Why schemas and not code?

[David Cantrell]

On Fri, Jun 30, 2006 at 09:23:11AM +0100, Peter Corlett wrote:

> CDDB genre data is complete rubbish. Well, so is the rest of the  
> data, but the genres especially so.

Even ignoring that (which I do; I classify everything myself), it still
sucks that I can't easily put something into two or more genres.  For
example, Hayseed Dixie should go into bluegrass and into rock.  Putting
it into "bluegrass/rock" sucks.  It's like putting CSV into a database.

And using "contains" to build playlists will fail the moment you have a
genre which is a substring but not a superset of another genre.  I can't
think of any at the moment, but I bet they exist.

Never mind that you might want to search for rock songs played in a
bluegrass style but not the other way round.

-- 
David Cantrell, who wishes that Doompah music existed