Each adc-number file corresponds to one tape. In most cases there
are two such files, as adc00216.bcd and adc00216a.bcd
where the second file is longer than the first. This is because
most of these tapes were used many times for different sets of data,
with each successive set of data files overwriting some or all of
the older data from the beginning of the tape.

The first adc-number file for each tape consists of the last
data written, and may be considered to the the entire intended
contents of the tape. All of these were fully recovered and
are error-free.

The second adc-number-a (a for "all") file consists of all data
recovered from the tape (up to EOT, a test pattern or a very long
blank section - indicating the furthest end of data ever written).
The first part of each "a" file matches the other file for the tape.
The rest is old data left on the tape, and typically contains
unrecovereable errors where old data first emerges from the final
erasure at the end of newer data, and possibly other errors I did
not trouble to clean up.

----


My regular file format for 7-track tapes is one character per byte, with a filename extension of ".bcd".
I'm now using this convention whether the data is BCD (even parity) or binary (odd parity.)
In this format each character from the tape takes up one byte in the file, stored in the low 6 bits.
The 7th bit is the parity, same as on the tape, and the 8th bit indicates the first character of a tape record.
A file mark is indicated by a record consisting of the single byte hex 8F.