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

[MMSEM-UC] Searching and Presenting News in the SemanticWeb use case: comments

From: VassilisTzouvaras <tzouvaras@image.ntua.gr>
Date: Tue, 28 Nov 2006 23:19:26 +0200
Message-Id: <200611282119.kASLJVIN003247@manolito.image.ece.ntua.gr>
To: "'RaphaŽl Troncy'" <Raphael.Troncy@cwi.nl>, "'Christian Halaschek-Wiener'" <halasche@cs.umd.edu>, "'Thierry Declerck'" <declerck@dfki.de>
Cc: "'MMSem-XG Public List'" <public-xg-mmsem@w3.org>
Dear Raphael, Christian, Thierry, Laurent and Misha,


Please find below my comments for the “Searching and Presenting News in the
Semantic Web” use case.


The use case is already in a good shape especially the introduction and
motivation sections. It is structured more or less according to the proposed
use case. Generally, I propose to keep a common structure in all use cases
in order to be homogeneous in the deliverable. Please make the appropriate
changes to be fully compatible with the proposed structure.


The Intro section presents well the general idea of the use case. The use
case is focused on the News area and more specifically in the IPTC standards
and the interoperability issues that caused from them. I suggest you also
include in the intro other standards that may be used for news metadata.
Also, you can present the past work in the area, if any. Finally, my opinion
is that the vocabularies (Kanzaki, Norm Walsh) of section 3 should be
included in the intro (a small description of them) and in the vocabulary
document the full description. 


The motivating example section contains three examples. In each of these
examples you should identify specific interoperability issues illustrated by
some examples, if possible. The query example is too sort and do not present
any specific issue. I suppose you are planning to extend it and provide
specific examples that an interoperability issue is identified. The second
example is much more detailed. An issue here is that permission must be
given by the broadcasting companies as soon as possible so you can display
more info about them. You identify the general interoperability issue and
mention that below we can find the large commonalities in the various
metadata documents. However, this is not obvious below. You describe how the
general structure looks like without giving specific examples where are
these commonalities and where the problem is occurred. I suggest including
specific examples (e.g. subset of the schema) from these standards and show
explicitly the problems. In this way you can better work in the possible
solution section where you have to show the solution in the specific
examples. The provided solution has to be easily scalable to all the
interoperability issues. Also, the last paragraph in this example can be
included in the possible solutions section since it gives a hint for the
solution. You may keep it if you do the same for the other two examples of
this section. The third example is under construction and there is not much
to say. You must be careful not to present too many interoperability issues
and then you have to deal with all of them in the possible solutions
section. I can see in the minutes of the last teleco that you are planning
to work with specific problems in NewsML1 and 2, DowJones ML.


I suggest that Section 3, as previously said, must be included in the intro
(a summary of it because the complete description should go on the
vocabulary document). 


Section 4, I suppose, is intended to be the possible solution section. Yet,
you are not providing any solutions apart from general ideas. The second
paragraph seems to me more appropriate to be included in the intro. In this
section you have to provide a more detailed description of the solution that
you more or less describe in the last paragraph. In order to do this, you
must work on the specific problems that should be identified in the
motivating section. Try to work initially with one issue and provide a
solution that could be theoretically used for the whole set of issues. If I
understood well, you are proposing to use formal semantics (such as OWL) and
maybe rules or DL-axioms for discovering subscription matches. I suggest
that you start making the solution more explicit by presenting which
semantic web technologies you are planning to use and how. Are you planning
to create mappings or a vocabulary that will harmonise (in an upper level)
the existing vocabularies/standards?



Received on Tuesday, 28 November 2006 21:20:26 UTC

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