W3C home > Mailing lists > Public > public-media-annotation@w3.org > November 2010

(no subject)

From: Ivan Herman <ivan@w3.org>
Date: Mon, 15 Nov 2010 11:59:12 +0100
Cc: "Evain, Jean-Pierre" <evain@ebu.ch>, "public-media-annotation@w3.org" <public-media-annotation@w3.org>
Message-Id: <3819733B-6D00-4D2D-BD08-4434A4DD3F73@w3.org>
Thierry et al,

first of all, thank you for having made this step forward. Unfortunately, I still have comments and, well, critiques on what is in the editor's draft. Overall, the document looks like the RDF/OWL formalism is put into the document very much as an afterthought, without making clear connections to the core text. Which makes it still very difficult to use the specification.

Some examples may make my concern clearer. 

I am looking at section 5.1.2 which, I expect, gives me the list of properties that I can use in my RDF annotation. First of all, I am not sure what the type description means. The way I make sense out of it for a complex type, in RDF terms, is that I have to create a blank node which may have additional attributes. Ie, is it true that, say, I can/should do something like:

<...> ma:contributor [
   identifier <URI>  ; # or a string
   role "SOMESTRING" ; # this is optional
] .

using an entry in the table of 5.1.3. This may makes sense, but it is not clear from the text. For example, in what namespace is 'identifier' and 'role'? Is this really the way the specification has to be translated into RDF? 

Well, I would expect having a specification for a ma:contributor in the OWL spec to look it up in the ontology. Hm. There is _no_ entry for #contributor! There is a #contributorIs though, so maybe that is the one and there is mismatch between the two parts of the text; that is fine, you guys will check that. So let us go for #contributorIs, it says its range is #Contributor. Ok, let us go there, maybe it says something about identifier and role. But it does not. And, of course it does not because an OWL ontology's owl:Restriction does not define a constraint on classes, just licenses to infer. So I would look for something like #role, but that is not there either... And then I give up and look for a different ontology for my application.

I hope you get what I say. The core text should make it very clear how that ontology is to be used by and for RDF, what the 'type definition' means in 5.1.2 and other places, how should I, mentally, translate what is in the tables for a real annotation using RDF. At the moment it does not. More examples would be really useful, b.t.w....

I also have some comments on the ontology itself

- Personally, I would prefer to have the OWL ontology in Turtle, too. Mainly for humans it is much more readable. But that is an editorial choice

- The starting comment includes 

    <song1.mp3> ma:hasGenre [ rdfs:label "Blues" ] .

which means the overall RDF/XML file invalid in XML:-(

- the datatype specifications are all in

  <rdfs:range rdf:resource="string"/>

but that is incorrect, the resource attribute should have the full XSD URI as a value (one of the features that would keep the specification much more readable in Turtle...)

- I would probably check whether the ontology remains in a profile of OWL, maybe OWL RL. In the application domain of this ontology, close to the RDF world of annotation, it may be very useful to make that sure. If there are features that would push the ontology out of RL, I would probably think through whether that is really necessary, ie, whether the inferences that can be drawn from that part of the ontology are really important and useful in practice...

I hope this helps

Thanks!

Ivan


On Nov 10, 2010, at 10:59 , Thierry MICHEL wrote:

> Ivan,
> 
> In your LC comment you had asked to see the new version of our rdf ontology
> 
> http://lists.w3.org/Archives/Public/public-media-annotation/2010Sep/0241.html
> 
> It now included in the latest Ontology draft (going for a  2nd LC).
> http://dev.w3.org/2008/video/mediaann/mediaont-1.0/mediaont-1.0.html#ont-RDF 
> 
> Please take a look at it and feel free to send feedback.
> 
> Thierry
> 
> 
> 
>> Le 10/11/2010 09:18, Evain, Jean-Pierre a écrit :
>>> Dear all,
>>> 
>>> I am pleased to send you the new version of our rdf ontology following
>>> the meeting at TPAC Lyon.
>>> 
>>> Thank-you to Pierre Antoine, Tobias and Yannick for their active
>>> support in this work.
>>> 
>>> During the TPAC meeting, it was decided that this should become a
>>> normative part of the MAWG ontology and therefore we need to respect
>>> the deadlines. As a consequence, final comments are due before Monday
>>> next week.
>>> 
>>> All the modifications made to the ontology are listed in comments
>>> within the rdf file.
>>> 
>>> Thierry, if you would be so kind to upload this as our new version.
>>> Thank-you in advance.
>>> 
>>> Best regards,
>>> 
>>> Jean-Pierre
>>> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf







Received on Monday, 15 November 2010 10:57:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 15 November 2010 10:57:52 GMT