Hi Pascal,

On 29 March 2012 11:37, Pascal Christoph <[email protected]> wrote:
> Am Wed, 28 Mar 2012 18:53:56 +0200 schrieb [email protected] :
>
>> On 28 March 2012 14:19, Pascal Christoph <[email protected]> wrote:
>>> Am Wed, 28 Mar 2012 13:19:46 +0200 schrieb [email protected] :
>>>
[snip]
>
>> Anyway, I wasn't proposing to go through all GND IDs and trying to
>> match them, just the ones that already have a link to VIAF. If <DNB
>> URI> owl:sameAs <VIAF URI> it is easy :) DNB offers a file with CC0'd
>> RDF about people at
>
> then you are on the good site because they identify only *persons*.
>
>> https://wiki.d-nb.de/display/LDS/Dokumentation+des+Linked+Data+Services+der+DNB
>> although I keep getting HTTP 500 errors trying to request those files.
>
> which one? (And also, think of comment that on the dnb-wiki).

I already sent an email to the webmaster (as it said on the error
page), but I may also add a comment.
The PND of April, part 1 doesn't work for me:
http://datendienst.d-nb.de/cgi-bin/mabit.pl?cmd=fetch&userID=testdat&pass=testdat&mabheft=PNDrdf_1.zip

>
> BTW, you may want to use our SPARQL-endpoint, we have an index of the gnd. It
> is based on an old dump, though, I think it was April 2011:
>
> curl -d 'query=
> SELECT * FROM <http://lobid.org/gnd/> WHERE {
>  <http://d-nb.info/gnd/100063462> ?p ?o .
> }
> ' http://lobid.org/sparql/
>
> (this is rather unofficial and experimental. Ping me if it breaks. )

It seems you have the data I wanted to download then :)

SELECT * FROM <http://lobid.org/gnd/> WHERE {
 ?s <http://www.w3.org/2002/07/owl#sameAs> ?o .
}
LIMIT 10
gives some promising results.

>> The example file does show that there are owl:sameAs links in the RDF.
>
>> So in lobid.org, you treat the name as an RDF resource (that has a GND
>> URI)? Interesting. No offense intended, but does that make more sense
>> than using the literal value?
>
> its all about linking, isn't it ;)
> I mean, if it *is* an RDF ressource we can lookup its literal (actually we 
> *do*
> that for subjects , see http://lobid.org/resource/HT014337912 (eng:subject =
> ger:Schlagwort)).
>  But you are right, if we cannot make a lookup it does not make sense. So we
> are in hope that the dnb will also provide rdf for these name-identifiers. It
> would have had the advantage that if there are more ways to write down a name
> ("Goethe, Gethe, Göthe ..." ) we would not have to curate that by ourself.
>  And we just telephoned with the dnb and there will be RDF for the
> not-individualised identifiers within the next month ! :)

Hmm, that does make sense :) And great news that DNB will offer more RDF!

>>  What if there are several Friedrich
>> Schillers and the names are split by the DNB into several persons?
>
> that is an important todo. Not that easy to do, though. If solved, it could
> help to create an authority file of publishers which I would love to see.
>
>> I have been thinking to do something similar when I'm ready to publish
>> my family tree as RDF, to relate names to their origins. But I am not
>> sure whether I would use foaf:name to connect name and person, as
>> foaf:name seems to be 'designed' for literals.
>
> Didn't get that. If you want to connect name and person that would be what you
> want: "name" is a literal and "person" the individualised URI of the person, 
> e.g.:
>
> <http://d-nb.info/gnd/135539897>
> <http://d-nb.info/gnd/preferredNameForThePerson> "Elgin, Catherine Z." .
>
I want to connect names to names ("this name comes from that name")
and names to persons, so I'll have to treat names as resources too.
But that's for another discussion/mailing list.

> -o

Ben
> _______________________________________________
> Ol-tech mailing list
> [email protected]
> http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
> To unsubscribe from this mailing list, send email to 
> [email protected]
_______________________________________________
Ol-tech mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to