[open-bibliography] New BNB sample data available

Thu Feb 3 21:38:56 UTC 2011

Hi Andrew,

>> From: public-lld-request at w3.org [mailto:public-lld-request at w3.org] On
>> Behalf Of Antoine Isaac
>> Sent: Thursday, February 03, 2011 13:54
>> To: open-bibliography at lists.okfn.org; public-lld
>> Subject: Re: New BNB sample data available
>>
>>       <dcterms:subject>
>>         <rdf:Description>
>>           <skos:inScheme
>> rdf:resource="http://id.loc.gov/authorities#conceptScheme" />
>>           <skos:prefLabel>Görner, Rüdiger--Travel--England--
>> London.</skos:prefLabel>
>>           <rdf:type
>> rdf:resource="http://www.w3.org/2004/02/skos/core#Concept" />
>>         </rdf:Description>
>>       </dcterms:subject>
>>
>>       <dcterms:subject>
>>         <rdf:Description
>> rdf:about="http://id.loc.gov/authorities/sh2008107012#concept">
>>           <skos:inScheme
>> rdf:resource="http://id.loc.gov/authorities#conceptScheme" />
>>           <skos:prefLabel>Literary landmarks--England--
>> London.</skos:prefLabel>
>>           <rdf:type
>> rdf:resource="http://www.w3.org/2004/02/skos/core#Concept" />
>>         </rdf:Description>
>>       </dcterms:subject>
>>
>> So I understand why you define "on-the-fly" (and "in-the-data") the
>> concepts that you can't find in the LCSH linked data. And I think this
>> is a reasonable solution.
>
> There is an issue with the way the RDF is specified IMHO and you have to
> read the SKOS specification to understand the implications of the above
> RDF:
>
>    <http://www.w3.org/TR/skos-reference/#L1567>
>
> S13 skos:prefLabel, skos:altLabel and skos:hiddenLabel are pairwise disjoint properties.
> S14 A resource has no more than one value of skos:prefLabel per language tag.
>
> If LC declares:
>
> @BASE<http://id.loc.gov/authorities/>
> <sh2008107012#concept>  skos:prefLabel "Literary landmarks--England--London."
>
> and the above RDF declares the same, then the integrity constraint S13 and S14 are
> violated because there now exists two triples in the combined graph of resources
> that say the same thing, thus S13 is violated because of the pairwise disjoint
> constraint and S14 is violated because there is more than one skos:prefLabel per
> language.
>
> If you want to do something like this then IMHO use rdfs:label instead of
> skos:prefLabel to get around integrity constraints S13 and S14.
>

A quite note, related to my comment on the data served by BL being potentially different from the one at id.loc.gov. The data above has:
<skos:prefLabel>Literary landmarks--England--London.</skos:prefLabel>
And id.loc.gov has
<skos:prefLabel xml:lang="en">Literary landmarks--England--London</skos:prefLabel>
Because of BL not committing to a specific literal language tag, S14 is in fact not violated: interesting side effect ;-)
But I agree, the best would be that BL publishes exactly the same prefLabel as id.loc.gov. That would ensure that there's never any issue!

As I said, I think that the "option" to serve concept data like this is reasonable only when these concepts are not already present on id.loc.gov. Otherwise it could raise quite many problems...

Best,

Antoine