Re: Southampton Pub data as linked open data

On 29 Jul 2008, at 11:59, Richard Cyganiak wrote:

> Bijan, Knud, Bernard, thanks for the clarification.
>
> I'm indeed surprised! Subclassing rdfs:label is okay in RDFS, and  
> it is okay in OWL Full, but it is not allowed in OWL DL.

Yes.

Working to change this, fwiw.

> The RDF consumers I'm working on (RDF browsers and the Sindice  
> engine) don't care if you're in OWL DL or not, so I'm tempted to  
> argue that it doesn't matter much for RDF publishing on the Web.

rdfs:label subclassing is almost always harmless. But, I would argue,  
it's still the wrong thing for things like UI management.

It's also a weird thing to talk about "users caring about whether you  
are in OWL DL or not", at least, in this way. Presumably, application  
users care about application functionality (as is proper), so the  
question, in general, is whether OWL DL can enhance (or inhibit)  
application functionality.

If it's just badge hunting then it's pointless.

> (IME, on the open Web, trust and provenance are much larger issues  
> than inference,

And yet, presumably, what you want from subpropertying rdfs:label is  
that that subproperty is treated the way rdfs:label is :)

Now, I believe using this sort of inference in this sort of context  
is sort of pointless (I'd much rather have Fresnel (well, done  
right ;)). That it's also harmless is a reason to handle it.

> and I don't believe that the open Web will ever be OWL DL, so why  
> bother.)
[snip]

This is an exceedingly narrow notion of "open Web". There are a large  
number of OWL DL ontologies on the Web which are, indeed, open. And  
this is a good thing. Getting the NCI thesaurus into OWL opened up  
that data (and helped NCI move to different toolsets!).

And there are plenty of cases where you don't need RDFS or OWL style  
anything. Which is fine.

Cheers,
Bijan.

Received on Tuesday, 29 July 2008 11:36:46 UTC