Slashdot has a thread on What Was Your Worst Computer Accident?. I have had a few disasters myself – but my real trial by fire came in 1986, in front of 250 of my colleagues.
I was 25 years old, recently graduated, and working in the IT department of the Norwegian School of Management, where I had also been a student. The school’s president had decided to haul the institution, kicking and screaming, into the modern world of computers and electronic communication – so we had installed PROFS, an IBM email and calendaring system, on our mainframe computer (a 4381 running VM/CMS, in case you’re interested.)
We then arranged an end-of-semester all-employee meeting in the grand auditorium, with 250 of the faculty and administration – colleagues and former professors of mine. My task was to convince this rather skeptical audience that email was the wave of the future, demonstrate how easy it was to use and how it would simplify your life. We had rigged an unpredictable and expensive projector, and I demonstrated how to log in, how to start PROFS, how to enter appointments into your calendar, and for the grand finale: How to arrange a meeting. I selected seven people (IT department colleagues with updated calendars) and a meeting room, asked the system to find a suitable time over the next ten days, picked one of the available times, wrote a short invitation, and pressed the F12 button to send the invitation and book the meeting.
The screen went blank, then a CP message (from deep in the bowels of the system) with incomprehensible error codes appeared. Our system programmer, who was sitting in the back of the auditorium, turned white, stood up and ran out of the room. And there was no response from the machine.
It turned out that I had found an undocumented error in PROFS – when choosing 7 people, for 10 days ahead (and some other conditions I don’t remember), PROFS would crash your virtual machine, in a way that had to be recovered from the system console. (This, incidentally, was very hard to do – VM/CMS was extraordinarily stable.)
Of course, I had to do this in front of all my colleagues.
We somehow recuperated, and PROFS became, over the next 2-3 years, a common tool (though, given its poor interface, nothing like the email we have today.) I have since had many instances of the “demo effect” (things breaking when you demo them), but this was the first major one. It inoculated me against technical embarrassment – things just can’t get much worse that bombing in front of the whole company.