W3C home > Mailing lists > Public > public-lld@w3.org > October 2010

Re: VIAF contributor model

From: Manue <manue@figoblog.org>
Date: Fri, 29 Oct 2010 17:07:11 +0200
Message-ID: <AANLkTimVvDJoRVc-JLCkAotk93L=B3AM8qdB7POChpNa@mail.gmail.com>
To: Karen Coyle <kcoyle@kcoyle.net>
Cc: public-lld@w3.org
I see skos:concept and foaf:person like the 2 sides of the same coin.

Take an authority record that is about a person. In MARC, we do have
sufficient info to know that 1/ its an authority, and 2/ its about a
person.
So, when we're going to translate that into a RDF model, we don't want
to lose that information.
If we use only skos:concept, we loose the "person" info.
If we use only "foaf:person" we loose the fact that's its meant to be a name.

So, the proposed solution is to say that it's a name (skos:concept)
about a person (foaf:person).
We could apply both types to the same resource, but we don't like that
very much because it creates confusion. So we prefer to coin 2 URIs,
that is, split our authority data (from 1 MARC record) into 2
resources.
Foaf:focus is the property used to link the 2, by saying that the
concept actually describes the person.

Martin suggested that to us and we found it very convincing at BnF. We
can split attributes and relationships (to speak the RDA language ;-)
for 1 authority between what is relevant for a concept and what is
relevant for a person. For instance, we dont have a skos:concept with
a birth date, or a foaf:person with an alternate label. We don't have
a skos concept which is the author of a work (I really didn't like
that at all ;-). And we can link the foaf:person to DBPedia.

Emmanuelle




On Fri, Oct 29, 2010 at 4:10 PM, Karen Coyle <kcoyle@kcoyle.net> wrote:
> Quoting Manue <manue@figoblog.org>:
>
>> Jeff,
>>
>> This sounds like a very good approach.
>>
>> As we were discussing during the F2F, authorities are about names, and
>> not about the real thing, so it makes sense to use SKOS. SKOS  is
>> perfectly fit for prefLabel, altLabel, etc. so why reinvent the wheel.
>> Then, it's also nice to be able to describe the thing that is named.
>> There, foaf:person and foaf:organisation are probably useful when it
>> comes to persons and corporate bodies, which is what VIAF is about
>> until now.
>
> If authorities are about names, and those names are what we include in
> bibliographic descriptions (in libraries), where in library data would
> foaf:person be used?
>
> (Somewhat answering my own question, I think that given this explanation,
> foaf:person would be used outside of the library data environment, so
> library authority data might link to other resources that focus on the
> person rather than the name. But I can't think of a place in library data
> would make reference to foaf:Person except, perhaps, in the administrative
> fields of the authority record.)
>
> kc
>
>
>
>
>>
>> So, what's important is that there are 2 resources, 2 different
>> entities with each its URI : the authority (a SKOS concept) and the
>> RWO (a FOAF agent).
>> We could even add a 3rd one for the record if needed to track
>> provenance metadata (see [1]).
>>
>> As for using RDA and FRAD, maybe they will be best fit for our data in
>> the end, but this should not prevent us to still use SKOS and FOAF if
>> we want a real uptake of our data outside the library community.
>> So back to a subject we touched during the F2F : these domain-specific
>> vocabularies should find a way to link themselves to the more general
>> ones, either by declaring equivalent properties or sub classes or
>> whatever. Otherwise, it will be very difficult for implementers (like
>> VIAF) to know how to articulate them.
>>
>> One of the added values of RDF, in my view, is that
>> ontology/vocabulary producers can provide insight on mapping their
>> entities to other standards, when relevant, in a pretty simple way.
>> Which was not really the case with MARC or XML formats. This should
>> help people who own the data create constitent mappings.
>>
>> Emmanuelle
>>
>> [1] http://lists.w3.org/Archives/Public/public-lld/2010Aug/0021.html
>>
>> On Fri, Oct 29, 2010 at 5:36 AM, Young,Jeff (OR) <jyoung@oclc.org> wrote:
>>>
>>> Karen,
>>>
>>> I assume you're talking about URI hash fragments here. I'm not using
>>> "#skos:Concept" for persons, I'm using "#skos:Concept" for skos:Concepts
>>> and "#foaf:Person" for foaf:Persons. These are two separate entities.
>>>
>>> Even though I identified these separately in VIAF early on, the
>>> need/purpose of doing so was unclear to me until Martin showed me the
>>> foaf:focus property last weekend. What it means is that SELIBR, DNB, and
>>> other VIAF contributors can disagree on the identity of the skos:Concept
>>> (including preferred and alternate labels) while still agreeing (via
>>> VIAF algorithms and owl:sameAs) on the identity of "the thing".
>>>
>>> (I wish I had gotten my bachelor's degree in the engineering college
>>> rather than the business college. What is the term we alchemists should
>>> use when we mean "axiom"?)
>>>
>>> Jeff
>>>
>>>> -----Original Message-----
>>>> From: Karen Coyle [mailto:kcoyle@kcoyle.net]
>>>> Sent: Thursday, October 28, 2010 8:46 PM
>>>> To: Young,Jeff (OR)
>>>> Cc: public-lld
>>>> Subject: Re: VIAF contributor model
>>>>
>>>> Quoting "Young,Jeff (OR)" <jyoung@oclc.org>:
>>>>
>>>> >
>>>> > http://viaf.org/viaf/102333412/#foaf:Person
>>>> >
>>>> > http://viaf.org/viaf/102333412/#skos:Concept
>>>> >
>>>>
>>>> I don't understand why you are using #skos:Concept for Persons/Agents.
>>>> Is it because they are marked that they can also be used in subject
>>>> headings in the MARC name authorities file? Or some other reason?
>>>>
>>>> kc
>>>>
>>>> --
>>>> Karen Coyle
>>>> kcoyle@kcoyle.net http://kcoyle.net
>>>> ph: 1-510-540-7596
>>>> m: 1-510-435-8234
>>>> skype: kcoylenet
>>>>
>>>
>>>
>>>
>>>
>>
>>
>
>
>
> --
> Karen Coyle
> kcoyle@kcoyle.net http://kcoyle.net
> ph: 1-510-540-7596
> m: 1-510-435-8234
> skype: kcoylenet
>
>
Received on Friday, 29 October 2010 15:07:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 29 October 2010 15:07:56 GMT