For the sake of argument... let's say you happen to have access to one of the Pioneer DRM-7000 cdrom/dvd jukeboxes and you are interested in using it to archive data, scientific data with reliably parsable header information describing the contents of each file. For the sake of argument.. lets just say that those files look a lot like ogg vorbis files. They aren't but I think everyone can imagine a 720 dvd scsi auto-changer full of music.

How would you go about indexing the contents of all the files on all the cds/dvds in the jukebox so that you could efficiently locate which disks had the files you wanted and then could direct the auto-changer apperatus to load those disks so the files could be retrieved.

I understand stand how to do the mechanism of the actually auto-changer operation. mtx supports this beastie, from an operations point of view. But... operating it is only part of the problem. I need to be able to build an index that maps file header information to physical cd in the library. is there any framework out there that exists for this sort of thing? Or am I basically going to end up writing some python glue around a sqlite db that I can then query via a homebuilt interface? I'm looking to leverage as much pre-existing code as I can.

And yes... i realize that in the final analysis just setting up a 2 terabyte raid of sata disks is probably more cost effective... but where's the fun in that. This thing makes some kickass noises when its moving its robotic arm around changing media...and if I can change out some of the paneling with see-through plexiglass and pimp it out with internal lighting...it will be super hot.

-jef