives/Public/www-archive/2001Jun/att-0021/00-part#33
> 
> Which text is wrong?

  or it can be a literal; i.e., a resource (specified by a URI) or a simple
  string or other primitive datatype defined by XML.

This text does not make it clear, at least to me, that a literal can have
a language component as described in:

  http://lists.w3.org/Archives/Public/www-archive/2001Jun/att-0021/00-part#221

  The xml:lang attribute may be used as defined by [XML] to associate
  a language with the property value. There is no specific data model
  representation for xml:lang (i.e., it adds no triples to the data 
  model); the language of a literal is considered by RDF to be a part 
  of the literal.

> 
> >     which should make it clear that a literal is not just a simple string,
> >     but a compound structure including an optional representation of a
> >     language encoding.
> 
> it already says that a literal isn't just a string.
> 
> >     It should be noted that other parts of the text of the specification
> >     may need similar clarification.
> >
> >   o that n-triple be modified to represent the language encoding of a
> >     literal
> 
> Why? n-triples can repesent pairs already, using triples.
> I find that option much more appealing.

The original WG have taken a different view.  Is there a more powerful
case that can be made to change what M&S specifies than we have so 
far heard.  

> 
> >   o that we delay determining the wording of the errata and the change to
> >     n-triple until issue
> >
> >     http://www.w3.org/2000/03/rdf-tracking/#rdfms-literal-is-xml-structure
> >
> >     is resolved as the outcome of that issue may further refine the
> >     definition of a literal.
> 
> I'm not interested
> in adopting it separately from a solution
> to rdfms-literal-is-xml-structure. 

Fair enough.

(I'm not interested
> in adopting it at all; but this issue allows me
> to give some justification, over and above 'yuk!')

I see 'yuk', what's the justification?

Brian

Received on Monday, 23 July 2001 18:11:43 UTC