In article ,
Larry Blanchard wrote:
On Tue, 13 Apr 2010 15:54:58 -0500, Robert Bonomi wrote:

It took a while to establish cause-and-effect, because *nobody* was
willing to believe that that 'innocent little job" -- nothing more than
4 standard statements in the system control language -- could _possibly_
be the culprit. Until they ran it as the _only_ job in the system, and
watched the machine crash.

The entire job consisted of:
1) request a tape mount
2) copy a file from disk to the tape
3) rewind the tape
4) copy from the tape back to a new file.

The original Univac I had tape drives that maintained tension with
springs and pulleys rather than vacuum columns. Those drives, as well as
the vacuum column model that replaced them on Univac II, could read both
forwards and backwards. There were 10 drives.

One of the programmers (no, not me) wrote a program that issued a write
command, followed by a read backwards, followed by a skip a block. That
sequence apparently exceeded the response of the strings and pulleys and
they wound up in a heap at the bottom of the drive.

The resident (yep, 24 hours a day) CEs wouldn't believe him when he
described the problem. So he wrote a little program to demonstrate the
problem, called in the CEs, and ran the program, dumping *all 10* tape
drives. He wasn't very popular with the CEs after that, but when he told
them he had a problem, they listened :-).

I sometimes think all us old computer nerds should start a website and
record all these stories before we all die and the stories are lost.

There is also the story about a university student who got the engineering
plans for an IBM mainframe disk drive (one of the washing-machine-size units),
_carefully_ calculated the mass involved, and wrote a channel program that
consisted of 'seek to outermost track', pause, 'seek to innermost track',
pause, and "repeat indefinitely". The 'pause' times were carefully calculated
to the 'resonant frequency of the drive unit. Reportedly, the unit 'walked'
almost *THREE*FEET* across the floor, _towards_the_operator_, before they
managed to find and kill the offending task. I'm given to understand that
the operators were 'a bit nervous' for some days thereafter.