Y2K
December 29, 2019 8:52 AM   Subscribe

‘Here We Go. The Chaos Is Starting’: An Oral History of Y2K - Once upon a time, we all thought the world was going to end on January 1, 2000. Two decades after the panic of the century, it’s time to finally hear from the people who spent years—and billions of dollars—making sure it didn’t.
posted by the man of twists and turns (76 comments total) 36 users marked this as a favorite
 
Joanna : So, where do you work, Peter?
Peter Gibbons : Initech.
Joanna : In... yeah, what do you do there?
Peter Gibbons : I sit in a cubicle and I update bank software for the 2000 switch.
Joanna : What's that?
Peter Gibbons : Well see, they wrote all this bank software, and, uh, to save space, they used two digits for the date instead of four. So, like, 98 instead of 1998? Uh, so I go through these thousands of lines of code and, uh... it doesn't really matter. I uh, I don't like my job, and, uh, I don't think I'm gonna go anymore.
I will forever think of this scene.
posted by Fizz at 8:57 AM on December 29, 2019 [32 favorites]


I’m currently getting lots of pings from a logging company to make sure I don’t have any two digit years in my logs, since apparently their fix was “A two digit year less then 20 is 20XX, 20 and above is 19XX” so now there’s a 2020 issue as those quick fixes that they’d have plenty of time to get around to fixing are hitting their end of life.
posted by mrzarquon at 9:05 AM on December 29, 2019 [45 favorites]


Happy Y2K20!
posted by fairmettle at 9:05 AM on December 29, 2019 [8 favorites]




so now there’s a 2020 issue as those quick fixes that they’d have plenty of time to get around to fixing are hitting their end of life.

*kicks can down the road*
posted by Fizz at 9:11 AM on December 29, 2019 [13 favorites]


(Also should clarify I don’t think it’s this vendors specific problem, more that they have customers who are still using systems with two digit years they’ve had to invest data from. That’s the worrying part to me.)
posted by mrzarquon at 9:14 AM on December 29, 2019


See also the 2-gigasecond unix clock rollovers in the 2030s.
posted by GCU Sweet and Full of Grace at 9:19 AM on December 29, 2019 [16 favorites]


I feel weirdly proud to be the only person I know who was affected by the millennium bug. I went to Journeys to buy a cool new pair of Doc Marten wingtips with my Christmas money, but when they ran my mom's card, it wouldn't go through. The cashier said they'd been having glitches all that day: Jan 1st. Never did get the Docs.
posted by erinfern at 9:25 AM on December 29, 2019 [16 favorites]


This is nothing, imagine when we switched from BCE to CE.

"Whaddaya mean the years go UP now? Everyone knows years get smaller!"
posted by oulipian at 9:26 AM on December 29, 2019 [60 favorites]


hey’d have plenty of time to get around to fixing are hitting their end of life.. And that is how Splunk got hit with a Y2K bug despite being founded in 2003.
posted by ocschwar at 9:30 AM on December 29, 2019 [5 favorites]


I learned my lesson, and use a 5-digit year now. Hope you all have a good 02020.
posted by pipeski at 9:35 AM on December 29, 2019 [27 favorites]


Once upon a time I worked in TV commercial production. I remember maybe a handful of the actual spots. This Polaroid Y2K is one of them. The future did not turn out as planned.
posted by dismitree at 9:37 AM on December 29, 2019 [4 favorites]


I have the fix to Y2K20: let's all set our calendars back to 2016 and then let President Clinton solve it.
posted by dances_with_sneetches at 9:37 AM on December 29, 2019 [16 favorites]


Clifford Stoll does not come off well in this article; he’s a retroactive Pollyanna throughout the whole thing: “it was obvious that no such apocalypse would happen. Computer time rollover just wasn’t an important issue. Nothing serious happened.” I recall he was similarly wrong about the viability of e-commerce in the 90s.
posted by migurski at 9:43 AM on December 29, 2019 [23 favorites]


I know at my dot com we had an intranet project for power plants across the country, where if machines on the east coast (or globally?) crapped out people in later time zones would have advance warning of the models/generators/servers involved. I was only a front end interface guy (and fuck that late 90s browser compatibility bullshit,) so I can't recall the scope of the thing, except that in the end nothing failed and everything went ok, and the site we built worked.
posted by Catblack at 9:48 AM on December 29, 2019 [1 favorite]


I remember working at a palce that used CADOL and the old-timer at the time said that some of the POS terminals that ran the older version of CADOL wouldn't work after 2000. Our version was updated but we still had to be
on-call even though we were the billing end of home health care and not the clinical end.

Spoiler alert: nothing happening
posted by 922257033c4a0f3cecdbd819a46d626999d1af4a at 9:49 AM on December 29, 2019


See also the 2-gigasecond unix clock rollovers in the 2030s.

2038 and already fixed.
posted by 922257033c4a0f3cecdbd819a46d626999d1af4a at 9:50 AM on December 29, 2019 [3 favorites]


I spent much of 1999 flying between North Sea oil rigs doing software and hardware Y2k updates - I could make more doing that than geology. A weird but lucrative year.
posted by thatwhichfalls at 9:55 AM on December 29, 2019 [11 favorites]


“Nothing serious happened“, because of the aforementioned billions of dollars spent to make sure nothing happened.

The father of a high school friend (I graduated in ‘99) was a retired COBOL programmer who got back in for one more “score”, and he bought himself two vacation homes out of Y2K. I think he was billing $300/hr towards the end.
posted by sideshow at 10:10 AM on December 29, 2019 [36 favorites]


so now there’s a 2020 issue as those quick fixes that they’d have plenty of time to get around to fixing are hitting their end of life.


*worried about y2karl now*
posted by Halloween Jack at 10:21 AM on December 29, 2019 [10 favorites]


2038 and already fixed.

How can you claim it is already fixed? Sure they patched the OS, but that doesn't fix all of the software that uses 32-bit integers or databases storing 32-bit integers.

And it doesn't have to wait until 2038. There are lots of programs that do computations projecting 20 or more years into the future. Your home mortgage for example.
posted by JackFlash at 11:00 AM on December 29, 2019 [1 favorite]


I started banking at my current bank in 2002.

I remember, around 2004, seeing a bank employee's screen, which said that I had been a customer since 19102.
posted by Hatashran at 11:15 AM on December 29, 2019 [14 favorites]


At my first programming job in 1989, one of the first code changes I made was changing a flat file format, and several mainframe COBOL programs that used the file, to expand a SINGLE-DIGIT year field.

In the first of several faux-pas I made at that job, I laughed to my boss about the foolishness of the programmer who'd made the decision at the beginning of the 80's to only use a single digit. Of course, my boss was the original programmer.

I suggested that we expand the date field to four digits, but he nixed that, and told me to use two, and that "we'll worry about the year 2000 or whatever" later on. He really hated to expand the sizes of fixed-width records.

He also hated my attempts at using "structured programming" techniques in COBOL, and would sometimes go behind my back and add GOTOs everywhere.
posted by JeffL at 11:16 AM on December 29, 2019 [20 favorites]


Use of GOTO Considered Hilarious!
posted by thelonius at 11:29 AM on December 29, 2019 [3 favorites]


It wasn't an issue because we decided to take steps to make sure it didn't become a huge problem. And a bunch of people worked their hinders off to do the work.

If only there were some other pressing issue that could warrant similar attention. Oh well, it's just where we live, I'm sure it will work itself out.
posted by fifteen schnitzengruben is my limit at 11:51 AM on December 29, 2019 [12 favorites]


Fix it in prod, ftw!
posted by Thorzdad at 11:57 AM on December 29, 2019 [4 favorites]


Most modern software is pretty resilient to date issues. People don't use fixed-field formats much anymore and 64-bit software won't notice 2038 at all.

But there's a lot of legacy software that was patched by making y < 20 mean this century. They had a whole twenty years to come up with a real fix! And then immediately forgot about it.

There are certain things that just can never have proper institutional support: renewing SSL certs, re-registering domains. They don't happen often enough to be staffed. And the staff that is supposed to "monitor" it turns over in the interim. These sorts of things will always be a problem.

I expect to see some fireworks this year. Probably nothing that can't be fixed with some frantic effort, but newsworthy nonetheless. If you support or interface with legacy systems, make sure your holiday plans don't extend into the new year.
posted by sjswitzer at 12:02 PM on December 29, 2019 [2 favorites]


Yeah, 64-bit software won't notice at all unless it relies on a library which is still 32-bit. I'm sure there aren't any of those around.
posted by Ickster at 12:15 PM on December 29, 2019 [4 favorites]


I spent New Years eve 1999 in an office of the CBC where I worked, waiting to intervene in the case of computer failures at 00:00 hours. Everyone else in town was partying like it was 1999. Here we were, a dozen or so managers, technicians and IT staff grimly waiting for what turned out to be nothing. The worst thing that happened? We momentarily lost contact with the WWV time signal, but our building's internal clock system free-ran for the seconds it took to re-obtain sync lock.
As I was a manager, I didn't even get OT, although we all got custom CBC Y2K T-shirts!
posted by Zedcaster at 12:46 PM on December 29, 2019 [5 favorites]


Oh, and why would we at CBC in Canada use the USA's WWV and not Canada's Dominion Time Signal, I hear you (not) ask? Because the time signal was sent via radio and Denver was closer to Vancouver than Ottawa so there was slightly less lag.
posted by Zedcaster at 12:51 PM on December 29, 2019 [7 favorites]


Somewhere I have the Utne Reader's Y2K possible-disaster prep booklet.
posted by tula at 1:28 PM on December 29, 2019 [6 favorites]


I saw this article linked on Facebook a day or two ago. It was the starting gun for The Usual Suspects to assure us all the Y2K was a huge hoax and obviously no airplanes fell out of the skies into nuclear reactors that we in mid-meltdown, so we can safely dismiss this climate change nonsense as well as another scam to prise cash away from the gullible. Okay then.
posted by ricochet biscuit at 1:43 PM on December 29, 2019 [6 favorites]


I recall he was similarly wrong about the viability of e-commerce in the 90s.

This one always makes me laugh because he said the problem is that people want to talk to clerks when they're making purchases. I feel a little bad that not having to talk to anybody is a selling point for e-commerce, for me, but I am also pretty certain I'm not the only one.
posted by atoxyl at 1:44 PM on December 29, 2019 [5 favorites]


They had a whole twenty years to come up with a real fix! And then immediately forgot about it.

Speaking as a software developer, this is depressingly common. Just as one example, one of the sets of integration scripts I work on is prefixed with "tmp" everywhere. They were written as a stopgap five years ago...and they're still one of the main jobs we run every day.

They're supposedly getting replaced in the new year, but I'm already dreading what fixes will be put in "just for now" that'll outlast me...
posted by Mr. Bad Example at 1:46 PM on December 29, 2019 [8 favorites]


Also:

I spent New Years eve 1999 in an office of the CBC

Mrs. Example and I used Y2K as an excuse to leave a dire, horrible party. I was on call that night but wasn't expecting anything to happen, but this party was so terrible that I snuck into a bedroom, called one of my coworkers who was in the office, and asked him to page me (this was the 90s, after all) with a fake crisis so we'd have an excuse to leave.

He did, and we did, and not having anything else to do, we drove over to my office to hang out with the people staffing the place for the Y2K rollover. We wound up having a better time--we had some drinks in the conference room, our boss played the mandolin he'd been learning for us, and at midnight we watched what we could see of the local fireworks from the building's loading dock. It was a much nicer night than we'd been having.
posted by Mr. Bad Example at 1:51 PM on December 29, 2019 [23 favorites]


Splunk issued a release note that two-digit dates beginning in 2020 are not processed properly (unless you upgrade). This is amusing because they chose the industry-standard 20 as the cutoff for rolling the date, but no log that Splunk will ever process could ever have had a date before the Unix epoch, 1 Jan 1970. So they could have chosen that.
posted by sjswitzer at 2:16 PM on December 29, 2019 [5 favorites]


Yeah, 64-bit software won't notice at all unless it relies on a library which is still 32-bit. I'm sure there aren't any of those around.

I know you're being cynical, but one example off the top of my head is mysql's timestamp data type is still implemented as a 32bit int. There's been an open bug for years, a decade maybe. Even a patch, as I guess this is one of those things that looks like a quick fix but will probably have a massive amount of potential edge case bugs.

The catch is that it's the only data type in mysql where you get the time zone conversion for free, that is: you're connected as not-UTC, but the timestamp value will be transparently converted to/from UTC as required. This makes queries using time ranges, which may or may not be in the time zone that the data was written the same as it being read, straight forward.

So if your app is time zone sensitive (hint, it probably is because: internet) then you either use a timestamp and have all that stuff for free but worry that it will break for dates post 2038, or you store a datetime *and* the time zone and handle the conversion yourself.

I mean, this is all easy stuff if you're out working in a greenfield, but imagine you've got 20+ years worth legacy schemas and code and all that stuff. And you do, because this is mysql the most widely deployed database server software on the web.
posted by lawrencium at 2:18 PM on December 29, 2019 [5 favorites]


I wish I could dig them up, but there were a couple of made-for-tv movies that dramatized the Y2K crisis and they were HILARIOUS. I worked with computers at the time, I did a bunch of Y2K-related fixing, and I knew it wouldn't be a big deal.

“Nothing serious happened“, because of the aforementioned billions of dollars spent to make sure nothing happened.

Well, sure, but also the drama was wildly exaggerated. People were seriously predicting millions of non-working cars sitting lifeless on the freeway, planes falling out of the sky, buildings somehow catching fire... the panic on the news was insane.

But there's a lot of legacy software that was patched by making y < 20 mean this century.

Crap, this rings a bell and I'm pretty sure I did this on one of my websites. I better find that...
posted by mmoncur at 2:24 PM on December 29, 2019 [1 favorite]


Yeah, I dunno. The information systems that underlie finance and logistics kinda make our way of life possible. More so now than 20 years ago. A massive failure would be a helluva speed bump at least.

Not to relitigate the past but this is a case of the prevention paradox. Any tragedy averted can be written off as Chicken Little. But was it really?
posted by sjswitzer at 2:33 PM on December 29, 2019 [6 favorites]


Well, sure, but also the drama was wildly exaggerated. People were seriously predicting millions of non-working cars sitting lifeless on the freeway, planes falling out of the sky, buildings somehow catching fire... the panic on the news was insane.

Well, sure, 99% of problems were going to be 'program displays 19100 instead of 2000', but it's the 1% of programs that take the system down with them that are going to get you.

It was an early lesson in how the news media aren't capable of describing risk, but it was helpful for getting companies that would have otherwise tried to ride it out to treat their software as part of their business.
posted by Merus at 2:35 PM on December 29, 2019 [3 favorites]


But there's a lot of legacy software that was patched by making y < 20 mean this century. They had a whole twenty years to come up with a real fix! And then immediately forgot about it.

That's the "windowing" approach (as opposed to the more permanent, but more difficult solution of actually expanding the size of the year field to four digits.)

The "20" in this case is the "pivot" year. As I recall, 50 was commonly used as the pivot... at least that's what we did in the mainframe software at a company I worked at in the late 90's.

A huge amount of mainframe/legacy/COBOL/whatever software relied/relies on fixed-width data fields. Changing all of them, and all the code that reads and writes to them can be a major undertaking. The "windowing" thing was usually more expedient.
posted by JeffL at 2:36 PM on December 29, 2019 [2 favorites]


As mentioned upthread, y2k forced a laser focus on information systems. And this happened just as the internet was exploding. So a lot of companies started investing massively in modernizing their information systems. That created a real boom in platforms and "middleware" software. Also, some of those companies (but only some!) came out with a competitive advantage.

How that all worked out was... mixed? But it really did shake things up.
posted by sjswitzer at 2:49 PM on December 29, 2019 [1 favorite]


> one example off the top of my head is mysql's timestamp data type is still implemented as a 32bit int. There's been an open bug for years, a decade maybe.

Is it cynical of me to think that Oracle's interest in fixing that can be correlated to when they can make the most money by fixing it?
posted by ardgedee at 2:57 PM on December 29, 2019 [3 favorites]


A weird but lucrative year.

Nothing to do directly with the so-called "millennium bug," but New Year's Eve 1999 also was a very good payday for a lot of jobbing musicians. This mostly was a supply-and-demand thing, with more big parties being thrown, and more of those parties hiring live bands instead of (or in addition to) DJs. The band I was in at the time made $1000 per player for a gig that in any other year would have paid us each $400 or so.
posted by Nat "King" Cole Porter Wagoner at 3:17 PM on December 29, 2019 [9 favorites]


I work on mainframe code. I've already had to repair code that I myself wrote with comments like "Picking an arbitrarily far year. If we are still using this code then, we have bigger problems."
posted by tofu_crouton at 3:55 PM on December 29, 2019 [25 favorites]


The NOC at work was all-hands-on-deck with senior engineers on site round the clock in shifts. And we were promised that there'd be catering and stuff. I guess a couple of half-bagels and a veggie platter were catering. Sort of.

And, yeah, that logging company with the 2020 bug. Come ON, man, what sort of SPeLUNKing did they have to do in their code to fix that, I wonder?
posted by rmd1023 at 4:04 PM on December 29, 2019


I worked for an "automated COBOL code remediation for y2k" company called MatriDigm (portmanteau of matrix and paradigm, for reasons) around that time frame. Almost everyone either fixed it in house or took the opportunity to upgrade mission critical systems. I think we got maybe one or two banks as customers and the money dried up pretty quickly. Office Space is not a bad movie, but it isn't my favorite either, if only because I've heard every variant of joke about TPS reports.
posted by BrotherCaine at 4:18 PM on December 29, 2019


A logging company? What, do the chainsaws stop working?
posted by BeeDo at 5:08 PM on December 29, 2019 [1 favorite]


Somewhere I have the Utne Reader's Y2K possible-disaster prep booklet.

I got that too, and at the time thought it was a little...much. I was working in a bank at that time, and the number of legal memos I saw flying around between different corporations from people who were swearing that their company had sorted out its own Y2K issues on their own end was oddly reassuring - if only because everyone was in such an all-hands-on-deck state about it that I figured that everyone was actually doing something about it for a change.
posted by EmpressCallipygos at 5:17 PM on December 29, 2019 [2 favorites]


Whether the dangers were exaggerated and overblown by sensationalist coverage in advance or, alternately, numerous potential problems were averted by long hours’ work by coders, I am not 100% sure. I do know that the sole major casualty of Y2K for me was the automated bus check phone line for the TTC (Toronto Transit Commission), which ceased functioning at midnight. For reasons that passeth all understanding, it took the TTC about six years to restore the service.
posted by ricochet biscuit at 7:44 PM on December 29, 2019 [3 favorites]


I do wonder if that paradox sjswitzer mentioned might contribute to the population-level sense of... lack of risk aversion, for lack of a better term, these days? I mean, I'm so glad there were enough resources being thrown to the problem that any major crisis was averted, but it didn't happen by magic. Yet, I wonder if that is an underlying dimension to say, the lack of institutional panic over the number of cities being hacked for ransomware. and we're talking about threats to the same kind of critical services that Y2K was supposed to impact.
posted by cendawanita at 8:55 PM on December 29, 2019 [2 favorites]


The thing about y2k is that it was a perfect storm of stupid and smart. It's a thing we could have foreseen, but didn't and then did.

If we'd not responded it would really have been terrible. But we did and it wasn't.

There are definitely other threats out there as cendawanita notes. Lots of work is going into countering those threats.

Will it be enough?

Probably!

Because the shit won't all hit the fan at the same time. That's the thing about y2k: the shitstorm would have been a gale.
posted by sjswitzer at 9:11 PM on December 29, 2019 [2 favorites]


MetaFilter: a perfect storm of stupid and smart.
posted by kirkaracha at 9:36 PM on December 29, 2019 [2 favorites]


A logging company? What, do the chainsaws stop working?

Payroll.

Inventory tracking.

100-year project(s) planning.
posted by porpoise at 10:13 PM on December 29, 2019


Counting logs.
posted by thatwhichfalls at 6:25 AM on December 30, 2019 [2 favorites]


That's the thing about y2k: the shitstorm would have been a gale.

We had updates to our 4GL, code updates, and pc bios updates to do for Y2K prep. Got it all done and tested, then walked into the board room, let everyone know we were ready, and I was going to Atlanta to see Widespread Panics NYE2K shows.

Phone call from my lead on 1/1 on the way back to Albany, while waiting for a plane in Chicago, everything was good.
posted by mikelieman at 6:30 AM on December 30, 2019


Hoo boy. I'm beginning to see lots of “Y2K Trutherism” showing up: because nothing cataclysmic happened, it was all a scam. Le sigh.

We weren't doing anything mission-critical at work, but we dedicated one member of our computer team from the beginning of 1998. A dedicated Sun workstation with all patches was set forward to 2000, and promptly wouldn't boot. So yeah, it needed investment and time.

Oh, and why would we at CBC in Canada use the USA's WWV and not Canada's Dominion Time Signal, I hear you (not) ask?

Good luck getting CHU on the west coast. I can barely pick it up in Toronto.
posted by scruss at 8:09 AM on December 30, 2019 [3 favorites]


I suggested that we expand the date field to four digits, but he nixed that, and told me to use two, and that "we'll worry about the year 2000 or whatever" later on. He really hated to expand the sizes of fixed-width records.

What the hell. Changing schema of fixed-width records is a pain in the ass, yeah, but going from 1 to 4 is no harder than 1 to 2.

Finally escaped the AS400 world earlier this year, but I still remember dealing with systems that had dates stored as 4 part century/year/month/day, sometimes as integers sometimes as chars, along with older tables that use JDE Julian, or newer tables that actually used date fields, and trying to get them all to work nicely with each other.
posted by kmz at 8:22 AM on December 30, 2019


I was working at Chase Bank in 1993, and my boss sent me to a meeting to hear about this Year 2000 problem. The presenter described the issue and the various solutions. The problem with two digit year fields would be patched with something he called "windowing". Any year prior to 35 would be 20XX, and 35 to (some threshold I don't remember) would be 19XX. He acknowledged that this was a temporary solution, but there would be plenty of time to devise a better solution in the future. The room full of about 200 people went quiet. You could tell that almost everyone there was calculating that they would be retired long before the 1935 / 2035 problem cropped up.

I currently work with an AS400 mainframe system at a different bank. We have loans that mature past 2034. When I pull data for various reports I have to do a find/replace on those maturity dates later than 2034 to correct them from 1935 to 2035, and so forth.

I assumed I would be retired, but it caught up with me.
posted by Midnight Skulker at 9:01 AM on December 30, 2019 [8 favorites]


I'm beginning to see lots of “Y2K Trutherism” showing up: because nothing cataclysmic happened, it was all a scam. Le sigh.

I think this has been a very common opinion for years, now - most of the time when I see Y2K come up it's somebody debunking it.
posted by atoxyl at 9:29 AM on December 30, 2019 [1 favorite]


Once every few years I'll preach a sermon where I really lay into the end-timers. I try to punch up as often as possible but these people make lose my mind.
And, invariably, I'll dust off an old chestnut:
"To this day, in the back of my uncle's old barn, there is a hand-carved wooden sign that reads: Y2K ready."
cue wry chuckles.
"Kids, you'll have to ask your parents to explain that one."

This is the thing. My uncle is a fool in the same, boring fashion as any other caucasian conservative male boomer anywhere in the Midwest.
The infuriating thing is his total confidence in his own abilities to survive anything.

He didn't understand a damned thing about Y2K. But, gawd dammit, he was gonna be ready unlike you effete liberal morons. I was 17 years old and his posturing and strutting was infuriating.

I came up with the category of white guys I now refer to as "know-nothing blowhards." Whether it's Y2K or something else - it's infuriating because they obviously know a little bit about the situation at hand, but they present themselves as absolute unquestionable experts.

I've heard it referred to as "engineer's syndrome." He is good at some things. He's pretty good on the grill. He's not a professional electrician but he's the first person I call for help. But it's like all of the actual knowledge he has is overshadowed by his boastful idiocy on whatever topic is being flogged to death on AM talk radio.

This is one of the things I hate most about my tribe. And I can see it, and I cringe for the world. White men. And I occasionally catch myself (a white man) behaving similarly and it makes me want to crawl into a hole in the ground. Loud, dumb, confidence.
posted by Baby_Balrog at 9:39 AM on December 30, 2019 [9 favorites]


Don't forget about all the front-end code parsing strings as dates going into the business code and database, and formatting dates as strings coming back out, and all the UI date and text box controls with their associated validation logic. There was a ton of code that did things like chop out certain characters from a string and assume that was the year, etc. It was a lot of work to fix all that.
posted by freecellwizard at 10:28 AM on December 30, 2019 [1 favorite]


I regularly remind people that Y10K is going to be huge. The date range of .NET and VBA ends on December 31, 9999. It usually gets a laugh, but I bet in a little less than eight thousand years, there will be someone digging out software manuals on how to fix C# programs.
posted by Xoc at 11:50 AM on December 30, 2019 [1 favorite]


I think this has been a very common opinion for years

Witness my comment history frothing at the mouth whenever someone dismisses Y2K as a big fuss about nothing. I'm thinking about commissioning a woodworker to make me a custom soapbox.
posted by Mr. Bad Example at 12:39 PM on December 30, 2019 [2 favorites]


Y2K20! My next cc payment is due 3rd Jan 1920. Whoohoo!
posted by jacanj at 5:14 PM on December 30, 2019 [2 favorites]


Consultants hyped the Y2K bug and IT departments got big budgets to deal with it, but it was never going to crash anything; only programs involving relative date calculations were going to be affected - for instance, some credit cards might have been temporarily rejected for wrongly calculated expiration dates - and there were easy code workarounds for that. Nowhere in the wold, even in places that didn't hype it as much as the USA, were there any consequential problems.
posted by yinchiao at 7:52 PM on December 30, 2019


Sorry, we've already seen first person accounts here of how even operating systems would crash along with all they support. Y2K was a real thing. If you don't want to believe it, ¯\_(ツ)_/¯.
posted by sjswitzer at 8:11 PM on December 30, 2019 [8 favorites]


One example. The software controlling the gas chromatograph on the rigs fell over at the Y2k turnover during lab simulations. No gas monitoring means, legally, no drilling. In financial terms, a rig spinning it's wheels like that is equivalent to setting fire to a $5 bill every second.
That's just from my small part of a complex and computer-ridden world.
posted by thatwhichfalls at 8:19 PM on December 30, 2019 [6 favorites]


but it was never going to crash anything; only programs involving relative date calculations were going to be affected

You are very, very wrong.
posted by JeffL at 7:51 AM on December 31, 2019 [5 favorites]


I regularly remind people that Y10K is going to be huge. The date range of .NET and VBA ends on December 31, 9999. It usually gets a laugh, but I bet in a little less than eight thousand years, there will be someone digging out software manuals on how to fix C# programs.

After 8 millennia, the changes in human culture and language and computing technology should be epic.
posted by ZeusHumms at 9:51 AM on December 31, 2019


Today I had a call saying all our clients were unable to log in. Turns out that for some reason I'd set the default account expiry date to 1st January 2020. I don't even remember why accounts had an expiry date in the first place. In the spirit of seasonal goodwill, I've changed them all to 1st January 2030. What larks!
posted by pipeski at 3:25 PM on January 1 [1 favorite]


Dates are just bad. We had a problem report for a thing that was a result of us not handling the end of the year gracefully. So do we fix it now or punt and hope that code is rewritten before the end of next year?
posted by fifteen schnitzengruben is my limit at 10:47 AM on January 2


So do we fix it now or punt and hope that code is rewritten before the end of next year?

That's a rhetorical question, right?
posted by thelonius at 10:53 AM on January 2


Yeah, Y2K definitely gets trotted out by our more disingenuous politicians and credulous relatives as an example of the scaremongering always being an exaggeration because everything just magically turns out okay in the end, ignore the muffled sounds of ass busting behind the curtain.
posted by lucidium at 1:10 PM on January 2 [3 favorites]


That's a rhetorical question, right?

Of course! In the proper spirit of Kicking the Can Down the Road. It is painted pink and covered in a Somebody Else's Problem Field.
posted by fifteen schnitzengruben is my limit at 1:19 PM on January 2 [1 favorite]




« Older You should be reading romance.   |   Clementine Paddleford's high-flying, deep-sea... Newer »


This thread has been archived and is closed to new comments