[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

RDF export of debbugs data and contribution to LinkedData for Debian bugs



Hi.

(I've discussed that with a few Debianers already, so apologies is this
is nothing new to you, and also, as I'm new to the list, pardon my
ignorance of previous discussions on same topic.)

We've been investigating the use of RDF and ontologies to represent
bugs, and have setup a demonstration server which publishes facts about
bugs in Debian (and associated entities, like foaf:profiles, and some
packages relations) as RDF, starting from UDD database, using triplify.

Here are more details :
https://picoforge.int-evry.fr/cgi-bin/twiki/view/Helios_wp3/Web/TriplifyUddToRdf (including a link to the demo server running on a recent UDD snapshot).

I'd like to discuss the idea of publishing such facts about Debian bugs
as RDF right from debbugs.

This would offer the advantage to allow debbugs to participate to the
Linked Data (http://linkeddata.org/) movement aka Semantic Web.

The idea is pretty simple :
 http://bugs.debian.org/534845, would be both a URI identifying :
        - a resource on the semantic web (subject of predicates in
        RDF), 
        - and a URL that could resolve to :
                . a RDF document (similar to the content of
                http://kilauea.int-evry.fr:8081/triplify/UDD/issue/534845 for instance), when requested by "machines"
                . a HTML document, when requested by humans, i.e.:
                http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=534845
                
This would allow to construct applications that could crawl the Semantic
Web and find <http://bugs.debian.org/534845> in other RDF documents
which would refer to such a Debian bugs, and all the possible
applications of the Semantic Web that one can imagine.

One such application could be the interlinking of bugs in different
bugtrackers through such semantically tagged relations, on the Semantic
Web, instead of in a semi-closed database (as is in LaunchPad ;), so
that tools like bts-link could make use of them.

I think our work on reusing the EvoOnt BOM ontology and extending it
with our Helios BT ontology could serve as a reference for an
implementation in Debbugs, as our server can be queried to check how we
produce RDF (which might be buggy, btw ;).

I'd like to get feedback, comments, and any suggestions on this idea.

Best regards,

P.S.: for those wondering the context, we're working on
https://picoforge.int-evry.fr/cgi-bin/twiki/view/Helios_wp3/Web/ and I'm
also interested in improving bts-link to which I have contributed in
recent weeks. And more about me at :
http://kilauea.int-evry.fr:8081/triplify/UDD/carnivore/949 or
http://www-public.it-sudparis.eu/~berger_o/foaf.rdf ;-)

-- 
Olivier BERGER <olivier.berger@it-sudparis.eu>
http://www-public.it-sudparis.eu/~berger_o/ - OpenPGP-Id: 1024D/6B829EEC
Ingénieur Recherche - Dept INF
Institut TELECOM, SudParis (http://www.it-sudparis.eu/), Evry (France)


Reply to: