Just a quick follow-up of a non-native english speaker, not to break but approve the consensus.

When a multilingual vocabulary has an original language, it should not be hidden but clearly assumed, to make clear that other language annotations are translations and only translations. Original suggestion of Alistair would make sense in a genuine multilingual vocabulary, with completely opaque URIs (no NL fragment identifiers), built by a multilingual community bringing about concepts issued of, and expressed in, various NL, and building consensus about their common semantics. Great idea, but I'm not sure such a thing exists, or has ever, or will ever. Maybe currently GEMET is the closest I see to such a dream, since it was really built at the European level out of hundreds of different, multilingual sources.

BTW, pointing the original language of a Concept or ConceptScheme ... How would you declare that in RDF? or SKOS?

Bernard


Bernard Vatant

Knowledge Engineering

Mondeca
3, cité Nollez 75018 Paris France

Tel. +33 (0) 871 488 459 

Mail: bernard.vatant@mondeca.com

Web: www.mondeca.com

Blog : universimmedia.blogspot.com



Dan Brickley a écrit :
* Alistair Miles <a.j.miles@rl.ac.uk> [2006-02-08 16:20+0000]
  
Hi Dan,

On second thoughts I agree, let's not remove english annotations from 
the main RDF description.

A link to the english annotations only is already in place:

http://www.w3.org/2004/02/skos/core_en
    

Consensus-tastic :)

(would be interested in Jeremy and other's take on this though...)

Dan


  
Cheers,

Al.

Dan Brickley wrote:
    
* Alistair Miles <a.j.miles@rl.ac.uk> [2006-02-03 14:45+0000]
      
Hi all,

Following one of Jeremy's suggestions at [1], I'd like to propose we 
factor all English annotations out of the main RDF description of the 
SKOS Core Vocabulary and into a separate resource, as is currently the 
case for all annotations in other languages. Jeremy's reasons:

  - yes english is the default language in W3C
        
<flamebait> And the world... </flamebait>

(Especially the technology world)

      
  - but also yes the english labels should be accessible using the
same mechanisms as any other supported language. This will allow tools
to not have to special case for english.
        
I 100% agree that the English labels should be accessible by a 
mechanism identical to the other language. But for the time being,
I suggest it would be counter productive to hide the English text 
      
>from tools. I don't know of any RDF or OWL tools that will go chasing
    
around rdfs:seeAlso links (sadly) when reading a vocabulary description.
I wish they did, ... but they way to achieve that imho is by patches
to opensource tools like Protege, rather than by removing triples and 
hoping that folks notice and write the code to go find where the 
triples are now hiding.

      
This change would mean removing all statements matching the triple 
patterns:

- (?x rdfs:label ?y)
- (?x rdfs:comment ?y)
- (?x skos:definition ?y)

... from the main RDF description of the SKOS Core Vocabulary, and into 
a resource named:

http://www.w3.org/2004/02/skos/core_en
        
+1 on adding the triples to core_en

-1 on removing them from the main description

      
This change would also mean adding the following triple to the main RDF 
description of the SKOS Core Vocabulary:

{
 <http://www.w3.org/2004/02/skos/core> rdfs:seeAlso 
<http://www.w3.org/2004/02/skos/core_en>.
}
        
+1 on the rdfs:seeAlso

      
Any objections to raising this proposal?
        
Yup sorry. If this new idiom / deployment style is going to get 
traction, it would need to be adopted by a few major vocabs. I don't 
think going it alone 1st with SKOS is of any great value, and will only
cause annoyance amongst puzzled users.

Here's another argument: the English version of the SKOS definitions 
really *is* privileged, because it is the primary version agreed on by the 
community, and the others are (perhaps lossily, fallibly) derrive from
it. Ideally this could be represented explicitly in RDF, and the 
English language text be managed as you suggest. But for now, nobody 
works that way. 

A vocab created and documented primarily in Japanese might make a
similar choice, but privilege the Japanese translations. I don't mean to 
suggest that the 'default' text should always be English, and I'm 
always delighted to find schemas documented in other languages...

Dan

      
Cheers,

Al.

[1] http://lists.w3.org/Archives/Public/public-swbp-wg/2005Nov/0082.html
-- 
Alistair Miles
Research Associate
CCLRC - Rutherford Appleton Laboratory
Building R1 Room 1.60
Fermi Avenue
Chilton
Didcot
Oxfordshire OX11 0QX
United Kingdom
Email: a.j.miles@rl.ac.uk
Tel: +44 (0)1235 445440
        
-- 
Alistair Miles
Research Associate
CCLRC - Rutherford Appleton Laboratory
Building R1 Room 1.60
Fermi Avenue
Chilton
Didcot
Oxfordshire OX11 0QX
United Kingdom
Email: a.j.miles@rl.ac.uk
Tel: +44 (0)1235 445440