David G. Wiseman

'cal 9 1752' explained

In <1992Jan30.205923.19385@helios.physics.utoronto.ca>
xiaoy@bullet.ecf.utoronto.ca (XIAO  Yan) writes:

>   September 1752
> S  M Tu  W Th  F  S
>       1  2 14 15 16
>17 18 19 20 21 22 23
>24 25 26 27 28 29 30

The guy that originally wrote the "cal" command on some old Version 7
machine had an off-by-one error in his code.  This showed up as some
erroneous output when a malloc'd variable overwrote 12 extra bytes with
zeroes, thus leading to the strange calendar output seen above.

Now, nobody in his right mind really cares about the calendar for
September 1752.  Even the *idea* of the year 1752 does not exist under
UNIX, because time did not begin for UNIX until early 1970.  As a result,
nobody even knew that "cal" had this error until much later.  By then
there were thousands of copies of "cal" floating around, many of them
binary-only.  It was too late to fix them all.

So in mid-1975, some high-level AT&T officials met with the Pope, and
came to an agreement.  The calendar was retroactively changed to bring
September 1752 in line with UNIX reality.  Since the calendar was changed
by counting *backwards* from September 14, 1752, none of the dates after
that were affected.  The dates before that were all moved by 12 days.
They also fixed the man page for "cal" to document the bug as a feature.

The 11 days from September 3 to September 13 were simply gone from the
records.  They searched the history books and found that fortunately
nothing of much significance happened during those 11 days.

Overall, this whole incident was pretty much a non-event.  One science
fiction author later heard about it, and blew the thing up into a
full-length work of science-fiction called "The Lathe of Heaven", a book
that in my opinion bears little resemblence to what really happened.
Rahul Dhesi 
UUCP:  oliveb!cirrusl!dhesi

Ha, ha, ha. Take me back to [ the alphabetic list ] [ the date-ordered list ].