[inforoots] Recovering data from old media (Was: "The digital Dark Age")

[John C Green Jr]

At 07:35 PM 9/27/05, Mike Albaugh wrote:
>Texx wrote:
>
> > Our own musuem collection contains boxes of mag tape that we havent 
> been able
> > to read yet.  Alas Im sure some of it has succumed to bit rit.
>
>  Perhaps, depenidng on storage conditions, but the older stuff
>was 200 or 556 BPI, and as long as its on "decent" media,
>it may surprise you. Paul Pierce was able to recover all but
>a few blocks of a tape with the sources for a CDC 3600
>Timesharing O.S. from a 7-track tape for me, about two years
>back. And since then, he has revamped and improved his
>read electronics.

A couple or three decades ago Datamation had an article about
a Navy guy who found a room of old tapes that had never been
intended to be archived e.g. had not been slowly wound end to
end in both directions to relieve tension and minimize print
through, rather had been high speed rewound.

He tried to recover data from tapes that were over a decade
old.  There was a mix of 200, 556, and 800 BPI.  As there
were mostly 556 BPI he only did them.

Results ...
Trying to read directly: Total failure
Cleaning with do-it-yourself tape cleaners available at the
time: Mostly failure
Sending out to commercial cleaners: Excellent results.

He had a table of the bit error rates by tape manufacturer.
One manufacturer had two grades: Normal and Millennium with
the premium supposed to last far longer than the normal.
The Normal had a much better bit error rate than the
Millennium under the conditions of his room of tapes i.e.
no special care having been taken to make the tapes readable
a decade later.

The vendor of the Normal and Millennium tapes: IBM.

Regards,
JC Green