W3C home > Mailing lists > Public > public-html-data-tf@w3.org > November 2011

Re: Draft Note for HTML WG

From: Dan Brickley <danbri@danbri.org>
Date: Mon, 14 Nov 2011 14:56:44 +0100
Message-ID: <CAFNgM+b+y=tTjVpeFOhgRW2hf4JYXK7PdqmifsFvZf3hgSNn_w@mail.gmail.com>
To: Ivan Herman <ivan@w3.org>
Cc: Jeni Tennison <jeni@jenitennison.com>, HTML Data Task Force WG <public-html-data-tf@w3.org>
On 14 November 2011 14:52, Ivan Herman <ivan@w3.org> wrote:
> On Nov 14, 2011, at 14:25 , Dan Brickley wrote:
>> On 14 November 2011 13:23, Ivan Herman <ivan@w3.org> wrote:
>>>> ... while also collecting mappings to the other widely used RDF vocabs. If RDFa parsers did
>>>> something useful with such mappings, that might help move things along too...
>>>
>>> I am not sure what you mean. What would you think an RDFa parser may do?
>>
>> I was thinking of the mechanism we discussed the other day. For
>> example, if http://schema.org/Person has an annotation saying
>> 'equivalentClass foaf:Person', then a post-processing option in a
>> parsing tool could allow that annotation to be used to expand out some
>> extra triples in the results. That said, I'm wary of something that
>> could put unpredictable load on vocab publisher's servers...
>>
>
> This is almost exactly the @vocab mechanism!

I thought vocab was just a new word used for namespace declarations,
now that XML stuff is uncool...

> Except that the minimal definition in RDFa does not rely (currently) on the owl vocabulary, only on the RDF(S) one. But if the schema.org vocabulary file has a subClass of foaf:Person, that will be used and any resource of type schema:person will also be a foaf:Person.

> equivalent class is very strong. That is a form of hijacking; after all, that would also mean that each foaf:Person is also a schema.org person. Which may be a true statement, but as a general mechanism it is heavy and dangerous, a reason why the RDFa group did not adopt that. (Another reason is that the spec can now refer to a small subset of RDFS entailment, but then it would have to refer to a small subset of OWL RL, which is heavier...)

I don't see things that way at all, sorry! sameAs is the nuclear one;
it states that there's just one thing, not too. So the hijacking
metaphor makes sense there.

As I understand equivalentClass (and equivalentProperty), you're just
saying that any world in which 'foo a Class1' is true, will also have
'foo a Class2' being true also, and so on. Cutting back to subClassOf
just means you're admitting the possibility that there are some things
that are in one class, but not in the other. Why force people to say
this if they mean otherwise?

Example: between FOAF and DC we agreed that we couldn't think of any
examples of dcterms:Agent that were not also foaf:Agent and vice
versa, ie. that our intent and wording and practice meant that they
were effectively equivalent. Why force us to say something other than
that in the markup?

Whether tools do anything complete with that is another matter. But it
seems fine to allow vocab owners to express equivalence directly (he
says, naively innocent of OWL RL details).

cheers,

Dan
Received on Monday, 14 November 2011 13:57:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 14 November 2011 13:57:16 GMT