Separate Interpretation/Manifestation into different groups for Events and Subjects

[Siegfried Gevatter]

Currently datamodel.Interpretation and datamodel.Manifestation contain both the ontologies for Events and for Subjects.

This is confusing in several ways:
 - In the docs (be it HTML or using Python's in-build help), as everything is mixed together.
 - In code, as it makes it impossible (without ugly hacks) to get, eg., a list of all Subject Interpretations, as they can't be separated from Event ones.
 - It just doesn't make sense.

Therefore, I propose that we separate those collections so that we have:
 - datamodel.EventInterpretation
 - datamodel.EventManifestation
 - datamodel.SubjectInterpretation
 - datamodel.SubjectManifestation

Also, while we are at it, we should consider hiding the INTERPRETATION_ID and INTERPRETATION_DOC variables, unless they are useful for something.

As this is an API break, I also propose that datamodel.Interpretation and datamodel.Manifestation should continue working (showing a DeprecationWarning) the same release as the change is done, and then dropped with the next one.

[Markus Korn (thekorn)]

I think we should not put much energy (and resulting API) changes into it for now, because the whole handling of `Symbols` will change once we start working on [0]. Over the last weekend I started thinking about writing a 'libzeitgeist' using vala, but one of the first annoyance was that I need to define our interpretations and manifestations over there again. I think we should start working on this spec ASAP, so we can fix two main points of this bugreport ("getting all {event/subject}interpretations/manaifestations" and "doc structure") 'by accident' ;)
Plus, everything will make much more sense.

[0] https://blueprints.edge.launchpad.net/zeitgeist/+spec/events-ontology

[Seif Lotfy (seif)]

I agree with markus. We should refocus on events-ontology since it might affect a bit of the datamodel. I will sit on it this weekend

[Siegfried Gevatter (rainct)]

>> Over the last weekend I started thinking about writing a 'libzeitgeist' using vala

Funny, I was about to do the same (in C) but didn't write anything because I didn't know where to start :P.

[Seif Lotfy (seif)]

seems like it gonna be done with the new ontology huh?

[Markus Korn (thekorn)]

We implemented this kind of differentiation with the ontology branch we landed some time ago.