W3C home > Mailing lists > Public > public-esw-thes@w3.org > January 2010

Re: SKOS primer section 3.3 and DCAM VocabularyEncodingScheme

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Wed, 27 Jan 2010 19:42:02 +0100
Message-ID: <4B6088FA.9030301@few.vu.nl>
To: "Houghton,Andrew" <houghtoa@oclc.org>
CC: public-esw-thes@w3.org
Hi Andrew,

> The SKOS primer, section 3.3 Subject Indexing and SKOS, points out that one can look to Dublin Core to dct:subject as a means to relate a resource to a SKOS concept.  I'm trying to reconcile this section with dct:subject and the DCAM notion of vocabulary encoding schemes, e.g., dcam:VocabularyEncodingScheme.  While the RDF definition for dct:subject doesn't include any formal domain or range dct:subject works in conjunction with dct:DDC, dct:MESH, etc.  One can say in DCAM:
> <dct:subject>
>    <dct:DDC>
>      <rdf:value>123.4</rdf:value>
>    </dct:DDC>
> </dct:subject>
> dct:DDC in DCAM is specified as:
>    dct:DDC a dcam:VocabularyEncodingScheme .
> and dcam:VocabularyEncodingScheme is specified as:
>    dcam:VocabularyEncodingScheme a rdfs:Class .
> Further DCAM specifies dcam:memeberOf similar to skos:inScheme where it represents a relationship between a resource and a vocabulary encoding scheme.
> I am trying to be true to the spirit of dct:subject, DCAM's dcam:VocabularyEncodingScheme and dcam:memberOf, and it seems to me that dct:subject should have a range of dcam:VocabularyEncodingScheme even though DCMI doesn't specify a range for dct:subject and dcam:memberOf should have a range dcam:VocabularyEncodingScheme even though DCMI doesn't specify a range for dcam:memberOf.
> What I would like to do is to use dct:subject with a URI to my SKOS concepts.  Due to the missing range on dct:subject I can do this without issue, but it seems like it's breaking the spirit of DCAM's vocabulary encoding schemes.  In order to reconcile this issue I'm looking at doing the following in my SKOS:
> @prefix dcam:<http://purl.org/dc/dcam/>  .
> @prefix dct:<http://purl.org/dc/terms/>  .
> @prefix foaf:<http://xmlns.com/foaf/0.1/>
> @prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#>  .
> @prefix rdfs:<http://www.w3.org/2000/01/rdf-schema#>
> @prefix skos:<http://www.w3.org/2004/02/skos/core#>  .
> @base<http://dewey.info/>  .
> <scheme/ddc22>  a skos:ConceptScheme ;
>    rdfs:subClassOf dcam:VocabularyEncodingScheme .
> <class/123.4>  a skos:Concept ;
>    skos:inScheme<scheme/ddc22>  ;
>    dcam:memberOf<scheme/ddc22>  .
> @base<http://example.org/>  .
> <resource/1>  a foaf:Document ;
>    dct:subject<class/123.4>  .
> Basically, a skos:ConceptScheme becomes a subclass of dcam:VocabularyEncodingScheme and each skos:Concept becomes a dcam:memberOf the skos:ConceptScheme.  Now I don't feel like I'm breaking the spirit of DCAM or SKOS, but allowing them to work together.
> I would like some feedback about adding these properties to my SKOS descriptions and whether making a skos:ConceptScheme a subclass of dcam:VocabularyEncodingScheme causes any issues for SKOS.

I would say that it's perfectly fine from a formal perspective. Except maybe that:

1. it would maybe call for a more precise characterisation: is skos:ConceptScheme really a subClassOf dcam:VocabularyEncodingScheme? Isn't it the opposite? Could they be equivalent? Honnestly I don't know enough DCAM to answer, but I'd be interested in having a clearer view ;-)

2. there's something weird with your examples. According to your definition of <class/123.4>, <scheme/ddc22> is an *instance* of both skos:ConceptScheme and dcam:VocabularyEncodingScheme. Which is fine. But in the lines above you say that <scheme/ddc22> is a subclass of dcam:VocabularyEncodingScheme, which does not fit.
Similarly, your first example (in XML) seems to hint that the object of your dct:subject statement is an *instance* of dct:DDC, not a *member of* it.


Received on Wednesday, 27 January 2010 20:55:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 January 2010 20:55:40 GMT