W3C home > Mailing lists > Public > public-esw-thes@w3.org > January 2008

[ISSUE-77] [ISSUE-48] Re: [Dbpedia-discussion] Skos subject properties are deprecated

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Thu, 24 Jan 2008 13:45:41 +0100
Message-ID: <47988875.6060301@mondeca.com>
To: Richard Cyganiak <richard@cyganiak.de>
Cc: Peter Ansell <ansell.peter@gmail.com>, dbpedia-discussion@lists.sourceforge.net, SKOS <public-esw-thes@w3.org>

Hello all

Some precisions before anyone gets carried away :-)
The latest SKOS draft Peter mentions is certainly the editor's draft at
http://www.w3.org/2006/07/SWD/SKOS/reference/20080118
This is only an editor's draft and has no official status whatsoever
The skos:subject property is mentioned as "at risk", which means its 
relevancy is questioned. It's not *deprecated* so far AFAIK, but under 
discussion.
There are two related "open issues" on this
http://www.w3.org/2006/07/SWD/track/issues/77
http://www.w3.org/2006/07/SWD/track/issues/48 
<http://www.w3.org/2006/07/SWD/track/issues/77>

I take, as Antoine (Isaac), that the later is a generalization of the 
former :
"The SKOS model should contain mechanisms to attach a given resource 
(e.g. corresponding to a document) to a concept the resource is about, 
e.g. to query for the resources described by a given concept."

I think this is obvious. Otherwise what is the point of SKOS altogether? 
The property skos:subject was (and still is) candidate to support this 
mechanism. As has been pointed in e.g., the other ongoing thread on 
dbpedia list [1], the term "subject" can appear to be too specific in 
meaning to cover all cases of linking  a resource to a concept, and 
strange in some borderline cases. But it's more a question of 
terminology than a question of need of such a generic property. In the 
referenced thread, I think the criticism should be more interpreted as a 
weird construction of Wikipedia categories (some are very weird indeed) 
than as a mistake in using skos:subject in DBpedia to represent the 
Wikipedia categorisation.

My take on this is that such a generic property is needed and should not 
be deprecated. Since a lot of people (including dbpedia folks, but not 
only) have started using skos:subject in the above quoted very generic 
sense, and I think they are OK to do so, it should be kept as is. But it 
should be put in best practices that whenever you want to specify an 
indexing property, you define a specific subproperty of skos:subject.

SKOS specification should stress and explain what the *functional* 
semantics of this property are, and are not. Simply to *retrieve 
resources* indexed on a concept. Not to infer any specific semantics on 
the indexing link. Just : "If you are interested in this concept, here 
are resources dealing about it in some way". No more, no less. If you 
want to be more specific, use a specific subproperty.

Bernard


[1] 
http://sourceforge.net/mailarchive/forum.php?thread_name=834575810801231916m4729f854lf34f47fe9af0a746%40mail.gmail.com&forum_name=dbpedia-discussion



Richard Cyganiak a écrit :
> Peter,
>
> On 24 Jan 2008, at 05:41, Peter Ansell wrote:
>   
>> I am new to this list, but in a discussion on another list we were
>> discussing the use of the skos:subject and related items, something
>> which dbpedia has invested in heavily to represent the wikipedia
>> category system.
>>
>> The latest SKOS draft has deprecated these properties.
>>     
>
> Can you give us some background on this decision? I have a hard time  
> understanding why this step was taken.
>
>   
>> What will dbpedia use instead?
>>     
>
> I don't know. Do you have any suggestions?
>
> Richard
>
>
>   

-- 

*Bernard Vatant
*Knowledge Engineering
----------------------------------------------------
*Mondeca**
*3, cité Nollez 75018 Paris France
Web:    www.mondeca.com <http://www.mondeca.com>
----------------------------------------------------
Tel:       +33 (0) 871 488 459
Mail:     bernard.vatant@mondeca.com <mailto:bernard.vatant@mondeca.com>
Blog:    Leçons de Choses <http://mondeca.wordpress.com/>
Received on Thursday, 24 January 2008 12:45:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:59 GMT