Re: SKOS transitive hierarchical relations

On 1/4/12 7:45 PM, Frederick Giasson wrote:
> Hi Antoine!
>
>> One use case for broaderTransitive is the one of a data consumer willing to use the information that a concept C is broader than concept A, whatever other concept exist "between them" in the conceptual hierarchy. This is especially useful for information retrieval.
>> One use case for broader is to be able to easily access the "individual hierarchy steps" which are present in a thesaurus. For example, when you want to display a thesaurus tree, it makes a big difference to know that there is a concept B which is broader than C but narrower than A.
>>
>> Now, if a data publisher releases only broaderTransitive, this is a pity because information is lost, as explained in my other mail today [1]. I know that in many cases you could retrieve the "original" skos:broader statements from skos:broaderTransitive one. But:
>> - this would require an algorithm which is not so trivial
>> - in some cases, the assumptions on which such an algorithm would typically rely on are in fact wrong. These cases are very rare, but still...
>>
>> So considering all these things, we opted for the current structure. Believe me, we have considered the issue quite seriously in the Semantic web Deployment group... I know that it may look sub-optimal, but this is really the best we could came up with at the time.
>
> Good thanks for this clear explanation. I will revise this, and we will revise UMBEL according to this new information, and we may release a new structure soon that would "fix" that issue.
>


Great! And sorry for not having spotted that in the UMBEL stuff. It's good that there's a community that looks at such things :-)

Antoine

Received on Wednesday, 4 January 2012 18:50:57 UTC