OTOH, it would be gross overkill to store the complete
bitstream
without clock separation of every track on disks that could be
defined as "5.25 DSDD 48TPI WD/IBM 5SPT 1024BPS, no known oddities",
just because SOME (a few thousand) disk formats DO have strange
things to deal with.
And herein lies the core of the electronic/data part of the
problem. Everything is perfectly orthogonal, except for the
exceptions. And there are so many exceptions as to rule the
roost.
The implication in this thread is there is some single
computer-based solution to solving the world's data loss problem.
It's silly on it's face. Look how hard it is with the black square
things with the rotating magnetic mylar inside, even of just one
size.
I think that the idea of always saving all data is naive. Even
within narrow bounds, say electronicizing printed manuals, there
are all sorts of compromises and data losses (resolution, color,
texture, ad nauseum) but the usage is often nerdly, not cultural,
eg. I wanna know what IC17 is and will interpolate between the
moire pattern artifacts to find out. The human eyeball/brain is
fixing the errors.
I predict that there will be no single solution to the current
crop of problems, which will expand with time.
Funny, I was making this argument in an automobile list, where
there's a similar implied goal of a magic fuel, motor, etc that
will "solve transportation problems". I think it's clear that
transportation is a cultural system, involving cars, trains,
public trans, urban planning, walkable neighborhoods, etc. One
size does not fit all. The problem has to be handled case by case.
While certainly lots of sharing is great, it might be that (made
up examples) Cromemco CDOS archiving is different from Cromemco
CP/M and other CP/Ms though there is overlap, etc and it's all
certainly different than archiving 1/2" tape, or paper tapes, or
disk packs, or EPROMs. A unified system seems like a 1890
solution.
Give a small boy a hammer, and everything looks like a nail.