Problems with NCO::IMAccount

[George Goldberg]

2010/5/11 Dario Freddi <drf54321 at gmail.com>:
> Hello people,
>
> I'm writing you about some issues I found with NCO::IMAccount. These days, I'm
> adding to Nepomuk the relevant information for Audio/Video chat support, and I
> found some problems with the ontology itself.
>
> Audio/Video chat support in Nepomuk is implemented through a different class,
> {Audio,Video}IMAccount. This creates the 2 following problems for me:

Agreed that this sucks.

>  - It is not guaranteed that an account capable of sending videos is capable
> of sending audio as well. I know, it's a weird case, but it might happen:
> however, this cannot be dealt with with the current ontology, where
> VideoIMAccount c AudioIMAccount.
>
>  - The second one, which is slightly more serious, is the fact that a contact
> capability of sending Audio or Video might change in time, and I really don't
> know if it is possible to change the class of a Resource. Also, is there a
> sane way, given a resource, to find out if it is a subclass of something?
>
> In the meanwhile (just for avoiding getting stuck), I added to Telepathy
> ontology 2 bool fields, supports{Video,Audio}, but I'd really like to see this
> fixed upstream, as I consider what I did in my ontology a temporary
> workaround. However, I think that something similar - more than adding some
> subclasses - would work out better.

The correct solution to this problem is to support "capabilities" on
an IMAccount, much like how Telepathy does it. "Audio" would be a
capability, as would "Video" and other things (can't think of any
examples off the top of my head, but there are plenty around).

I believe tracker have come up with some ontology stuff for
Telepathy-like "capabilities" - we should look at what they've done
and see if we can use it/base work off it to avoid reinventing the
wheel.

Cheers,

George