- From: Stasinos Konstantopoulos <konstant@iit.demokritos.gr>
- Date: Wed, 24 Oct 2012 06:09:04 +0300
- To: Phil Archer <phila@w3.org>, Richard Cyganiak <richard@cyganiak.de>
- Cc: Public GLD WG <public-gld-wg@w3.org>
Richard, Phil, all, On 24 October 2012 03:16, Richard Cyganiak <richard@cyganiak.de> wrote: > > On 23 Oct 2012, at 15:55, Phil Archer wrote: > >>> 22. “The identifier relationship should not be used to link to the identifier issued by the authority that conferred legal entity status on an organization.” -- But it's a subproperty of rov:registration, so by using rov:registration, one automatically uses rov:identifier! Change this to say that the more specific sub-property rov:registration is available for identifiers that confer legal entity status. >> >> One of us has this the wrong way round. rov:registration is a sub property of the more general rov:identifier. I think the text is correct, no? > > ?a rov:registration ?b implies ?a adms:identifier ?b. One cannot use rov:registration without also implicitly asserting the rov:identifier relationship. > > This is like saying: “Don't type a resource as an Agent if it actually is a Person.” You can't do it -- typing something as a Person also types it as an Agent. Phil, I think that specifying that "?a adms:identifier ?b" should not be asserted (explicitly, it has to get inferred) on top of "?a rov:registration ?b" is an unnecessary and potentially harmful condition. In many situations the flow of code and information might be such that "?a adms:identifier ?b" is asserted first and only later is it discovered that also "?a rov:registration ?b"; retracting a statement can be a major issue in many RDF stores, especially in the presense of forward-chaining reasoning. Thus, I think that the explict statements: ?a adms:identifier ?b ; rov:registration ?b . although superfluous, should be perfectly happy to co-exist. Richard, I think that "is available" is too weak a wording; the intention is to force authors to use rov:registration where it is known to hold between ?a and ?b. A situation where one always uses adms:identifier (for whatever reason) should be made non-conformant; although, as you already pointed out, it is obviously correct as per the semantics of rdfs:subPropertyOf . Since adms:identifier is a foreign term and there is the general clause that "[a conforming data interchange] does not use terms from other vocabularies instead of ones defined in this vocabulary that could reasonably be used", isn't this enough to make everbody happy? It seems to me that this clause is sufficient to make the use of rov:registration obligatory in all situations where it is applicable and the addition of adms:identifier admissible even where superfuous. My suggestion is that the wording is changed in a way that it does not regulate, but rather reiterates or reminds the conformance requirement that rov:registration must be asserted where applicable and that only asserting an adms:identifier statement with a subject-object pair that is in the extension of both properties is non-conformant. I would further point out that the wording should be such that it does not make non-conformant situations where rov:registration is inferred rather than explict. Eg., if I define a subproperty of rov:registration that is restricted to range over VAT numbers, I can only expect that an instantiation where only my subproperty is explictly asserted is conformant with ROV. Best, Stasinos
Received on Wednesday, 24 October 2012 03:09:37 UTC