Re: [MMSEM-UC] Questions on the tagging use case

Dear Jeff,

To complement the answers already given by Thomas:

> 1) This is a very exciting use case. Am I right in saying that  the
> interoperability here is among different tagging systems, rather than
> that of existing multimedia standards? Or there are some implicit
> relations between the two?

I would answer yes and no :-) The use case clearly aims to solve
interoperability problems between various tagging platforms, but implicitely
between the underlying way they handle the metadata too ! So, as you suggest
in the second part of your question, there is clear relationships between
the two.
Actually, each platform provides functionalities, interfaces for tagging,
searching and retrieving multimedia resources. However, you can't use the
metadata (the tags) used in a particular platform (e.g. Flick) to search
items in another one (e.g. YourTube). But you could expect that a end user,
that captures an event with photos and videos, and even though he stores
them on different platform, would prefer to make this link explicit !
This is what this use case report: the partial solution being providing a
unified and uniquely way for referencing the tags used, thus allowing to use
the same tags on various platforms. Which brings to the second question:

> 2) I am not an expert of SKOS and could not quite see how SKOS can help.
> Could you give an example or clarify this point?

Yes ! So, in this use case, the asumption is that the user annotates the
multimedia items with "tags". If by definition, the tag is a free keyword
term (or an expression), people think that it would be useful to structure
the tags. However, they do not want to model "ontologies", since the tags
should remain a term (and not a concept).
SKOS is the perfect language for representing this Simple Knowledge
Strcuture (as the name suggest).

And I would recommend you to follow the interesting work of the SWD WG that
will push SKOS to a recommendation.
For your personal interest, there is outstanding issue of how SKOS could be
modeled in OWL DL (status and semantics of owl:AnnotationProperties, etc
...).

More about all that in Athens.
Best.

    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/

Received on Sunday, 3 December 2006 20:05:32 UTC