[oslc-core] Fwd: [oslc-cm] Proposed Discussion/Comment resource model for inclusion into Core

[Dave]

Sorry, I meant to respond publicly to Olivier...


Le mercredi 28 avril 2010 à 09:33 -0400, Dave a écrit :
> On Wed, Apr 28, 2010 at 5:01 AM, Olivier Berger
> <olivier.berger at it-sudparis.eu> wrote:
> > Le mardi 27 avril 2010 à 12:35 -0400, Steve K Speicher a écrit :
> >> The CM workgroup has defined a resource model for Discussions and
> >> Comments see:  http://open-services.net/bin/view/Main/CmComments
> >>
> > Sorry I haven't kept up with the ongoing works in Cm recently and didn't
> > raise this earlier.
> >
> >> This appears to be very applicable to many domains.  I'd propose we
> >> move this to Core spec/guidance and it live in the core namespace.
> >> Any objections?
> >
> > May I suggest to adopt more generally used/popular Ontology for these
> > like SIOC [0], as is done for DC for other "common" properties ?
>
> Is SIOC really popular or even generally used? I've been aware of SIOC
> for a number of years now and considered supporting it in Apache
> Roller, but it never seemed to have much traction

I'm not completely sure, and it probably depends what mesure mechanism
would be used, which environments, etc. But as far as I know, speaking
about RDF ontologies, it seems to me.
http://sioc-project.org/blog and their tales from the SIOCosphere will
probably give more details... unless it's shameless marketing ;)

>  -- and it reinvents
> Atom format, which does not really seem like a good thing.

Unless that Atom is no XML/RDF :-( And I believe RDF is a very doog
design decision for OSLC (let aside collections that use the non-RDF
Atom XML ;).

>
> Have any major social software sites or products adopted it yet? I see
> that the SIOC project has developed plugins for some major software
> packages, but I don't see signs of real adoption in Wordpress.com,
> Typepad.com, Blogger.com, etc.

Well the same stands for the Semantic Web as a whole in any case.

>
> Regardless of popularity, adopting properties from SIOC may still be a
> good idea.
>

It will nevertheless be more popular than OSLC ;)

>
> > SIOC already pretty well defines a lot of posts, messages, comments and
> > likes, and already is quite used in the Semantic Web applications (blog
> > tools, etc.).
>
> What specific SIOC properties would you propose that we use to model comments?

sioc:Post mainly ... see http://rdfs.org/sioc/spec/#sec-overview and its
example.

As you may notice in the EvoOnt BOM Ontology [0] the bom:Issues were
inheriting from a sioc:Thread, i.e. an Issue / Change Request could as
well be the starting point of a conversation.

Same would probably hold for any Resource that users may
discuss/comment, etc.

Such inheritence design decision would probably be good for OSLC too.

>
>
> > I'm not sure it's really ALM specific to model comments and
> > conversations, so I wouldn't think it worth reimplementing as part of
> > OSLC, which should focus primarily to real domain-specific concepts.
>
> This is true. Then again, maybe we don't need a full distributed,
> threaded, etc. conversation model -- we're just after simple flat
> comments and that's what Steve has proposed using standard properties
> except for oslc_cm:body and oslc_cm:inReplyTo.

Hmmm... it it needs to fit any tool (thinking about the various
different ways bugtrackers manage bug discussions for -CM here), then an
existing model that already seems to fit popular applications would
probably be useful... even if more difficult to implement than a simple
one, but it will have less shortcomings from the start (assuming that
SIOC is full-featured in my reasoning).

>
> - Dave

Hope this helps.

Best regards,

[0]: http://baetle.googlecode.com/svn/evoont/trunk/evoont.png
--
Olivier BERGER <olivier.berger at it-sudparis.eu>
http://www-public.it-sudparis.eu/~berger_o/ - OpenPGP-Id: 2048R/5819D7E8
Ingénieur Recherche - Dept INF
Institut TELECOM, SudParis (http://www.it-sudparis.eu/), Evry (France)