Disk archival techniques

On Tue, 17 May 2005, Al Kossow wrote: I was attempting to be faintly poetic; I meant only that the meta-data doesn't need to be overdetermined. For a floppy archive, for example, a tabular table would contain most of it, with comments on oddities. Stored on a computer as a text file, for example, but I don't totally discount paper copies. Here we slightly diverge in that I would like to see them separate but highly correlatable. I understand the desire to have the data and metadata in one place, but all of the mechanisms to do so are ephemeral. Tar, maybe less so. And even with that, non-technical organization techniques used in a consistent way, eg. a TAR with a directory containing two files (dumb example) DATA and METADATA, and the TAR file having the same name as the directory inside it, that sort of excessive, trivial and brutal redundancy. Ugh. But this is an excellent example: it's not a problem of portability of data or metadata, really. But if the tapes could magically be turned into character streams [in the originals leading and trailing NULs were generally ignored] I imagine the tapes are delicate, cracked/torn, etc) and the the index number recorded as metadata (and maybe a photo of the tape leader, we all know how crucial those hand-written notes on tape leaders are!!) you at least have staved off lossage. Assuming the new data can be read in 10, 20 years... And the scribbles-on-tape-leader thing points out the futility of nerdly computer approaches to archiving; it's not a computer problem. Teledisk version 99 or FutureIncompatibleData won't solve obtaining and preserving data that's not electronically readable in the first place. On my 100 or so 8" floppies the sticky, fading inked label contains utterly crucial data, and has to be hand transcribed. As a(n increasingly ex-) nerd, I increasingly run into the fact that few real problems are solved with computers.