W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > March 2003

Re: Proposal to close tex-01

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Thu, 27 Mar 2003 14:19:05 +0000
To: Jeremy Carroll <jjc@hpl.hp.com>
cc: w3c-rdfcore-wg@w3.org
Message-ID: <13569.1048774745@hoth.ilrt.bris.ac.uk>

>>>Jeremy Carroll said:
> 
> 
> Summary: duck and swerve - neither accept nor reject
> 
> Background: anxiety that normalizing language tags to lowercase is
> non-standard.

I'd prefer we did this.  We've greatly variety on the abstract RDF
syntax and after requiring XC14N with comments, this seems the last
one.  Would it be so bad to just require it?  I can't weigh up the
downsides of that since I'm not an I18N expert, I can't recall the
detail of their anxiety.

> solution: clarify that implementations should preserve case, leave abstract
> syntax in lowercase.
> 
> Old text:
> http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#section-Graph-Literal
> 
> Add new Note before the other two
> [[
> <a name="lang-implementation-note"> </a>
> IMPLEMENTATION NOTE:
> This section has described part of an abstract syntax (see
> <a href="implementation-note">below</a>), not an implementation.
> Implementations should preserve the case of language identifiers for:
> + round-tripping.
> + interaction with other software that conforms with [RFC 3066]
> Document authors are reminded that [ISO 3166] recommends,
> but does not require, that country codes be in uppercase,
> as in the RFC-3066 language identifier
> usually written as "en-US".
> Implementations may, for example, choose to implement the abstract syntax
> by using case insensitive comparison for language tags, rather than also
> storing a case normalized form.
> ]]

[aside:
I've been wondering about having excess NOTEs in our WDs.  Is it just
clarifying a point or are we leaving a normative section?  Seems that
if it is important, it should be in the main body.  Here an
implementation note, which is a good reason, but there are many
others.
]

> 
> The new ref to ISO 3166 could be normative, since it is already part of the
> normative closure of RDF Concepts, or it could be informative, to stress
> that this implementation note is informative.

It is?  Does it say that?  I was assuming it was normative.

I will have to review what all NOTEs say and consider moving them
into the body.


> Proposed new test case (in directory tex-01?)
> 
> <rdf:Description>
>   <eg:prop xml:lang="en-US">Hello world</eg:prop>
> </rdf:Description>
> 
> entails, and is entailed by
> 
> 
> <rdf:Description>
>   <eg:prop xml:lang="En-us">Hello world</eg:prop>
> </rdf:Description>

Possibly.  An RDF/XML to Ntriples test case is the only thing a
parser can test (and there are more of them than RDF inferencers) so 
would be better.  I don't think you necessarily need entails & entails-by
to get the point across.

 <rdf:Description rdf:about="http://example.org/">
   <eg:prop xml:lang="en-US">Hello world</eg:prop>
 </rdf:Description>

entails
  <http://example.org/> <http://example.org/prop> "Hello world"@en-us.


and
 <rdf:Description rdf:about="http://example.org/">
   <eg:prop xml:lang="en-us">Hello world</eg:prop>
 </rdf:Description>

entails (the same file)
  <http://example.org/> <http://example.org/prop> "Hello world"@en-us.


This would require ntriples to be modified to require languages in
lowercase which might be compatible with this proposal, since it is a
concrete version of the abstract syntax.

Dave
Received on Thursday, 27 March 2003 09:20:03 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:56:15 EDT