Data Engines [WAS: KDE/kdebase/workspace/plasma/dataengines/nowplaying]

[Jason Stubbs]

On Tuesday 15 July 2008 02:51:53 JST, Aaron J. Seigo wrote:
> On Monday 14 July 2008, Jason Stubbs wrote:
> > But! If the above is correct, I wonder if it is not possible to make it
> > so that a an applet can (only?) request a list of data engines that
> > report "FOO" data. In the case of "now playing", each data engine (for
> > each player) could report data such as:
>
> this would expose to the visualization the implementation detail that there
> are N different sources of data being agregated. in this case, there wasn't
> even a change in the dataengine's exported set of sources (which is the
> whole point of dataengines).
>
> in the case of weather, which is the same concept, the engine itself
> supports plugings and things are agregated in the dataengine that the
> visualization interacts with.

Hmm.. perhaps that plugin model could be generalized and supported by plasma 
then, rather than each data engine having to roll its own? I'm playing 
devil's advocate here, but if hordes of people start coding applets left, 
right and centre there's bound to be a subset also wanting to code data 
engines and/or data engine extensions - which is why I'm questioning the 
above.

I just realized there's also the problem of user provided data engines 
identifying themselves by the same name (unless it's already covered by the 
APIs) but the issue I'm currently asking about, using "now playing" as an 
example, goes something like:

1) KDE provides a "now playing" data engine that aggregates and reports the 
   status of several different media players.
2) Third parties provide hundreds of applets that utilize this data engine.
3) Joe Bloggs creates a new media player and would like the hundreds of third 
   party applets to be able to report the status of his media player too.

Is the only way for Joe Bloggs to achieve that by getting support for his 
media player integrated into KDE's data engine?

-- 
Jason Stubbs