[Feedback] Feature suggestion

[Colin Guthrie]

Jeff Mitchell wrote:
> Colin Guthrie wrote:
>> Jeff Mitchell wrote:
>>   
>>>>> I've been talking about this holy grail of media device metadata backend
>>>>> for *years*, but have given up hope if it happening in the short term
>>>>> (seeing as I've personally not got time to do it!).
>>>>>
>>>>> IMO there should not be any database but everything should run off a
>>>>> metadata tracker (e.g. Beagle, Tracker, Spotlight) which would be
>>>>> detailed and flexible enough to allow all the information to be
>>>>> aggregated.
>>>>>         
>>> I believe 2.0 is using Xesam to query Strigi for metadata.  Or will be.  In 
>>> which case, it's a lot closer to what you talk about than anything else seen 
>>> so far...
>>>     
>> In the words of the great thinker, Eric Cartman, "Kick Ass" :)
>>
>> It will be interesting to see how this all pans out and I really hope
>> media apps of the future will work (in some capacity) in this way. Not
>> that I'd ever convert, but it would be cool to fire up e.g. Rhythm Box
>> and get the same playlist, playcounts and such as I have in Amarok. If
>> Strigi or Tracker or Beagle or any other Xesam compatible system is used
>> to provide the source of metadata then this is in theory possible.
>>   
> No, you're mixing up different things here.  Strigi and Beagle (and
> parts of Tracker) aren't used for random storage, they index files.  So
> they could serve as sources of data that we Xesam query to get a list of
> what tracks are around the filesystem that we can play.  But random
> metadata storage would be more of a Nepomuk function, which Tracker
> could maybe do too.

Granted, but things like playcount etc. could in theory be stored in
e.g. ID3 tags, vorbis comments etc. in which case Strigi, Beagle etc.
*could* return that metadata (that was the vague logic behind my
comments above!).

Col