W3C home > Mailing lists > Public > public-media-annotation@w3.org > November 2009

RE : [mawg] RE: [q] MAWG: Definition of subproperties

From: Evain, Jean-Pierre <evain@ebu.ch>
Date: Fri, 20 Nov 2009 10:16:09 +0100
To: Pierre-Antoine <pierre-antoine.champin@liris.cnrs.fr>, Felix Sasaki <felix.sasaki@fh-potsdam.de>
CC: "Bailer, Werner" <werner.bailer@joanneum.at>, Joakim Söderberg <joakim.soderberg@ericsson.com>, Tobias Bürger <tobias.buerger@sti2.at>, "public-media-annotation@w3.org" <public-media-annotation@w3.org>
Message-ID: <7D1656F54141C042A1B2556AE5237D6001006EB6CC71@GVAMAIL.gva.ebu.ch>
PA,

do you mean a property/sub-property like title / (title) type? or contributor / role? without specifying what the type or role is but allow mapping to what is available from other descriptions.

JP

________________________________________
De : Pierre-Antoine [pierre-antoine.champin@liris.cnrs.fr]
Date d'envoi : vendredi, 20. novembre 2009 09:30
À : Felix Sasaki
Cc : Bailer, Werner; Evain, Jean-Pierre; Joakim Söderberg; Tobias Bürger; public-media-annotation@w3.org
Objet : Re: [mawg] RE: [q] MAWG: Definition of subproperties

Hi all,

For the record, I also agree with the last exchanges btw Werner and
Felix, as well as Felix's proposal below.

My only concern is : shouldn't we make the API ready for subproperties
right now? If we don't, we might find out that we made some choices that
prevent anyone to add them...

As I already suggested, a very basic use of subproperty would be not to
standardize any, but consider them as a placeholder for information
about the *actual* property/sub-property used in the underlying format
-- and even leave it to the implementation to decide how to expressit.
Later on (possibly only in version 2.0), we might still decide to
standardize *some* sub-properties to be put in this placeholder.

  pa

Le 20/11/2009 08:52, Felix Sasaki a écrit :
> Hello Werner,
>
> thank you for the clarifications. There seems to be one open question to
> me: Is it OK for us to start work on subproperties although we have not
> reached our goals even with the simple properties? IMO, we should first
> for all simple properties
> - define the API methods (done to some extend)
> - provide test cases and test suite material (not done at all)
> - run the tests with at least two implementations
> so that we can be sure to declare victory, even if we don't get to the
> sub property part.
> So I am saying not "no" about sub properties, but postpone work on them
> until we have done our basic job.
>
> Would you and others agree with that?
>
> Best,
>
> Felix
>
> 2009/11/20 Bailer, Werner <werner.bailer@joanneum.at
> <mailto:werner.bailer@joanneum.at>>
>
>     Dear Felix,
>
>     >> > it is a matter of scope as I said before. If you want a
>     vocabulary in
>     >> > RDF to which people map, fair enough and I was hopping this was the
>     >> > original task of the group.
>     >> I agree with Pierre-Antoine that defining an RDF vocabulary to map to
>     >> and defining an API to access the properties in this vocabulary
>     are not
>     >> contradicting goals. In fact, this is how some of the mapping
>     >> approaches that several people in the group have implemented so far
>     >> work internally: map to the properties defined in our vocabulary
>     >> (represented in OWL) and then provide getter functions to access
>     these
>     >> values. So an implementation getAsRDF() function that gives you the
>     > source metadata expressed with our vocabulary.
>     >
>     > well, my implementation is XML-based. So I have a very high preference
>     > to not have to implement RDF-processing to be a conformant
>     > implementation. So if we create an RDF vocabulary, it should be only
>     > one serialization of our ontology, and support for it should not be
>     > mandatory. Do you agree with that?
>
>     Yes, I agree. As Doug suggested at the F2F, we can state in the
>     recommendation that the RDF representation is not a mandatory part,
>     however, if someone uses an RDF representation, it needs to conform
>     to the definition in the rec.
>
>     > All questions we discuss below about subproperties are IMO independent
>     > of the question "RDF or not".
>
>     I agree.
>
>     Best regards,
>     Werner
>
>     > > More specifically sub properties will either inevitably go
>     further in
>     > > the direction of one existing vocabulary (then why not adopt) or
>     > > diverge hence significantly reducing the chances oif mapping in
>     > profit
>     > > of a 'nice new vocabulary'.  By the way, by which magic this
>     group in
>     > > particular would come with the super nice new solution?  All
>     those we
>     > > are mapping to tried to achieve that goal.
>     > I absolutely agree that we face these risks. However, as you said, we
>     > aim at mapping existing formats. If we do not consider subproperties,
>     > we can only map on a very coarse level, leaving it to the application
>     > developer to handle the semantics of subproperties coming from
>     > different formats. In my opinion supporting a defined set of
>     > subproperties is useful for the following reasons:
>     >
>     > - Allowing users of the API to have well defined semantics of
>     > subproperties independent of the source format. If the
>     subproperties of
>     > a property agree in many of the formats, then we do not invent
>     anything
>     > new and we do not complicate mapping for these formats. If they do not
>     > match, we have to decide which format to follow and map the others.
>     >
>     > - There are cases where properties are defined on a different
>     > granularity level. In a format like MPEG-7, we could say we just
>     map to
>     > creator, and leave the subproperty to a role classification scheme.
>     > However, ID3 for example has distinct properties for composer, text
>     > writer, etc., so in such a case we have to make a decision about
>     how to
>     > map to subproperties of creator.
>     >
>     > - All subproperties that we do not define in our set can still be
>     > passed on, leaving it to the application developer what to do with it,
>     > but not losing the information.
>     >
>     > Best regards,
>     > Werner
>     >
>     > > ________________________________________
>     > > De : Pierre-Antoine [pierre-antoine.champin@liris.cnrs.fr
>     <mailto:pierre-antoine.champin@liris.cnrs.fr>]
>     > > Date d'envoi : mercredi, 18. novembre 2009 19:00
>     > > À : Evain, Jean-Pierre
>     > > Cc : 'Felix Sasaki'; Joakim Söderberg; Tobias Bürger; public-media-
>     > > annotation@w3.org <mailto:annotation@w3.org>
>     > > Objet : Re: [mawg]  RE: [q] MAWG: Definition of subproperties
>     > >
>     > > Jean-Pierre, Felix,
>     > >
>     > > If I understand your point, our focus should be on enabling existing
>     > > formats to map to the API, not on hoping that people will drop those
>     > > format in favor of RDF with an ad-hoc new vocabulary. And I agree
>     > with
>     > > that.
>     > >
>     > > However, I don't think that defining an API and defining a language
>     > are
>     > > so different tasks as you seem to imply. An API is a vocabulary
>     for a
>     > > programming language. An ontology is (in the SemWeb world) a
>     > vocabulary
>     > > for RDF. We are to focus on the API, granted, and the ontology is
>     > here
>     > > only to formally define the mapping. But let's not discard a
>     proposal
>     > > just because it is about a "vocabulary".
>     > >
>     > > I think subproperties would help provide a fine-grain access to the
>     > > underlying metadata. I think that subproperties should be an
>     optional
>     > > attribute of our return values. So users could just stick to the
>     > value
>     > > (interpreted in the context of the ma: property they retrieved, or
>     > they
>     > > could try to make more sense out of the provided sub-property.
>     > >
>     > > Whether we should define our own set of subproperty or just reflect
>     > > those provided by the underlying format... I'm still not sure. Both
>     > > options have their advantages.
>     > >
>     > >   pa
>     > >
>     > >
>     > >
>     > > Le 18/11/2009 15:28, Evain, Jean-Pierre a écrit :
>     > > > HI Felix,
>     > > >
>     > > >
>     > > >
>     > > > That was also my perception.
>     > > >
>     > > >
>     > > >
>     > > > R, JP
>     > > >
>     > > >
>     > > >
>     > > > *From:* felix.sasaki@googlemail.com
>     <mailto:felix.sasaki@googlemail.com>
>     > > [mailto:felix.sasaki@googlemail.com
>     <mailto:felix.sasaki@googlemail.com>]
>     > > > *On Behalf Of *Felix Sasaki
>     > > > *Sent:* mercredi, 18. novembre 2009 15:20
>     > > > *To:* Evain, Jean-Pierre
>     > > > *Cc:* Joakim Söderberg; Tobias Bürger; public-media-
>     > annotation@w3.org <mailto:annotation@w3.org>
>     > > > *Subject:[mawg] * Re: [q] MAWG: Definition of subproperties
>     > > >
>     > > >
>     > > >
>     > > > Hi Jean-Pierre,
>     > > >
>     > > > we are definitely working on a tool to map with a number of
>     > existing
>     > > > formats. Note also that all the concrete implementations we
>     have so
>     > > far
>     > > > go this route, and the browser scenario proposed by Silvia does
>     > that
>     > > > too. The "new" properties are no new vocabulary, but just a means
>     > to
>     > > > make clear what the smallest common nominator between existing
>     > > formats is.
>     > > >
>     > > > Best,
>     > > >
>     > > > Felix
>     > > >
>     > > > 2009/11/18 Evain, Jean-Pierre <evain@ebu.ch
>     <mailto:evain@ebu.ch> <mailto:evain@ebu.ch <mailto:evain@ebu.ch>>>
>     > > >
>     > > > Joakim,
>     > > >
>     > > > Here we are.  Are we redefining a new Dublin Core ("a nice
>     > vocabulary
>     > > > that others would like to use" as you call it) or are we
>     working on
>     > a
>     > > > tool to map with a number of existing formats?
>     > > >
>     > > > It is a matter of scope.
>     > > >
>     > > > In any case, if we want to do the former then we'd better work on
>     > RDF
>     > > > and leave those who have used other formats map to it. Then we
>     > would
>     > > > really be working on an ontology although archaic for an ontology.
>     > > >
>     > > > As I said during the F2F, we would now badly need to know exactly
>     > > what
>     > > > we are trying to achieve.
>     > > >
>     > > > Regards,
>     > > >
>     > > > Jean-pierre
>     > > >
>     > > >
>     > > > -----Original Message-----
>     > > > From: public-media-annotation-request@w3.org
>     <mailto:public-media-annotation-request@w3.org>
>     > > > <mailto:public-media-annotation-request@w3.org
>     <mailto:public-media-annotation-request@w3.org>>
>     > > > [mailto:public-media-annotation-request@w3.org
>     <mailto:public-media-annotation-request@w3.org>
>     > > > <mailto:public-media-annotation-request@w3.org
>     <mailto:public-media-annotation-request@w3.org>>] On Behalf Of
>     > Joakim
>     > > > Söderberg
>     > > > Sent: mercredi, 18. novembre 2009 14:57
>     > > > To: Tobias Bürger
>     > > > Cc: public-media-annotation@w3.org
>     <mailto:public-media-annotation@w3.org> <mailto:public-media-
>     <mailto:public-media->
>     > > annotation@w3.org <mailto:annotation@w3.org>>
>     > > > Subject:[mawg] RE: [q] MAWG: Definition of subproperties
>     > > >
>     > > > Jean-Pierre did not like it because he doesn't believe that there
>     > > will
>     > > > be mappings to all sub properties from all formats. Ex. "album
>     > title"
>     > > in
>     > > > TVA, ID3 etc.
>     > > >
>     > > > He has a point but I think that if we define a nice vocabulary
>     that
>     > > > becomes popular, more mappings will follow from several
>     > contributors.
>     > > > Which by the way inclines that we should make it possible in the
>     > > future
>     > > > to (easily) update the ontology. But I guess that has to do with
>     > the
>     > > > implementation.
>     > > >
>     > > > /Joakim
>     > > >
>     > > > -----Original Message-----
>     > > > From: Tobias Bürger [mailto:tobias.buerger@sti2.at
>     <mailto:tobias.buerger@sti2.at>
>     > > > <mailto:tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>>]
>     > > > Sent: den 18 november 2009 13:57
>     > > > To: Joakim Söderberg
>     > > > Subject:[mawg] Re: [q] MAWG: Definition of subproperties
>     > > >
>     > > > Hi Joakim,
>     > > >
>     > > > I agree, having subproperties is like extending the core set with
>     > > > qualifying terms for each of the attributes we defined. I think
>     > that
>     > > > having subproperties could give us a more precise way to
>     define the
>     > > > mappings and not to end up in being too generic in parts where
>     most
>     > > of
>     > > > the formats we have in scope are more specific.
>     > > >
>     > > > I have seen that there were some people that did not like the idea
>     > of
>     > > > subproperties @ the F2F. What were their arguments (if you
>     > remember)?
>     > > >
>     > > > Best regards,
>     > > >
>     > > > Tobias
>     > > >
>     > > > Joakim Söderberg wrote:
>     > > >> Hi Tobias, thanks for accepting the AP.
>     > > >>
>     > > >> I had a look at the properties defined in AMG (All Media Guide;
>     > see
>     > > > Video_tables) and EBU (see zip file).
>     > > >>
>     > > >> The more I think about the sub-properties the more I think they
>     > are
>     > > an
>     > > > integral part of the Ontology. It's like extending the core set
>     > with
>     > > > qualifying terms for each core attribute, or what do you think?
>     > > >>
>     > > >>
>     > > >> All the best
>     > > >> Joakim
>     > > >>
>     > > >> -----Original Message-----
>     > > >> From: Tobias Bürger [mailto:tobias.buerger@sti2.at
>     <mailto:tobias.buerger@sti2.at>
>     > > > <mailto:tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>>]
>     > > >> Sent: den 18 november 2009 08:32
>     > > >> To: Joakim Söderberg
>     > > >> Subject:[mawg] [q] MAWG: Definition of subproperties
>     > > >>
>     > > >> Hi Joakim,
>     > > >>
>     > > >> as you might know I got the action to work on the subproperties
>     > for
>     > > > the properties we defined in the ontology. I started to read what
>     > you
>     > > > discussed during the F2F and also talked to Florian yesterday to
>     > > discuss
>     > > > this.
>     > > >> I have seen that you have already started on this issue at
>     > > >> http://www.w3.org/2008/WebVideo/Annotations/wiki/Sub_Types
>     > > >> I just wanted to ask you which sources you already considered,
>     > i.e.
>     > > >> where you looked for possible subproperties? Based on that I can
>     > > start
>     > > > of working on the subproperties.
>     > > >>
>     > > >> Thanks a lot in advance for your answer!
>     > > >>
>     > > >> Best regards,
>     > > >>
>     > > >> Tobias
>     > > >>
>     > > >> --
>     > > >> _________________________________________________
>     > > >> Dr. Tobias Bürger
>     > > >>
>     > > >> STI Innsbruck
>     > > >> University of Innsbruck, Austria
>     > > >> http://www.sti-innsbruck.at/
>     > > >>
>     > > >> tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>
>     <mailto:tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>>
>     > > >> __________________________________________________
>     > > >>
>     > > >
>     > > > --
>     > > > _________________________________________________
>     > > > Dr. Tobias Bürger
>     > > >
>     > > > STI Innsbruck
>     > > > University of Innsbruck, Austria
>     > > > http://www.sti-innsbruck.at/
>     > > >
>     > > > tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>
>     <mailto:tobias.buerger@sti2.at <mailto:tobias.buerger@sti2.at>>
>     > > > __________________________________________________
>     > > >
>     > > >
>     > > >
>     > > >
>     > > >
>     -------------------------------------------------------------------
>     > --
>     > > ---
>     > > >
>     > > > * ************************************************** This
>     email and
>     > > any
>     > > > files transmitted with it are confidential and intended solely for
>     > > the
>     > > > use of the individual or entity to whom they are addressed. If you
>     > > have
>     > > > received this email in error, please notify the system manager.
>     > This
>     > > > footnote also confirms that this email message has been swept by
>     > the
>     > > > mailgateway ************************************************** *
>     > > >
>
>
Received on Friday, 20 November 2009 09:18:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 20 November 2009 09:18:35 GMT