Preamble

Since you've become involved with computers and digital audio, your actual CD-collection has become a burden. Per some arguments, you want to keep the pristine originals (e.g. because you know your backup methods suck), but then again -- it all just takes up space.
This howto briefly -- but adequately -- covers how to easily set up ``abcde'' to allow you to semi-automatedly backup your CD collection to the Free Lossless Audio Codec ``FLAC''.

The custom configuration file

Put the following into ~/my-cd-backupping-abcde.conf:
FLACOPTS="--best"
ACTIONS=cddb,move,playlist,clean
OUTPUTTYPE=flac
OUTPUTFORMAT='${ARTISTFILE} - ${ALBUMFILE}/${TRACKNUM}.${TRACKFILE}'
VAOUTPUTFORMAT='Various-${ALBUMFILE} - ${TRACKNUM}.${ARTISTFILE}-${TRACKFILE}'
PLAYLISTFORMAT='${ARTISTFILE} - ${ALBUMFILE}.m3u'
VAPLAYLISTFORMAT='${ARTISTFILE} - ${ALBUMFILE}.m3u'
EJECTCD=n
mungefilename ()
{
	echo "$@" | tr \*\/ +_ |  tr -d \[:cntrl:\]
}
This snippet will:

The call

If you're very trusting of your CDDB-provider -- which I am not -- you could add a -N to obtain non-interactive mode. I however, prefer to be very careful to edit the tracklistings into place before ripping.
while true; do abcde -c ~/my-cd-backupping-abcde.conf && eject; sleep 5; done
This call should run indefinitely (use Ctrl-C to interrupt it), trying to rip the disk in the drive every five seconds.
The reason that abcde shouldn't eject the CD from itself is that it would do so before program termination -- as a matter of fact, a lot of post-processing (compression, mostly) may be pending at that time. By ejecting from bash, we're using the sound of an ejecting drive to prompt us to insert the next disk.
While my system has some trouble keeping up the --best-compression while ripping as fast as possible, your's may do well. (This one has 2000 bogomips). However, you could use --fast at this point, returning later to do a completely noninteractive re-compression of the whole of your .flac-collection. However, that would require some scripting. But note that as long as metadata is preserved, nothing is lost in the recompression -- it's not a lossless codec for nothing.
Remember that this process is highly paralellizable and it would be beneficial if you used every machine you could get a hold of in the process. I'm currently ripping on two machines in parallell, but I also only have a modest CD collection.

If you have any suggestions, corrections or complaints, I'm reachable at skrewz@skrewz.dk.