[nflug] Off Topic: Building a digital music collection

[Joe]

As this is somewhat off topic, replies might be better off list.

I'm starting to build a digital music collection on Linux - mostly from
p2p sources (and CD's I own).  I'm looking for strategies for organizing
and storing it - with pro's and con's.

* Any good books / tutorials on this that are relevant to doing it on Linux

* How do you structure your collections on disk (file hierarchies and
file names, etc.) *and* why

* Why does Amarok say tagging of mp3 files using musicbrainz is not
supported (is it one of those proprietary formats issues)?  Picard says
it can do it (as well as for wma's and m4a's!)

* I'm on a multi-boot notebook with a 100GB drive and I'm already
starting to run out of room.  (I have around 70GB tied up with a Windoze
partition and an old Linux (dapper) partition I haven't gotten rid of
yet).  If I move the music out to a USB drive, then what do I do when
it's not connected?  Are there any tricks to reduce redundant work when
any new files collected are later transfered to the USB drive?  Are
there any problems with the p2p software when the USB drive is disconnected?

What I'm doing now:

So far, I have two main directories set up, one that is shared (p2p) and
one that is private (for my CD's, etc.)

All new p2p tracks get stored in the shared folder.  Once they've been
"normalized", I put them in subdirectories of the shared folder roughly
by genre.  These directories are currently flat (no subdirectories).

By "normalized", I mean hand editing the tags in Amarok and then
renaming each track in the format Artist - Album - Track - Title.mp3  .
(So far, I'm converting any m4a's and wma's to mp3's so Amarok can edit
the tags and so the unwashed can play them if they download them from me.)

For p2p, I'm using Limewire (free version).  I tried gnutella, but the
user interface is overly complex, cluttered, and I couldn't find good
explanations for what a number of the features did.

I just installed Picard (musicbrainz collection manager / tagger),

http://musicbrainz.org/doc/PicardTagger

but I haven't tried it yet.  By default, it wants to fix all the tags
for me from the web and then store the tracks in a hierarchy of
Shared-Folder/Artist/Album with tracks renamed to Track - Title.  I'm a
bit ambivalent about this.  It makes the track names simpler, but when a
track is moved, the rest of the info about it is in the tags only and is
harder to sort and search for using normal Linux tools instead of
specialized software such as Amarok or Picard, etc.

Also, I'm a little leery of having an application automatically (I know
it can be set to manual) grab files, rename them and the squirrel them
away somewhere that might be unexpected and hard to find later.

Any feedback would be appreciated.  I'm sure some of you have been
managing your music on computers for quite some time and have already
gotten past the initial learning curve.

Thanks.

Joe