On Mon, 19 Jun 2006 12:15:19 +0200, Bogdan Butnaru wrote:

On 6/19/06, Chris Bransden <[EMAIL PROTECTED]> wrote:
by saying that "all
indentically named tracks are indenticle in contet" you require users
to have heard all instances of the track in question.

It's not a bijection, it's an injection from tracks to titles. It
means that "one song" (the exact same version) should (almost always)
have a single title. There can be multiple versions with the same
title (though that's usually avoided using version info; each song
should have exactly one "title+version info" pair, though the version
info can be empty for some, preferably only one).

There is no entity in the MusicBrainz database representing "songs".

I think that is the flaw in this argument. Yes, if there was an entity representing a song, then it would be a good idea to have only one name for this entity (indeed it would be stupid to give one entity several names). But *there is no such entity*! There is only an entity that represents tracks.

So, please do not write "song" when you should write "track".
For more details please read ObjectModel on the wiki.

A track is always on an album. If you force the title of two tracks to be identical if the tracks are based on the same song, then you force editors to remove all information from the track title that is contextually relevant to one release only. This is what the debate is about.

IMO it is a very bad idea to force an entity to semantically represent something which differs considerably from its structural role in the relational database.

  DonRedman



--
Words that are written in CamelCase refer to WikiPages:
Visit http://wiki.musicbrainz.org/ the best MusicBrainz documentation around! :-)

_______________________________________________
Musicbrainz-style mailing list
Musicbrainz-style@lists.musicbrainz.org
http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style

Reply via email to