Re: [Rhythmbox-devel] Proposal for first-generation Musicbrainzintegration

[Joshua Haberman <joshua haberman com>]

On Thu, 2003-11-20 at 21:11, Jeremy Nickurak wrote:
> On Thu, 2003-11-20 at 16:58, Joshua Haberman wrote:
> > bstacles are there to a simple level of musicbrainz integration
> > like this?
> 
> UI-wise: allowing the user to perform such operations on multiple files
> simultaneously.
> 
> I'd really really like to be able to:
> 1) Apply musicbrainz info to many/all of my files in one UI process
> 2) Apply selective changes to multiple files in one UI process. (Eg:
> Select all instances of Weird Al music, and in one swoop, change all
> their Artist tags to be consistent

Baby steps!  :)

Using musicbrainz in batch is difficult because for each song you look
up there can be several songs in the db that match.  You want to select
the right one from the results that the db returns, and a GUI to do this
for many songs at once is nontrivial.

Regarding the second idea, this is *exactly* the reason that I think
files should have musicbrainz track/artist/album ids as the primary way
for determining if two artists are really the same thing, as opposed to
capitalization-disregarding/"the"-disregarding/etc hacks.  See
http://www.musicbrainz.org/db_structure.html for the database model,
which provides unique ids for every artist, album, and track.  See
http://www.musicbrainz.org/docs/specs/metadata_tags.html for how you can
put these ids in MP3/Ogg/FLAC files.  If the browser could use these ids
for artist/album/track, then you would never have the problem of a
single artist under different names ("Wierd Al", "Weird Al Yancovich",
etc.) *ever* *again*.  It's the right solution to the problem.

Josh