W3C home > Mailing lists > Public > public-swd-wg@w3.org > May 2008

RE: [SKOS] SKOS-XL & label relations

From: Alistair Miles <alistair.miles@zoo.ox.ac.uk>
Date: Fri, 2 May 2008 11:04:46 +0100
To: "'Reul, Q. H.'" <q.reul@abdn.ac.uk>, "'SWD WG'" <public-swd-wg@w3.org>
Message-ID: <000d01c8ac3b$f40d7630$dc286290$@miles@zoo.ox.ac.uk>

Hi Quentin,

Yes, in a sense there is duplication. But XL gives you a convenient way to
make statements about the lexical entities used in a knowledge organisation
system, which vanilla SKOS doesn't. It also gives you a way to "dumb-down"
to vanilla SKOS, to get the simplified representation.

E.g. using SKOS+XL ...

<C1> rdf:type skos:Concept;
  xl:prefLabel <L1>;
  xl:altLabel <L2>;
  skos:inScheme <S>.

<L1> rdf:type xl:Label;
  xl:literalForm "animals"@en;
  skos:inScheme <S>;
  dc:source "Another thesaurus."@en. 

<L2> rdf:type xl:Label;
  xl:literalForm "fauna"@en;
  skos:inScheme <S>;
  dc:source "Yet another thesaurus."@en. 

... which can be "dumbed-down" via the XL data model to vanilla SKOS as ...

<C1> rdf:type skos:Concept;
  skos:prefLabel "animals"@en;
  skos:altLabel "fauna"@en;
  skos:inScheme <S>.

... does that make sense? 

I suppose you could achieve the same goal without xl:prefLabel, xl:altLabel
and xl:hiddenLabel, e.g. ...

<C1> rdf:type skos:Concept;
  xl:prefLabel "animals"@en;
  xl:altLabel "fauna"@en;
  skos:inScheme <S>.

<L1> rdf:type xl:Label;
  xl:literalForm "animals"@en;
  skos:inScheme <S>;
  dc:source "Another thesaurus."@en. 

<L2> rdf:type xl:Label;
  xl:literalForm "fauna"@en;
  skos:inScheme <S>;
  dc:source "Yet another thesaurus."@en. 

... however the connection here between <C1> and <L1> is now much weaker,
via the literal "animals"@en. In fact, because two different instances of
xl:Label can have the same literal form, there is no definite connection
between <C1> and <L1> at all.

Cheers,

Alistair.

--
Alistair Miles
Senior Computing Officer
Image Bioinformatics Research Group
Department of Zoology
The Tinbergen Building
University of Oxford
South Parks Road
Oxford
OX1 3PS
United Kingdom
Web: http://purl.org/net/aliman
Email: alistair.miles@zoo.ox.ac.uk
Tel: +44 (0)1865 281993


> -----Original Message-----
> From: Reul, Q. H. [mailto:q.reul@abdn.ac.uk]
> Sent: 15 April 2008 09:58
> To: Alistair Miles; SWD WG
> Subject: RE: [SKOS] SKOS-XL & label relations
> 
> Hi Alistair,
> 
> I had a look @ both the SKOS Reference [1] and the SKOS XL document
> [2].
> I was wondering why we needed to create xl:prefLabel, xl:altLabel and
> xl:hiddenLabel?
> 
> I see that the range for these is xl:Label which uses xl:literalForm to
> give the literal form. However, the range of xl:literalForm is the
> class
> of RDF plain literals, which is exactly the same as the range of
> skos:prefLabel, skos:altLabel and skos:hiddenLabel. Is this not
> duplicating properties?
> 
> Cheers,
> 
> Quentin
> 
> [1] http://www.w3.org/TR/skos-reference/
> [2] http://www.w3.org/2006/07/SWD/SKOS/xl/20080414
> 
> 
> -----Original Message-----
> From: public-swd-wg-request@w3.org [mailto:public-swd-wg-
> request@w3.org]
> On Behalf Of Alistair Miles
> Sent: 14 April 2008 17:14
> To: 'SWD WG'
> Subject: [SKOS] SKOS-XL & label relations
> 
> 
> Hi all,
> 
> Considering Antoine's comments at [1], it seems there are three
> alternate
> patterns for representing relations between lexical entities. Here, I
> call
> these three patterns "n-ary literal relations", "binary (XL) label
> relations" and "n-ary (XL) label relations". Currently, the SKOS
> Reference
> provides support for the first, my original "XL" sketch [2] supported
> the
> third, and nothing supported the second.
> 
> Given the amount of implementation experience, it is difficult to make
> any
> judgments about which of these three patterns is "best". Most likely,
> each
> pattern will be suited to different situations, and there may be a need
> for
> all three. I had a chat with Sean about this this morning, and he felt
> the
> same way.
> 
> So I can see two options open to the WG.
> 
> OPTION 1 - Move all support for relations between lexical entitites
> *out* of
> the SKOS Reference. (If time, publish a note illustrating three
> alternative
> patterns with some vocabulary.)
> 
> OPTION 2 - Leave the current SKOS Reference features in place. (If
> time,
> publish a note illustrating the two remaining patterns with some
> vocabulary.)
> 
> I had originally favoured option 2, however I am beginning to see that
> to
> favour any one pattern by placing it in the SKOS Reference does not
> accurately reflect the state of standardisation and consensus.
> 
> If we chose option 1, we could then consider publishing a note on the
> three
> design patterns. For illustration, I've sketched the outline of such a
> note
> at:
> 
> [3] <http://www.w3.org/2006/07/SWD/SKOS/xl/20080414>
> 
> Best wishes,
> 
> Alistair.
> 
> [1]
> <http://lists.w3.org/Archives/Public/public-swd-wg/2008Apr/0041.html>
> [2] <http://www.w3.org/2006/07/SWD/wiki/SkosDesign/SKOS-XL>
> 
> --
> Alistair Miles
> Senior Computing Officer
> Image Bioinformatics Research Group
> Department of Zoology
> The Tinbergen Building
> University of Oxford
> South Parks Road
> Oxford
> OX1 3PS
> United Kingdom
> Web: http://purl.org/net/aliman
> Email: alistair.miles@zoo.ox.ac.uk
> Tel: +44 (0)1865 281993
> 
> 
> 
> 
> 
> 
> The University of Aberdeen is a charity registered in Scotland, No
> SC013683.
Received on Friday, 2 May 2008 10:05:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 2 May 2008 10:05:30 GMT