Ryan Lee wrote:
> Chris Bizer wrote:
>
>> No direct benefit for <inter:uri>, but a benefit for <inter:type> because
>> you can have several <inter:type> elements but only one type=""
>> attribute.
>> And as I had to model type that way, I also modelled URI that way.
>
>
> I'm not sure if you intended it to sound this way, but it does sound
> like your 'type' is inteded to list every single type of a resource
> (apologies if that wasn't your intent).
It sounds like that to me.
> I think it's worthwhile to draw
> a distinction between the type(s) that was (were) used to determine the
> lens and all the other types.
It might be. Could you give an example of use of this disctinction?
> So instead of, say, listing foaf:Person
> and ex:Researcher because they're both types for a resource, as below:
>
> <inter:Resource>
> <inter:type>http://xmlns.com/foaf/0.1/Person</inter:type>
> <inter:type>http://example.org/#Researcher</inter:type>
> </inter:Resource>
>
> one would get this instead:
>
> <inter:Resource>
> <inter:type>http://xmlns.com/foaf/0.1/Person</inter:type>
> <inter:Property>
> <inter:uri>http://www.w3.org/1999/02/22-rdf-syntax-ns#type</inter:uri>
> <inter:value>
> <inter:Resource>
> <inter:uri>http://example.org/#Researcher</inter:uri>
> </inter:Resource>
> </inter:value>
> </inter:Property>
> </inter:Resource>
>
> because a foaf:Person-based lens was used.
I understand this example. I like:
- the fact that you do not introduce another XML element for type
properties that were not used in the lens selector
I dislike:
- the verbosity,
- the inconsistency in presenting similar information (type properties),
- inter:type is no longer a good name for this element as it only
contains type properties that were involved in the lens selector, and
thus not all type properties of the resource (if any).
Considering this, I'm not really in favour of making this distinction
until someone provides a strong use case of such a distinction.
> There would be no inter:type if the type wasn't used in determining the
> lens.
>
> I'm not sure how useful that constraint is in the face of the more
> advanced selectors...
Yes, and this gives me more ammunitions from the
inconsistency-of-type-information-modeling-in-the-tree point of view.
Again, I am not strongly opposed to this. I'm just not yet convinced
that making this distinction is worth the trouble it brings.
--
Emmanuel Pietriga
INRIA Futurs - Projet In Situ tel : +33 1 69 15 34 66
Bat 490, Université Paris-Sud fax : +33 1 69 15 65 86
91405 ORSAY Cedex FRANCE http://www.lri.fr/~pietriga
Received on Wed May 04 2005 - 06:12:55 EDT