Re: Section 4: LDPR/non-LDPR formal definitions

On 22 Mar 2013, at 12:59, Martynas Jusevičius <martynas@graphity.org> wrote:
> Let me rephrase -- is the following true
> in the LDP ontology?
> 
>  ldp:Resource rdfs:subClassOf foaf:Document

Personally I think it is. However, that implies consent with the TAG's httpRange-14 decision, which is, shall we say, controversial. I think it would be ideal to define LDP in a way that is agnostic with regard to httpRange-14.

I'm also unsure whether we want to explicitly assert relationships with FOAF, given that LDP as W3C product has a good chance of outliving FOAF.

>> I cannot think of any use case for the class of non-LDPRs. You seem to think that there is a use case for it, that's why I asked you to state the use case.
> 
> My use case would be to extend LDP ontology and/or create instances of
> its classes.

That's not a use case.

Why would you want to extend the LDP ontology in a way that requires a non-LDPR class? And why would you want to create instances that require a non-LDPR class?

>>> There is an ldp: namespace
>>> and seems like there is an ontology also, why shouldn't it include a
>>> definition of ldp:Resource alongside of ldp:Container, ldp:Page etc.?
>> 
>> Of course, but why should it include ldp:NonLDPR?
>> 
>>> And if LDPRs and non-LDPRs are disjoint, why not make it explicit as
>>> well?
>> 
>> Of course LDPRs and non-LDPRs are disjoint. Again, that's trivially true, it's a tautology. X and non-X are disjoint regardless of what X is. I don't see the point in explicitly stating trivially true facts.
> 
> Trivial for humans, but not for machines?

A machine has no trouble whatsoever figuring out that the classes X and (not X) are disjoint.

Stating tautologies to machines is even less pointless than stating them to humans.

> Isn't that the purpose of ontologies, to explicitly define facts? Even if they're trivial.

The purpose of every ontology is different. Some ontologies have no purpose at all.

You still haven't given any reason why the class of "things that are not LDPRs" would be useful. It seems to me that you're trying to evade the question by stating generalities. Can you give a concrete use case for such a class that helps in delivering some concrete benefit to some class of end-users?

Best,
Richard

Received on Friday, 22 March 2013 13:20:21 UTC