Re: ISSUE-2 (olyerickson): dct:language should be added to DCAT [Best Practices for Publishing Linked Data]

Richard, all,

I also find the use of URIs more agreeable than plain literals.

Notice that lingvoj.org recommend using the lexvo.org vocabulary over
their own. lexvo.org have published URIs derived from ISO-639 codes,
eg http://www.lexvo.org/page/iso639-3/ell for Modern Greek. These are
instances of http://lexvo.org/ontology#Language which is, conveniently
enough, a subClassOf http://purl.org/dc/terms/LinguisticSystem [1].

Stasinos

[1] http://www.lexvo.org/ontology


On 8 December 2011 18:32, Richard Cyganiak <richard@cyganiak.de> wrote:
> On 8 Dec 2011, at 14:01, Maali, Fadi wrote:
>> Currently, I suggest using language codes as literals similar to what RDF uses for tagging plain literals (http://www.ietf.org/rfc/rfc3066.txt)
>
> Note that RFC 3066 is obsolete and is being superseded by RFC 4646 (a.k.a. BCP 47):
>
>   http://www.ietf.org/rfc/rfc4646.txt
>
> RDF 1.1 will reference that spec for the normative definition of language tags.
>
> Conveniently, the definition of dcterms:language also says:
>
> [[
> Recommended best practice is to use a controlled vocabulary such as RFC 4646 [RFC4646].
> ]]
> http://dublincore.org/documents/dcmi-terms/#terms-language
>
> On the other hand, the range of dcterms:language is defined as dcterms:LinguisticSystem, which seems to contradict the quoted statement above, and seems to discourage the use of a simple string as a value of this property.
>
> I note that the old dc: namespace (/elements/1.1/) doesn't contain that range declaration.
>
> There are already datasets that define URIs for languages, like http://www.lingvoj.org/lingvo/en-us so recommending the use of a URI would be another option. (In the case of Lingvoj, these URIs are again based on RFC 4646.)
>
> Best,
> Richard
>
>
>
>>
>> Details at:  http://www.w3.org/2011/gld/track/issues/2
>>
>> Regards,
>> Fadi Maali
>>
>>> -----Original Message-----
>>> From: Government Linked Data Working Group Issue Tracker
>>> [mailto:sysbot+tracker@w3.org]
>>> Sent: 23 November 2011 02:14
>>> To: public-gld-wg@w3.org
>>> Subject: ISSUE-2 (olyerickson): dct:language should be added to DCAT
>>> [Best Practices for Publishing Linked Data]
>>>
>>>
>>> ISSUE-2 (olyerickson): dct:language should be added to DCAT  [Best
>>> Practices for Publishing Linked Data]
>>>
>>> http://www.w3.org/2011/gld/track/issues/2
>>>
>>> Raised by: John Erickson
>>> On product: Best Practices for Publishing Linked Data
>>>
>>> (Added as an issue at the suggestion of @cygri)
>>>
>>> dct:language should be added to DCAT <http://t.co/loHgJejt>
>>>
>>> Discussion: We (TWC RPI) are refining our IOGDS <http://t.co/7HPOldN>
>>> and will be adding dct:language the Catalog level, first for statistics
>>> purposes and then as a browsable facet. dct:language is in our metadata
>>> model, as a conversion "enhancement" that is added after scraping and
>>> first-pass RDF conversion. The person doing that work has noted the
>>> predominant languages (approx 16 across 116 catalogs from over 36
>>> countries and international organizations) but we don't have it "in the
>>> graph" (yet)...
>>>
>>>
>>
>
>

Received on Thursday, 8 December 2011 20:37:19 UTC