Re: [dxwg] prof:isTransitiveProfileOf needs more convincing case and/or example (#643)

@nicholascar it conveys useful information, thanks!

But it may have one flaw:
> 
> The ISO thing that the NZ Profile profiles is actually a Web Ontology Language
> OWL) version of the original ISO addressing standard 
sounds really loose. Can "the ISO thing" be something like "ISO19160-1"? 
And here's where the flaw could be: in this case, is `<http://linked.data.gov.au/def/iso19160-1-address>` really a profile of ISO19160-1:2015? I think that one could see in it a resource descriptor (in OWL) for that base ISO standard, not a fully-fledged profile.

Also, in terms of formulation, my expectation would be that an example for a transitive property should not hint that data publishers have to materialize the statements of that property. In SKOS we were rather clear that the general expectation is that data publishers assert `skos:broader` statements and that `skos:broaderTransitive` are then infered, possibly by others, mainly for data consumption scenarios (i.e., querying). 
Of course this is not a general rule. Still, the intro and formulation for your example could be changed.

Here's what I would suggest:

1. The example without the transitive statements.

2. A sentence that introduces the inference/semantics. Like "From the `prof:isProfileOf` statements in this example, and according to the semantics of `prof:isTransitiveProfileOf`, one can infer the following statements:"

3. The transitive statements infered.

4. A recap/guidance for data consumers and implementers, leaving options open. Like 
"These statements can for example help data consumers to identify that a piece of data that conforms to a very narrow profile actually conforms to a more general specification that they can handle, independently from the complexity of the 'profiling chain' that relates these two specifications. At this stage however we do not make recommendations on who should materialize the results of inference, and where they shall be published".
*unless* there is a requirement that says that the hierarchy of profiling should be published?

I hope this helps.

-- 
GitHub Notification of comment by aisaac
Please view or discuss this issue at https://github.com/w3c/dxwg/issues/643#issuecomment-460593461 using your GitHub account

Received on Tuesday, 5 February 2019 10:50:41 UTC