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

RE: IBBT Metadata model + return values

From: Chris Poppe <Chris.Poppe@ugent.be>
Date: Mon, 5 Jan 2009 13:54:33 +0100
To: <public-media-annotation@w3.org>
Cc: <fsasaki@w3.org>, "'Rik Van de Walle'" <rik.vandewalle@ugent.be>, <erik.mannens@ugent.be>
Message-ID: <3814017939054C87872247C6D276BDEF@elis.UGent.be>

Dear,

Please find my answers enclosed below,

Kind regards,

Chris Poppe
 
Ghent University - IBBT
Faculty of Engineering
Department of Electronics and Information Systems (ELIS)
Multimedia Lab Gaston Crommenlaan 8 bus 201
B-9050 Ledeberg-Ghent
Belgium
 

t: +32 9 33 14959
f: +32 9 33 14896
t secr: +32 9 33 14911
e: chris.poppe@ugent.be

-----Original Message-----
From: Felix Sasaki [mailto:fsasaki@w3.org] 
Sent: dinsdag 16 december 2008 15:09
To: Erik Mannens
Subject: IBBT Metadata model + return values

Hi Erik,

Chris presented the IBBT metadata model to us in Ghent, and we would
like to have information about the mappings it defines. Are these
available somewhere?

[CP] The mappings are currently not available since these are only partially
implemented. Note that the mappings are defined to use OWL constructs like
equivalentClass, so using standardized ways to link different ontologies
with each other (The IBBT project required only a minimal mapping). 
Additionally, we looked into the usage of rules to define instance
equivalence (e.g., to define that a person (read the instance of a class
Person_ontolgoy_A) described in one ontology is equivalent to a person
described in another ontology  (read the instance of a class
Person_ontolgoy_B). So, I see the mapping as the combination of
OWL-constructs and rules that would allow to map instances on each other.
(These are currently not supported in the system since it is hard to
incorporate a rule engine in the framework used within the IBBT project).


 Also, we are having a question on return values:
for each property is there just one type of return value, or are there
several ones? Eg. if you query for the name of the creator of a media
object, would you always get a string value, or sometimes something else
(e.g. an URI)?

[CP]Our metadata model defines the actual type of these
(DataType)properties. So for instance the name of the creator would in our
case be a string. The metadata service (which contains a web service holding
methods relevant for the use cases of the IBBT project) can of course
convert/interpret this string to whatever is needed... 
In the model we make the separation between a name (information about the
person) and an identifier (which refers to the actual person).
(What's in a name :)).
I guess I would favour multiple properties when confusion is possible.
Although this might increase the size of the resulting ontology.

Agreeing on a common format for a property, which can not be misinterpreted,
is necessary for interoperability between systems.



It would also be great to have Chris directly particpating in the group,
if that would be possible ...

[CP]In fact, I just joined the mailing list. I remember that I have an
action point in this WG, namely the mapping table for DIG35. I was wondering
in what format we need to distribute this and how?

Many thanks for the information in advance,

[CP] You are most welcome.

Felix
Received on Monday, 5 January 2009 13:44:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 5 January 2009 13:44:30 GMT