subset and superset use case

[Alasdair Gray]

Keith, Richard, Michael, Jun, hello.

I know you are that there is a discrepancy in the usage of void:subset W3C note and its definition. 

I believe that there should be a inverse property provided so that if you are browsing information about the subset you can jump to the superset. Thus, I think there should be a superset predicate that is the inverse of the current subset predicate. I believe that the misuse in the document of the existing subset predicate and this browsing use case support this need.

I also wonder if it would aid users, as I have encountered several folk misunderstanding the predicate, if it would be better use hasSubset and hasSuperset (or isSubsetOf). This means the directionality of the predicate is captured in its name.

Best regards,

Alasdair

[Richard Cyganiak]

Hi Alasdair,

Quite frankly, I'm not convinced by your argument.

On 25 Mar 2013, at 09:58, Alasdair Gray <alasda...@gmail.com> wrote:
> I know you are that there is a discrepancy in the usage of void:subset W3C note and its definition.

Yes. It's an error in the document.

> I believe that there should be a inverse property provided so that if you are browsing information about the subset you can jump to the superset.

I don't follow here.

The “direction” of the property (what's subject and what's object) and the navigability (in which direction you can “jump” from the one to the other) are two separate issues.

In the case of void:subset, to ensure navigability from subset to superset, the only question is whether your toolkit includes that triple into the description of the subset.

This is a toolset issue, not a vocabulary issue.

> Thus, I think there should be a superset predicate that is the inverse of the current subset predicate. I believe that the misuse in the document of the existing subset predicate and this browsing use case support this need.

Consider rdfs:subClassOf. There is no rdfs:superClassOf. It seems to work out fine.

In general, some vocabularies define inverses for everything, some don't. I don't think there's universal agreement on this. Personally, I'm concerned that defining inverses for everything makes data processing and querying harder because you always have to account for both options. (It's less of an issue if your toolchain includes a reasoner, but most toolchains don't.)

> I also wonder if it would aid users, as I have encountered several folk misunderstanding the predicate, if it would be better use hasSubset and hasSuperset (or isSubsetOf). This means the directionality of the predicate is captured in its name.

There's a high cost to renaming deployed predicates.

By convention, a predicate "xxx" means "has xxx". See rdf:type, foaf:name, dc:creator and so on.

Best,
Richard


>
> Best regards,
>
> Alasdair
>
> --
> You received this message because you are subscribed to the Google Groups "void-discussion" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to void-discussi...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>