I can clearly remember recommending that new systems being built 10 years before Y2K begin utilizing 4 digit years. The counter argument at the time was that switching from 2 to 4 digits would require changing policy and that changing policy would require acknowledging that the current policy was wrong. Changing policy would mean that all pre-existing systems were not in compliance. Much better to wait until the Y2K situation was a generally acknowledged crisis and could be attacked on an emergency basis (no time then for sterile finger pointing exercises to assign blame). Of course, it costs a lot more to do things right on the second try; but its not your own money and there is a lot of no-questions-asked paid overtime available.
-- iws
(To reply, click here)
In my experience systems built around 1990 weren't the major problem (though there were of course some that still incorporated 2 digit years).
The biggest problem were computer programs written decades before (mostly in the '70s), that no one realistically believed would still be in use 25 years later at all, much less at the scale they stayed in use. And as it turned out, the oldest systems were often the most critical ones, e.g. utility control programs, air traffic monitoring systems, etc., because the costs to develop new ones were high due to the amount of testing required.
I think the topic is an interesting one, but ultimately unanswerable. I'll be interested to see if Farhad comes up with any good insights, but I really don't expect any major revelations.
-- racerx
(To reply, click here)
American Radio Works (the best damned documentary series around) covered this ground very nicely back in 2005.
-- Olaf
(To reply, click here)
Is it technically accurate to call it the Y2K "bug?" I think of a bug as an accidental error contained within a program. When people coded only the last two digits of the year in software years ago, they very well knew the limitation they were putting in. The thinking was the software was probably going to be superseded or updated before that limitation would manifest.
It's sort of like how they changed now when daylight savings time happens. Some older devices still use the old calculations to switch the times. I don't know if I'd call those bugs.
-- b_r
(To reply, click here)
What did you think of this article?
Join The Fray: Our Reader Discussion Forum
PRINT
E-MAIL
RECOMMEND...
RSS
Facebook
MySpace
Mixx
Digg
Reddit
del.icio.us
Furl
Ma.gnolia
Sphere
StumbleUpon
Sarah Palin's Brand of Populism Is Dangerous and Deceptive
New York Times Amazed To Find Jews in Montana
All Your Nagging Amanda Knox Questions Answered
No One Writes a Great Letter Anymore
New Evidence That Buying Green Products Can Make You a Worse Person
Could a Tea Party Candidate Actually Win an Election?
I can clearly remember recommending that new systems being built 10 years before Y2K begin utilizing 4 digit years. The counter argument at the time was that switching from 2 to 4 digits would require changing policy and that changing policy would require acknowledging that the current policy was wrong. Changing policy would mean that all pre-existing systems were not in compliance. Much better to wait until the Y2K situation was a generally acknowledged crisis and could be attacked on an emergency basis (no time then for sterile finger pointing exercises to assign blame). Of course, it costs a lot more to do things right on the second try; but its not your own money and there is a lot of no-questions-asked paid overtime available.
-- iws
(To reply, click here)
In my experience systems built around 1990 weren't the major problem (though there were of course some that still incorporated 2 digit years).
The biggest problem were computer programs written decades before (mostly in the '70s), that no one realistically believed would still be in use 25 years later at all, much less at the scale they stayed in use. And as it turned out, the oldest systems were often the most critical ones, e.g. utility control programs, air traffic monitoring systems, etc., because the costs to develop new ones were high due to the amount of testing required.
I think the topic is an interesting one, but ultimately unanswerable. I'll be interested to see if Farhad comes up with any good insights, but I really don't expect any major revelations.
-- racerx
(To reply, click here)
American Radio Works (the best damned documentary series around) covered this ground very nicely back in 2005.
-- Olaf
(To reply, click here)
Is it technically accurate to call it the Y2K "bug?" I think of a bug as an accidental error contained within a program. When people coded only the last two digits of the year in software years ago, they very well knew the limitation they were putting in. The thinking was the software was probably going to be superseded or updated before that limitation would manifest.
It's sort of like how they changed now when daylight savings time happens. Some older devices still use the old calculations to switch the times. I don't know if I'd call those bugs.
-- b_r
(To reply, click here)