RE: skos:hasTopConcept mandatory?

 Let me also mention qSKOS [1], which checks for "omitted top concepts" 
 and "top concepts having broader concepts".
 Recently we did a survey containing the question whether controlled 
 vocabularies should have "at least one concept that constitutes an 
 'entry point' to the contained information". Our contributors noted that 
 it depends on the intended usage scenario, however most of them perceive 
 top concepts as useful in hierarchical vocabularies, as already argued 
 in this discussion. Having "entry points" in vocabularies was (strongly) 
 agreed by 17 or 28 contributors, 6 (strongly) disagreed, the rest 
 selected neutral or didn't know. Furthermore, we asked if these entry 
 points "should not have broader concepts". 12 of 28 (strongly) agreed, 7 
 (strongly) disagreed, the rest selected neutral or didn't know.
 So imho there is no rule of thumb when it comes to using top concepts. 
 But having a usage scenario for their vocabularies in mind, it would be 
 important for developers to know about (lack of) top concepts and their 
 position in a hierarchy.

 best,
 Christian

 [1] https://github.com/cmader/qSKOS

 Am Mittwoch, den 04.09.2013, 06:57 +0200 schrieb "Johan De Smedt" 
 <johan.de-smedt@tenforce.com>:
> Hi Thomas,
>
> I agree with all your remarks.
>
> Though concerning the remark below (a top concept having a broader
> concept) though, I would like to nuance.
>
> - SKOS though leaves this open as it allows to represent directed
> graphs where top concepts are just entry points.
>
>   (I do not know if there are useful applications for this.)
>
>   See: http://www.w3.org/TR/skos-reference/#L2446 [1]
>
> - For common hierarchical schemes and thesauri (standard ISO 25964)
> indeed indicate that a top concept cannot be a narrower concept of 
> any
> other concept.
>
> (Validation rules may differ per application. Uniqueness preferred
> labels per language is another such characteristic.)
>
> Kind Regards,
>
> JOHAN DE SMEDT
>
> _Chief Technology Officer_
>
> _ _
>
> mail: johan.de-smedt@tenforce.com [2]
>
> mobile: +32 477 475934
>
> FROM: Thomas Francart [mailto:thomas.francart@sparna.fr]
> SENT: Tuesday, 03 September, 2013 19:51
> TO: oreste.signore@isti.cnr.it
> CC: SKOS
> SUBJECT: Re: skos:hasTopConcept mandatory?
>
> Hello
>
> _[JDS:>] _[…cut…]
>
> Is it acceptable that a TopConcept is narrower than another
> Concept/TopConcept?
>
> Certainly not. If a top-level concept is moved down the hierarchy, 
> its
> topConceptOf/hasTopConcept properties should be deleted.
>
> [...cut…].
>
> Thomas
>
> [1] SKOS Play : http://labs.sparna.fr/skos-play/ [3]
>
> --
>
> THOMAS FRANCART - Sparna
> Consultant Indépendant
> Data, Sémantique, Contenus, Connaissances
> web : http://sparna.fr [4], blog : http://francart.fr [5]
> Tel : +33 (0)6.71.11.25.97
> Fax : +33 (0)9.58.16.17.14
>
> Skype : francartthomas
>
> Links:
> ------
> [1] http://www.w3.org/TR/skos-reference/%23L2446
> [2] mailto:johan.de-smedt@tenforce.com
> [3] http://labs.sparna.fr/skos-play/
> [4] http://sparna.fr
> [5] http://francart.fr

Received on Thursday, 5 September 2013 10:56:21 UTC