Hector Martin

Agreed - though, I think, given a format with "complete" data it should be possible to build tools to extract a subset and dump it elsewhere as a different format.

I pitched an IFF style one a while back but after thinking about it, this is actually very bad for streaming and it should be an interleaved format (at least).

So the question is, do you want to go all out and design a format that can be used on cheaper microcontrollers or FPGAs to implement optical disc emulators that interface at an analog level? Or stick to archiving/emulators leave that as extra challenge for someone to do later?

I feel that optical disc emulators of that sort would have a implementation-specific format due to the performance constraints; and trying to constrain a format intended for archival to those limitations would be unnecessarily restrictive.

I think it's important to recognize that the purpose of archival is not to perfectly replicate the original, but to perfectly replicate the *useful information* encoded in the original. I.e. there is no point in storing voxels of atoms.

In that sense, it probably makes more sense to have a standard but *extensible* format that can encode all useful data at a common abstraction level for optical media *and* platform-specific metadata.