W3C home > Mailing lists > Public > public-esw-thes@w3.org > September 2013

Re: Question: What is Best Practice for a Catalog of ConceptScheme mapping

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Tue, 10 Sep 2013 09:41:11 +0200
Message-ID: <522ECD17.80408@few.vu.nl>
To: <public-esw-thes@w3.org>
Dear all,

EDOAL is a very good piece of work, but be careful. Its main drawback is the focus on application that consume alignment data in an advanced way. In particular, all individual link are reified.

I.e the data does not say

C1 skos:exactMatch C2

It states instead something like

cell1
edoal:source C1  ;
edoal:target C2  ;
edoal:relation skos:exactMatch  .

In the case EDOAL targets there are not many other ways to proceed--only named graphs come to mind for representing such context info as EDOAL does. But this could be an obstacle for applications that are focused on simpler (SKOS) data.

Best,

Antoine



> Thanks for the suggestion Vladimir.
>
> As I understand it, EDOAL is an alignment specification format.
>
> It details the alignment between instances of 2 vocabularies.
>
> EDOAL is a powerful formalism at this level.
>
> The specification can be applied and the application result could be expressed in owl:sameAs or in SKOS mapping or otherwise.
>
> What I am seeking best practice on is on how to publish the alignment specification or the result of an alignment application.
>
> DCAT is not very formal about detailing the alignment characteristics:
>
> - the matched vocabularies
>
> - the alignment vocabulary (EDOAL alignemnt specification, skos:mapping properties, ...)
>
> Imagine:
>
> - some agency published a mapping between AGROVOC and EUROVOC (maybe using EDOAL) and resulting in SKOS mapping properties
>
> - another mapping details the alignment between ELI (1) on one hand and DCMI (2) on the other hand resulting in equivalent properties
>
> What would be the publishing vocabulary (my request suggested VoID) to publish this alignment work?
>
> Imagine I have 2 sets of annotated document resources,
>
> - one set is classified with AGROVOC + annotations from dublin core (2) and from bibo (3) ontology.
>
> - the another set is classified EUROVOC + annotations from the ELI specification
>
> It would be interesting to find results of alignments published in catalog.
>
> And agent could then select the available alignments depending on the alignment characteristics it needs and understands.
>
> This can be used to make relations between the documents in the two sets.
>
> Kind Regards,
>
> Johan De Smedt
>
> (1) ELI: http://new.eur-lex.europa.eu/legal-content/EN/TXT/?qid=1377769883350&uri=CELEX:52012XG1026%2801%29
>
> (2) DCMI: http://dublincore.org/documents/dcmi-terms/
>
> (3) bibo: http://bibliontology.com/specification
>
>> -----Original Message-----
>
>  > From: Vladimir Alexiev [mailto:vladimir.alexiev@ontotext.com]
>
>  > Sent: Monday, 09 September, 2013 17:07
>
>  > To: public-esw-thes@w3.org
>
>  > Subject: RE: Question: What is Best Practice for a Catalog of ConceptScheme mapping
>
>  >
>
>  > > From: Johan De Smedt [mailto:johan.de-smedt@tenforce.com]
>
>  > > Sent: Monday, September 02, 2013 2:23 PM
>
>  > > I have considered VoID (http://vocab.deri.ie/void) as one of the vocabularies allowing to express the
>
>  > relationship(s) between the mapped concept schemes and the used mapping property
>
>  >
>
>  > Hi Johan!
>
>  >
>
>  > Perhaps you should consider the matching info formats used by OAEI:
>
>  > http://alignapi.gforge.inria.fr/format.html
>
>  > http://alignapi.gforge.inria.fr/edoal.html
>
Received on Tuesday, 10 September 2013 07:41:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:46:26 UTC