I tested a 386 luggable once to see how it would handle Y2K. I set it for 23:59 12/31/99 and it worked just fine. Only problem was that it only accpeted two digits for the year. Any attempts to set the year from the date prompt would have resulted in 1900.
One thing you might notice from these posts is that, given the wide variety of situations and ways and means of “fixing” the Y2K problem, a number of other date related issues were recognized/created. Derleth, for example, touched on the 2038 issue earlier.
This site is a decent starting point for the technical details on everything from Y2K to 2038.
However, none of these dates have quite the same massive “appeal” as Y2K. Each issue affects a critical handfull of systems. And the media has already been burned once on Y2K, so none of these should get the same kind of press attention.
We wouldn’t want to burden the public with actual technical details after all. :rolleyes:
What? you can patent source code now? excuse me while i get the rights to
10 print “hello”
20 goto 10
“Y2K advocates were like newly hatched mosquitoes. They only have a short amount of time to suck blood before they die.” - Paul Kedrosky
Y2K was a non event for my company, Phelps Dodge, because we took steps to prepare for it. We still had a mainframe and spent major bucks moving the applications to other platforms ahead of time. A core of the IS department had to hunker down in the corporate office overnight to monitor our offices in Rotterdam and the UK and shut down systems before midnight. It was mostly a party and the only things that happened were a couple of smelter scales that printed tickets wrong. My only regret was not bringing an air mattress to put my sleeping bag on in my cubicle. I had considered building a sandbag bunker and investing in a belt fed weapon but was glad I didn’t need them. Not saying what was in my overnight bag 
My favorite Y2K joke was about Colt’s development in smart gun technology. There was an onforseen Y2K glitch and the weapon could only be used to shoot William McKinley.
Yes, a lot of PC’s CMOS only accept two digits for the year; newer CMOS sometime take four digit dates, and some computers use drivers to deal with the issue.
However, that’s only a small part of the problem. The system time and date is typically only used to get today’s date, and in general, software can make informed guesses about what century we’re in right now. Where the real Y2K issues came up is when software itself stored only 2 digits for the date, and made the assumption that the first two digits were always 19. On PCs, this was pretty atypical - most languages used on PCs have some sort of esoteric date/time storage mechanism that boils down to storing the number of milliseconds after January 1, 1900 or some such goofy format. Like other have pointed out, some databases stored only 2 digits for the year.
Speak for yourself matey, I didn’t fix diddly. I’ve been hiding under my desk since July 1999.
Is it safe to come out now?
:: looks at news site, to find myriad reports of war, terrorism, new viruses and economic collapse ::
Erm, is there any spare room under that desk of yours?
The General Accounting Office reported in report B-285601 that there have been 18 court cases that invoked the Y2K act. Hence there have been some problems, though not nearly as much as has been predicted.
I agree with the OP that one of the reasons that there was no disaster was because things were fixed beforehand. But from what I’ve seen people overestimated the amount of dependence on continually operating systems. I’ve worked for several years in IT in the petro-chemical industry and in finance. System failure happened quite regularly. The thing you did was sit down and try to figure things out. No real harm was done (except some angry bosses) if you didn’t manage to fix it in a day. Of course for the important systems (such as the plant controllers) it was different, but those naturally did not have any reference to date. From what I’ve heard anecdotally there may have been a number of companies that suffered minor glitches; these were quietly fixed without any external effects visible.
DO NOT ATTEMPT TO ENTER THE SPACE UNDER MY DESK, I AM ARMED AND PREPARED TO DEFEND MYSELF. I HAVE NO SPARE FOOD OR WATER TO GIVE YOU.
I work for a large software company, partially doing mainframe support. It was a big deal for us for most of 1999 – when we were fixing the legacy crap. It’s not like we were stupidly waiting around doing nothing, as the press sometimes implied. OTOH, 90% of the problems were display issues, where a date would’ve been 1/1/100 or 1/1/10 instead of 1/1/00. In our case not much mission critical stuff was ever that dependent on dates. It also helped that most databases enforce four digit years, so the data was stored properly.
Although I was on-call for that night, just in case.
I worked as a Y2K software tester and still work in QA/CM for an IT shop working with government. We renovated code and through the renovation found many Y2K problems in code, enough to have created significant problems and interruptions of service to government beneficiaries. That was just in-house code. I can also confirm that at least one software package by a major vendor was non-Y2K compliant (even though their website said it had been ‘certified’ Y2K compliant and we had a letter from them stating same). I was here at work when the date rolled and the software absolutely quit working.
Minor nitpick: COBOL wasn’t and never has been the problem. Many languages were succeptible to the problem. It all has to do with how the code was written. The problem was most visible with COBOL because it was one of the oldest languages in really widescale use and many of the programs were written non-Y2K compliant with the thinking “ah, surely they will convert this to something else by then”. I am aware of programs that have been running for almost 30 years. The last compile date was 1974. How many people reading this on a Windows machine think W2K or Win 95/98 will still be in use in 2025? Ya don’t think so?.. neither did they, the mentality is still around.
The problem sort of solved itself - storage is cheap these days so the extra two bytes are not going to kill you. If 5MB hard drives cost $5,000 we might still be dealing with this issue.
With respect to language and storage:
I’m still surprised that this bogus idea of storage shortage was believed to such an extent. When I was programming it wasn’t storage in terms of bytes that was the problem, it was other limits, such as an 8-character file name limit. If you wanted to put the date in the file name, you quickly ran out of characters if you wanted to have a Y2K-compliant filename. So you needed to have a conversion routine of your own.
Another reason I’ve personally coded programs in the '90s that were not Y2K-compliant was those stupid IBM date subroutines that up in the late '90s were not Y2K-compliant. I’m talking RPG on an AS/400. How the heck are you supposed to make a Y2K compliant program if the standard subroutine you have to use only allows 2 characters for the date?
If those kind of limits wouldn’t have existed I’d always have programmed Y2K-compliant as a matter of course. Unfortunately, these limits did exist.
I’m guessing you are young enough that you have never dropped a deck of cards and had to re-index them by hand? Go and find yourself some crusty old Mainframe guy who worked in a Data Processing shop back in the 60’s, they will tell you have every character was sacred in both processing and DASD terms. I for sure wasn’t around but I’ve had enough conversations with enough people who were around when the business was in its infancy to believe that it was an issue.
I tried to look up the cost of MF storage but don’t get any good hits. I know what our client pays for storage but I’m having trouble converting that into useful figures because I’m not sure of how the devices are configured. If my numbers are right I get about $2000 per gig of storage per month but somehow I think that is way too high, probably by a factor of 2 or 4.
Anybody know what a new 3390 cost these days?
Personally, I’m pissed that you all didn’t fix the Y10K bug at the same time. I know you all don’t believe we’ll still be using the same legacy systems in 10,000 AD, but no one expected to be using them in 2000 AD either. Would it have killed you to add an extra digit to the date field? It would be nice to know that after civilization collapses and thousands of years later savages dig up our databases they won’t crash when reaching 10,000 AD. When Og the barbarian boots up an ancient laptop, he’s going to be pretty disappointed that it doesn’t work, all because some programmer wasn’t farthinking enough.
Am I the only one who pictures some sort of violent soccer/rugby match involving laptops and barbarians? 
Og’s got the laptop… he shoots… he scores!
I started writing a parody song in the early days of 2000 about Y2K, but I never got beyond the beginning of it, and after a while it seemed pointless to finish it. Since you folks have heroically brought this topic back from the dead, I thought I’d let you see what I have so far.
Y2K Guy
(to the tune of “American Pie” - duh)
*Long, long time ago…
Way back last November,
On the news I saw a talking head
And he said that there was a chance
We’d wake from our collective trance
And all our home computers would be dead…
His tale of terror made me shudder,
We were a boat without a rudder…
I shouldn’t wait 'til later
To buy a generator…
The food supplies would soon be gone,
So stockpile Spam and Evian…
But New Year’s came, and with the dawn
We knew…the hype…had died.
(chorus)
So bye bye, newsroom Y2K guy,
We’ve heard hype of every stripe but now it really won’t fly.
You good old boys have tried to scare us - but why?
Could it be to send your ratings sky-high?
Did you want your ratings sky-high…*
Some people seemed to be worried about the Y2000K bug back in 1999. Now THAT’S looking ahead!
Revtim, I think I got that spam!