Re: Usage of language tags in SKOS

I just looked at it in some more detail, and it's wrong on at least 3 
points:

1) Subtags have to be 8 letters or less, but all their examples contain 
subtags longer than 9 letters.

2) Use of private use tags in an official spec by a Standards 
Organization is simply "not eating your own dogfood".

3) Of course, the notations involved aren't natural human languages at all.

Regards,    Martin.

P.S.: If somebody is or knows an AC Rep, please (tell them to) vote 
"yes, BUT" (or "no") on the involved specs. That's the fastest way to 
fix the problems.

On 2009/06/23 14:15, Felix Sasaki wrote:
> Unfortunately, they do the same in the SKOS proposed recommendation:
> http://www.w3.org/TR/2009/PR-skos-reference-20090615/#L2655
> So this seems to need some action ASAP from our side - agenda+ for the next
> call.
>
> Felix
>
> 2009/6/23 "Martin J. Dürst"<duerst@it.aoyama.ac.jp>
>
>> I very much agree with Felix on this. Something like
>>    skos:prefLabel "FRA"@en-x-notation-threeletter
>> is totally weird. Just because RDF does allow to tag strings with languages
>> doesn't mean that this should be misused for anything and everything. RDF
>> also has typing, and typing would be much more appropriate.
>>
>> [As shepherd of 4646bis, I just was telling the responsible AD (who was
>> thinking about going as far as disallowing them) that use of private use
>> tags for language tags is very infrequent. So much for that :-(]
>>
>> Regards,   Martin.
>>
>>
>>
>> On 2009/06/23 6:55, Felix Sasaki wrote:
>>
>>> Hi all,
>>>
>>> looking at
>>> http://www.w3.org/TR/2009/WD-skos-primer-20090615/#secnotations
>>> I am wondering if SKOS recommends the usage of language tags in an
>>> appropriate manner. I am not sure about this passage:
>>>
>>> [
>>> However, the management of such datatypes can be cumbersome. Further, the
>>> previous pattern is not really needed when publishers consider the
>>> notations
>>> themselves to be simple language-independent labels. In such cases, it is
>>> possible to use one SKOS labeling property, for instance skos:prefLabel,
>>> in
>>> combination with private use language tags (or subtags) as defined by RFC
>>> 4646 [RFC4646]. This pattern was first proposed for a list of coded
>>> countries [COUNTRYCODES-SKOS] from which the following example is adapted:
>>> iso3166:FR skos:prefLabel "France"@en ;
>>>    skos:prefLabel "FRA"@en-x-notation-threeletter ;
>>>    skos:prefLabel "250"@x-notation-numerical.
>>> ]
>>>
>>> First, private use language tags are recommended to be used for a
>>> "language-independent-label", which seems to be a contradiction in itself.
>>> Second, the labels are used for specifying the language of notations as
>>> part
>>> of e.g. classifications. Notations for classifications are not natural
>>> language, but rather closer to identifiers in programming languages, so it
>>> seems to me that usage of language tags is not appropriate for this
>>> purpose.
>>> Any comments?
>>>
>>> Best,
>>>
>>> Felix
>>>
>>>
>> --
>> #-# Martin J. Dürst, Professor, Aoyama Gakuin University
>> #-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp
>>
>

-- 
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp

Received on Tuesday, 23 June 2009 10:57:11 UTC