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

about representing persons (and other things) in our ontology

From: Pierre-Antoine <pierre-antoine.champin@liris.cnrs.fr>
Date: Tue, 17 Nov 2009 15:22:40 +0100
Message-ID: <4B02B1B0.9070907@liris.cnrs.fr>
To: "public-media-annotation@w3.org" <public-media-annotation@w3.org>
Hi all,

I'm continuing that we started today during the teleconf.

Several ma: properties relate to complex objects, like persons and
organizations for the fields ma:creator, ma:contributor...

As I see it, the in-scope format range from representing those object as
opaque labels (e.g. ID3), to structured embeded data (e.g. TV-anytime
?), to a URI representing the object itself, relying on other formats to
describe it (e.g. Media RDF).

This leads, in my view, to two consequences :

- our ontology should acknowledge the fact that *there is such thing* as
an agent (person or organization), which is the range of ma:creator and
ma:contributor, and that can descrived in a more or less structured way
(from plain label to external linked data).

- our API should provide ways to reflect the different kinds of
representations. My suggestion is to allow two representations :
 * a text label (which can be constructed by aggregating the fields
   of a structured representation, if any)
 * a URI (if one is provided by the underlying format, which may
   give access to a rich description according to the linked data

I don't think we should commit into any more structured description,
which is bound to hinder interoperability.

On the other hand, I don't think we should give up the URI -- when it is
provided, of course; I have no intent to *produce* linked data from
legacy formats! But *if* linked data is available, we should give access
to it. This is an API for the Web, and linked data is the *webbish* way
of providing additional metadata. And finally, it *is* in the scope of
the WG, since DC and Media RDF are.


Received on Tuesday, 17 November 2009 14:23:16 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:17:35 UTC