W3C home > Mailing lists > Public > public-xg-mmsem@w3.org > November 2006

[MMSEM-UC] Music Use case: comments

From: VassilisTzouvaras <tzouvaras@image.ntua.gr>
Date: Tue, 28 Nov 2006 00:45:40 +0200
Message-Id: <200611272245.kARMjh8Z018422@manolito.image.ece.ntua.gr>
To: <christian@deit.univpm.it>, <g.tummarello@toomuchspam-removeme-gmail.com>
Cc: <public-xg-mmsem@w3.org>
Dear Giovanni and Christian,


Please find below my comments for the Music Use case.


The Music use case is very interesting and promising. It covers the
interoperability issues that caused by the different tagging standards and
the different RDF vocabularies that are used in music applications. The
usecase presents interesting ideas and covers a broad range of
interoperability issues. However, is not yet easy to read and specific
interoperability issues are not presented. Firstly, I suggest structuring
the usecase according to the structure that has been proposed for the use
cases. You must start with the intro by presenting the usecase, the area
that the usecase belongs to and past work. Then is the motivating example
presenting the specific interoperability issues. These issues must be
concrete and not too many (one or two). Following this, you provide possible
solutions using semantic web technologies. The possible solutions should no
be too complicated but as simple as possible showing the added value of
using semantic web technologies. Finally, the conclusion and the references.



In the current version of the usecase, the intro starts with the motivating
example ("Mary has a large music collection."), which should be in the
motivation section and then is presenting some info about ID3 standard
without giving an intro to this area (apart from 2 sentences in the
beginning of the document). I suggest starting by giving some general info
about Music applications and the general idea of your usecase. ID3 and other
standards (e.g. Ogg Vorbis, WMA tagging format, APEv2 tag) that are used to
annotate music metadata as well as tools and repositories can be included in
the intro section.



In the motivation example section, you can start with the paragraphs of
"Mary has a large music collection" and what Mary requires from the
application (following track). The "Identifications" and "Obtaining
metadata" sections can also be included in the motivation section. You
should then focus on the interoperability issues caused by the different
standards that each tool/repository is using, which result in non-sharing
metadata and content. The interoperability issues should be concrete. If I
understood well, there are two kinds of interoperability issues. The first
issue is on the tagging format that is used (e.g. ID3, Ogg Vorbis). The
second issue is on the RDF vocabularies that they use to share their
metadata. You should identify a *specific* interoperability issue (e.g. in
two specific vocabularies and/or tagging standards) and not trying to cover
all the issues. I suggest starting with metadata expressed in RDF and not
including in this stage metadata expressed in XML that have to be
transformed in RDF using an ontology. Also, the option of creating from
scratch an ontology covering all the aspects of musical objects might not be
feasible in the framework of this XG since it requires a lot of time
creating it.  You may also present small examples showing where the problems


Possible solutions:

In the possible solutions section, you have to provide a possible solution
for the specific interoperability issue using semantic web technologies
(e.g. create mappings between terms of the two standards or use an ontology
to harmonise the two schemas). I suggest providing a sample of the solution
(e.g. create mappings for a subset of the two schemas) since it might be
difficult to provide a complete solution. A suggestion is to work on  the
vocabularies that MusicBrainz and Bitzi use. Examine how you can share and
collect metadata from these repositories. These repositories contain only
basic metadata in RDF and may not be too hard to map/harmonise them. 



You can include all the urls from the tagging formats and repositories in
the references section.




Received on Monday, 27 November 2006 22:45:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:01:24 UTC