Re: [SKOS] proposed resolution for Issue 4 - BroaderNarrowerSemantics

On 15 Jan 2008, at 21:00, Antoine Isaac wrote:

>
> Hi,
>
> First, just to ward off any ambiguity: I support Guus' resolution

I would also like to add my support for this resolution.

> Second: just a bit of explanation why I wanted in the post-telecon  
> discussion to defend the transitiveBroader as a subproperty and not  
> as a super-property, in spite of what is in the part-of pattern of [1]
> (I'll let you decide whether this nail is a matter of scientific  
> interest or personal pride. But I had never misunderstood the  
> pattern of [1] ;-)
>
> The point is that there is two possibilities for this transitivity  
> effect:
> 1. It is controlled by the one who publishes the data:
> If transitiveBroader as a subproperty of broader, then it amounts  
> for the one who uses it just creates a KOS that includes all the  
> broader links that can be inferred from the hierarchy. Even if the  
> consumer of the data can still retrieve the "direct" broader by  
> some procedure, this is a situation where the consumer is strongly  
> encouraged to adhere to the point of view the publisher adopts on  
> the transitivity of "his" broader statements.
>
> 2. It is controlled by the one who consumes the data:
> If transitiveBroader as a superproperty of broader, then the  
> publisher has a more neutral stance with respect to the way the  
> hierarchy will be accessed. The consumer can decide whether he  
> wants to get the transitive closure or the direct broader, by  
> querying for the corresponding statements.
>
> So the decision we are making here in favor of 1 is not neutral.  
> But I won't argue against it, as it now seems to me to have strong  
> application motivations.
> (and the publisher really motivated for option 1 can still create  
> his own transitive specialization of skos:broader...)

There is also the consideration that Alistair raised during the post- 
telecon discussion [1] of compatability with existing data. Adopting  
the transitive subproperty means that we don't have a standardised  
mechanism for retrieving transitive closure of broader. Adopting the  
transitive superproperty means that we do.

I accept that the decision we make here is not neutral, but I believe  
that the pattern suggested by Guus fits much better with the pattern  
of "assert direct, query over closure" that I described last night.

Cheers,

	Sean

--
Sean Bechhofer
School of Computer Science
University of Manchester
sean.bechhofer@manchester.ac.uk
http://www.cs.manchester.ac.uk/people/bechhofer

Received on Wednesday, 16 January 2008 09:23:23 UTC