and if you had a Fortran/COBOL program, the last 7 columns were
reserved for a sequence number.
Actually: COBOL defined the first
six columns as the sequence number, and reserved the last eight for user use.
They had no meaning to the program. Often they were left blank; some
organizations used the to identify the program or program fragment the card
belonged to.
FORTRAN used all of the first 72 columns, and ignored the last
8. Sometimes, these eight were subprogram id; sometimes (as encouraged by the
Burroughs Large Systems compilers, for instance) it was sequence numbers.
The
reason that the early programming languages only used the first 72 card columns
is purely technological: the IBM systems of the middle 1950s read cards in "row
binary" format, the holes in each row (12-row, 11-row, 0-row, 1-row, thru 9-row)
were read into bits of words. The words were 36 bits, so the card appeared in
memory as [9-row, cc 1-36], [9-row, cc 37-72], [8-row, cc 1-36], 8-row, cc
37-72], ... [12-row, cc 1-36], [12-row, cc 37-72] . There was no place for the
bits for columns 73-80 to go.
Univac adopted the 72-column convention for
their "90-column" cards, for compatibility with IBM. --- --Bill. NAL:
question the answers, especially mine. [ Reply to This | Parent | # ]
|