Xref: utzoo rec.humor:19088 rec.humor.d:1663 comp.misc:5190
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!amdahl!pacbell!ptsfa!dmt
From: dmt@ptsfa.PacBell.COM (Dave Turner)
Newsgroups: rec.humor,rec.humor.d,comp.misc
Subject: Re: Looking for Computer Folklore
Message-ID: <4689@ptsfa.PacBell.COM>
Date: 18 Feb 89 01:53:50 GMT
References: <7143@pyr.gatech.EDU> <4744@sfsup.UUCP> <2887@sybase.sybase.com> <1912I78BC@CUNYVM> <1036@tutor.tut.fi> <6761@pogo.GPID.TEK.COM> <557@rpi.edu> <6321@saturn.ucsc.edu> <83525@felix.UUCP>
Reply-To: dmt@ptsfa.PacBell.COM (Dave Turner)
Organization: Pacific * Bell, San Ramon, CA
Lines: 34

Whenever we used to make major changes to our operating system or transaction
processing system we were required to repeat a prior day's business to prove
the the system was ready for production.

Until about 10 years ago, we would do this by copying all the databases and
tapes for a day and run a series of tests on Saturdays. All the production
terminal operators would be at their terminals typing exactly the same things
that they had typed on the day being repeated.

All this was very expensive and error prone. Usually the tests would cause
a crash a few minutes after they started.

On one memorable day in 1976 the test was running very smoothly.
The computer room was filled with onlookers: operations people, systems
programmers, bigshots, vendor representatives, etc.

The console operator was continuously displaying the status of the system.
One common command was to display all the jobs in the system:

	$dj 1-999

Everyone was pleased that the test was going so well until around
4 PM when all the jobs suddenly stopped running.

Concern turned to elation when the console operator confessed that he
had mistakenly typed:

	$cj 1-999

Which *cancelled* all the jobs in the system!


-- 
Dave Turner	415/542-1299	{att,bellcore,sun,ames,pyramid}!pacbell!dmt