On Fri, 1 Oct 2010 09:26:25 -0400, "Stormin Mormon"
wrote:
Were there other things that were going to fail?

As others have said, as of 1990 nearly everything would have failed in
2000. The percentage of programs which do not use dates in any way is
small. In 1970, Y2K was not even on the perceptual horizon for most
programmers, including me. (I wrote my first program in 1966, full
time starting in 1971.) A few businesses which project decades into
the future (such as life insurance) were mostly OK because they were
dealing with 21st century dates even in 1960.

In many cases, the errors would have been trivial, such as printing
1900 instead of 2000 on a report. Even those errors needed to be
corrected, though -- most programmers would not dream of releasing a
program into production which printed incorrect dates. If nothing
else, the result is that the users lose confidence in the program.

In a pretty good percentage of cases, programs would obviously crash
or give incorrect results in very serious situations.

The majority of cases, though, were the ones where analysis was
required even to determine what would happen and what fix was needed.
The actual fixes in most cases were much simpler than the analysis.
Software errors can ramify in complex and unexpected ways, so it was
essential to be thorough in the analysis.

So, your car would still have run -- the microcontroller there was not
date-dependent. Traffic signals might have worked, but many vary their
patterns on weekends -- they might have gotten the day of the week
wrong, or the software might have crashed and left all the lights
blinking. The Internet would have stayed up, but most web sites would
have gone down. And this is before we even start talking about
banking, all business systems (including point of sale), etc. Even
industrial control systems, airplane control systems, medical device
controllers, etc had the potential to fail -- many of these embedded
systems would have been OK, but nobody could tell until they were
analyzed. Again, Y2K just wasn't in the field of view when much of
this was written.

Realize that in general, even a minor error in a computer program can
make the entire program fail. It's as though omitting one nail from a
rafter might make an entire building collapse. Digital systems are far
more sensitive in this respect. Thus even minor date-related errors
had to be found and fixed because they had the potential to bring down
the entire edifice.

So, people who say "January 1, 2000 wasn't a big deal, why did we
spend all that money" are wrong. January 1, 2000 wasn't a big deal
*because* we spent all that money. Without the remediation effort, the
human world, the first world anyway, would pretty much have come to a
stop that day.

Edward