id3, playlists and DB

[Ivan Sergio Borgonovo]

I'm currently using amarok 1.4.8 (sid) and while I think it is one of
the best players around with a lot of interesting features etc...
that makes me even more greedy ;)

I'd like to understand the logic behind id3, the DB and playlists.

I've a centralised postgres DB and an nfs share that I share across
people/client at home.

Currently I generally:
- put new stuff in the collection folder
- add new stuff to playlist
- edit "proprieties" eg: genre, artists

a) stuff is hard to relocate
if I decide that an album/artist have to be moved in a "genre"
directory, the DB is not updated, and I lose scores and edited "tags"
I could use the manage file, but that is not always viable if all the
tags haven't been fixed [1]
b) tags don't get updated in the file, but just in the DB
c) if I access the same collection from other workstations it is a
PITA to see changes to tags.
  1) delete all entries in playlist
  2) rescan collection
  3) add all collection to playlist
when collections are huge it is a very slow process and it is prone
to freeze/crash

Maybe I'm just using the "wrong" approach, maybe there is something
that can be improved.

thx


BTW
There are some problem of DB sync when you update tags.
Sometimes
Edit 'XXXX' for Selected Tracks
just update *some* of them.
You see "Writing..." some get written with the new value few don't
get updated or pick up a "random" value.

[1] there are too many path/url columns in the DB to update them by
hand through SQL even if I did a global transfer of my collection on a
different path, this is not a viable method if you've to move just
bunch of your collection around.

-- 
Ivan Sergio Borgonovo
http://www.webthatworks.it