Re: [Tracker] Zeitgeist ontology and Tracker



On 31/03/11 08:39, Ivan Frade wrote:
Hi,

  The big difference between that original NEO (Nepomuk Event Ontology)
and the zeitgeist proposal is how to represent an instance of an
event.

  In NEO we propose:
<x>  a neo:Event ;
    neo:hasManifestation<neo:pre-defined-manifestation-access>
    neo:hasInterpretation<neo:pre-defined-interpretation-user-activity>
    neo:hasActor<urn:software:123123123>    # This being a neo:software

  What ZG is proposing is more in the line of:
<x>  a neo:Event, neo:AccessEvent, neo:UserActivity
     neo:hasActor<blabla>

  [They do it "overriding" rdf:type property, but at the end is
translated into this]

The first option is consistent with our use of the ontology everywhere
else, and Interpretations and Manifestations don't have specific
properties, so there is no need of a class for them. It is also closer
to the Zeitgeist model, which is a good hint that it goes in the right
direction.

I think Mikkael overall agrees on this, but then comes the problem of
API/ABI in zeitgeist. IIRC they are exposing the event
manifestations/interpretations via #defines (which is ok to update)
but some applications are using directly DBus... so we are not
completely free there. We should take a look into this more carefully
and hopefully we can make a smooth transition. This is a ZG-tracker
combined effort.

Thanks for the comments Ivan.

Actually, the first thing that comes to mind is, can we alias the ontologies to allow both names for smooth transition (without duplicating the data in the db). I haven't seen anything like this from what I remember. Juergbi, any comment there?

--
Regards,
Martyn



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