W3C home > Mailing lists > Public > public-xg-mmsem@w3.org > February 2007

Re: Interoperability Framework and Vocabulary

From: Tobias Bürger <tobias.buerger@deri.org>
Date: Thu, 22 Feb 2007 07:37:56 +0100
Message-ID: <20070222073756.rxiy3qpvi840k8oc@mail1.novara.ie>
To: "Jeff Z. Pan" <jpan@csd.abdn.ac.uk>
Cc: Raphaël Troncy <Raphael.Troncy@cwi.nl>, public-xg-mmsem@w3.org

Dear Jeff, Raphaël and all,


Zitat von "Jeff Z. Pan" <jpan@csd.abdn.ac.uk>:

>
> Dear Raphaël and all,
>
> Raphaël Troncy wrote:
>> Dear Jeff,
>>
>>> After some thoughts, it seems to me that there is quite a connection
>>> between the interoperability framework and the vocabulary deliverable.
>>
>> Yes, there is certainly strong links :-)
>
> Good :-)

I also see the strong links because we want to achieve  
interoperability on the descriptive level. And the vocabularies are  
the tools to provide these descriptions at the moment.
>
>>
>>> The goal of the former one is to provide an integrated set of vocabulary
>>> within a simple extensible framework. In fact, an integrated set of
>>> vocabulary is
>>> the most needed component in the framework.
>>
>> Hummmm, I would not say that. From my perspective, the goal of the
> former is
>> to show how to use together different vocabularies in the context of a
>> specific use case
>
> It might not be proper to design the interoperability framework for *a
> specific use case*. In the framework, we use the bottom-up approach to
> support a set of use cases identified in the XG. As we presented to the
> SWCG in the telecon, our use cases cover the following three layers (which
> are based on our FTF2 discussions):
>
> - Content layer: such as Algorithm representation
> - Media+domain layer: such as Photo, Music and News
> - End-user layer: such as Tagging
>
> Note that although we expect the framework to hopefully cover many
> applications related to the use cases, we won't claim that the framework
> would cover all possible related applications.
>
>> even though these vocabularies have been made for different
>> purposes. So it is rather, for practical applications, what motivate the
> needs
>> to use several vocabularies and how combine them.
>
> Yes, the goals are to identify the needs ("why") and to provide a
> mechanism to integrate/combine the vocabulary ("how"). As we all see the
> "why" bit, the mentioned goal is focus on the "how" part. In particular,
> the keywords are "integrated" (as discussed) and "extensible" (how to make
> the framework at least slightly extensible to support some
> standard/vocabulary that we don't explicitly cover in the framework), as
> well as "simple" (we are aiming to provide a framework which is easy to
> understand and use).
>
>

I strongly support that argument. I think the goal of the deliverable  
can or should not be an integrated set of vocabularies that does well  
for one or two use cases. We should rather try to come up with a  
general method how the vocabularies could be integrated and  
exemplarily show it at one or two use cases.

>>> The simplest form of
>>> integration is to assert samePropertyAs; i.e., we should be able to say
>>> property P1 in standard S1 is the same as the property P2 in standard S2.
>>
>> He he he!!! But that works only in the easy case where you assume that your
>> vocabularies strongly overlap.
>
> Not necessarily. It would be indeed helpful as long as there exists any
> overlap.
>
>>  This is unfirtunately now always our case.
>
> I wouldn't say it is *now always" our case, but this is one of the cases
> that the simple framework can/should cover. And an integrated set of
> vocabulary is indeed what the Photo use case asks for. It is important to
> note that we want to include in our simple framework some useful *and*
> easy-to-support features, such as the samePropertyAs relation.
>
>> We
>> rather use different vocabularies that complement each other ... so not
> really
>> "owl:sameAs" or "owl:equivalentClass" or "owl:samePropertyAs" or ....
>

> Indeed, the framework should/could  *also* cover some other simple
> features. The intention of my previous email is to encourage some
> discussion on what else we should consider in order to improve
> interoperability towards this direction.
>
>>
>>> Given the use cases that we have, I wonder how hard it is to identify such
>>> equivalent relations.
>>
>> Oscar did that to some extent with the music ontologies. Is it what you
> have
>> in mind?
>>
>
> It would be nice if Oscar could kindly share with us his opinions w.r.t.
> the music use case. Authors of other use cases are very welcome join the
> discussions too :-)

I don't think that it should be the goal to come up with the least  
common denominator of the vocabularies relevant for one use cases.  
This could lead to the use Dublin Core as the vocabularies are very  
heterogeneous. Thus it would really be good to hear Oscars opinion on  
the music ontologies. Furthermore it is not very easy to come up with  
sameAs or equalAs - Statements between different properties of  
different vocabularies. The problem we would face here is the one of  
schema integration that is researched since many years now. Especially  
in the database community, later for XML Schemas and since a couple of  
years also for ontologies (ontology alignment / mapping).

>
>>> Furthermore, given the use cases, I wonder if there
>>> are any other cross standard/vocabulary relations that we have to handle
>>> in the simple framework.
>>
>> I don't get this question :-(
>
> I meant other simple features that the framework should cover, as
> discussed above.
>

What we went for in one of the European projects that we were involved  
in in the past few years was to define a foundational ontological  
layer that identified different properties of content in general  
(presentation, description, ...) and that then could be used to align  
different vocabularies to it. And thus to achieve interoperability in  
this way.

Best regards,

Tobias


> Best regards,
> Jeff
>
>
>> My 2 c.
>> Best regards.
>>
>>     Raphaël
>>
>> --
>> Raphaël Troncy
>> CWI (Centre for Mathematics and Computer Science),
>> Kruislaan 413, 1098 SJ Amsterdam, The Netherlands
>> e-mail: raphael.troncy@cwi.nl & raphael.troncy@gmail.com
>> Tel: +31 (0)20 - 592 4093
>> Fax: +31 (0)20 - 592 4312
>> Web: http://www.cwi.nl/~troncy/
>>
>>
>>
>>
>
>
> --
>
> Dr. Jeff Z. Pan (http://www.csd.abdn.ac.uk/~jpan/)
> Department of Computing Science, The University of Aberdeen
>
>
>
>
>



-- 
________________________________________
Tobias Bürger, http://www.deri.org
skype: tobitrautich
Received on Thursday, 22 February 2007 06:44:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:21 GMT