On 5/26/14 3:59 AM, Peter Coghlan wrote:
If the data
is obscure and there are no running or emulated systems which dealt with it
available, there may be no way of verifying that a correct and complete image
has been obtained.
If it is tape, there is a high probability there will be dropouts. If you can't
verify what you've read, you save the tape until you can, perhaps with better
data recovery equipment. If the tape is so bad physically that a second pass is
impractical, you do the best you can in tape prep (baking, retensioning) or if
it is a large batch, you weigh recovery probabilities with the value of the data
you're recovering. Some batches aren't worth it (accounting records with no
provenance) and some are (Doug Englebart's SRI lab backup tapes).
At least in the case of 1/2" magnetic data tape, there is at least parity and a
block check character, and established standards for on-tape block formats, so if
you can get a decent image of the flux levels and where they were on the tape you
have a pretty good idea if what you read was what was recorded even if you can't
decode it past the block level. The best luck I've had with tape recovery has been
with the self-clocking formats (1600/6250). The last generations of drives were quite
good. I wish the same were true for 800, and in 20+ years I've never had a working
solution for 7-track media.
What you don't know at that point is if what is on the tape or diskette matches
what the paper label says should be there. This is one of the reasons I read multiple
physical copies of what should be the same thing from different tapes/disks.
Copy-protection is another rathole. The best I can do is try to identify if something
is known to be copy-protected or not. This tends to be PC software, and there is no
shortage of hobbyists dealing in that space.