Probably because we're at the 20th anniversary of the Y2K non-event, a column about the subject that I posted a while back spawned a bit of a Twitter storm of criticism. Normally I'd just tweet a response, but Twitter isn't a good forum for nuanced explanations.
At the end of the post, I'll explain why Y2K is still important but for now I'll turn to the complaints. The following two tweets are representative:
In other words,
- "We fixed some Y2K bugs, so Y2K wasn't a hoax"
- "There weren't any disasters because we fixed all the bugs."
Let's look at each complaint in detail.
How Dangerous Was the Y2K Bug?
Nobody seriously questions the existence of Y2K bugs. The question is whether Y2K bugs were serious enough to cause the predicted disasters, which included:
- Power grid failures.
- Planes falling out of the sky.
- Home appliances exploding.
- A total collapse of the financial system.
- Meltdowns of nuclear power plants.
- Etc., etc.
Y2K, in other words, if not fixed would be "the end of the world as we know it."
In September of 1998, twenty-one Y2K experts "ranked the problem ranged from 0 for absolutely no concern to 10 for a belief that the problem is so serious that major worldwide social, economic, and technological disruptions will occur."
That list of "experts" included the most widely-quoted Y2K sources including Dr. Ed Yardeni, who ranked the seriousness at 8.0, Ed Yourdon, who ranked it at 8.0, and Dr. Gary North, who ranked it at 10.0. (Put a pin in that Gary North name.)
All but one expert rated the seriousness of Y2K at a 6.0 or higher, with the bulk of the rankings clustering around 8.0. So, the consensus--at least among experts who were building their careers around Y2K--was that the problem represented a major, world-changing threat.
The Gartner Group (a big computer industry analysis firm) was recommending that people stockpile food and water for at least two weeks, get some cash so you have money when the banking infrastructure goes down, and so forth. But Gartner was considered rather conservative when it came to ringing the klaxon; the real consensus was "we are totally screwed."
So, why did the experts think the situation was so dire? Read on:
The Origins of the Y2K Disaster Scenarios
There were three explicit threads of reasoning behind the Y2K disaster scenarios:
- Catastrophic Y2K pre-testing. There had, according to numerous reports, been several roll-the-clock-forward tests of Y2K problems, including one that caused a nuclear power plant to malfunction.
- Unfixable Systems. The Y2K bug existed inside embedded systems (some of which were in the power grid) where date checking was inside the hardware. Fixing the Y2K bug thus meant replacing millions of semi-computerized devices located who knows where--an impossible task.
- Cascading Faults. When Y2K hit, there would be a snowball effect (aka "cascading faults") that would spread from system to system, causing additional geometrical and exponential system failures, even in applications that did not have the Y2K bug or where it had been fixed.
Together those threads of reasoning comprise a pretty good case for a "head for the hills" strategy. But in fact none of them hold up under logical examination.
1. Those "Catastrophic Test" Disasters Never Happened.
When I was asked to write a feature article on Y2K for Upside magazine (I eventually wrote several others), I did what other reporters failed to do: try to find an original source for the roll-forward disaster stories. None of them panned out. In fact, what testing I could find that was actually done turned out to be duds: the roll-forward created no problems.
So, then, from whence did those dire stories emerge? Somebody, inserted them into the discussion, resulting in a round robin of media outlets (including such luminary publications as the Smithsonian) quoting each other, without a full fact-check.
2. The "Unfixable System" Problem Was Imaginary.
If the Y2K experts (some of whom had a software background but none a hardware background) had actually asked some electrical engineers about date checking in embedded systems, they would have learned that only a complete idiot would do anything resembling the conversion and comparison of calendar dates inside a chip. You use elapsed time, which is a simple, single counter; it takes ten seconds to add to a circuit.
I'm may oversimplifying but ultimately the reasoning doesn't matter. It's now clear that the unfixable systems problem either wasn't real or wasn't significant enough to spawn a disaster. Because there weren't any.
3. The "Cascading Faults" Concept Was Frankly Idiotic.
If system "A" passes a crappy piece of data (say a record with a date of 1900 rather than 2000) to system "B," it's not going to crash the system. Depending on how system "B" is programed, it will either 1) pass the record through (maybe resulting in, say, the system issuing a check dated 1/1/1900), 2) reject the record, or 3) flag it as flawed. And that's it.
One bad field in a data record isn't going to spread like like a virus.
The Hidden Origin of the Y2K Disaster Scenarios
In addition to the three reasons above (which were widely publicized), there were two less-publicized reasons the "experts" were so convinced a Y2K disaster was inevitable:
Reason 1: the Y2K "experts were making big money selling books, seminars, and consulting. The more dire the danger, the more money they could charge. That's why the "experts" quickly wiped the Y2K stuff from their bios and websites when nothing happened. What had been an asset--"I'm a Y2k Expert" became an embarrassing liability. (Uh oh... I printed hard copies and filed them away.)
Reason 2: some Y2K "experts" predicted an apocalypse is because the Bible told them so.
(OK, remember the pin you put in the name "Gary North"? We're going there now.)
While there had been several bestselling books published on the Y2K threat, the nexus of online information about Y2K was GaryNorth.com, whose site was like a Reddit thread (this was before Reddit) where he posted links to every article or news report about Y2K that hit the Web, along with his commentary.
Because North was at the center of the online Y2K world, North became the "go-to" source for reporters looking for new wrinkles on the Y2K beat, where he was usually quoted in Y2K "the end is near" articles as a "computer expert" or "programming expert."
Few of those reporters ever bothered to check on North's background. A better description of North than "computer expert" would be "world-class wing nut."
For example, in 1976, North published None Dare Call It Witchcraft, which asserted that witchcraft was real--the real-live broomstick-flying kind--who were summoning the devil on a daily basis. This is not a joke. North thought witchcraft was as real as science.
But that's not all. North's other writings identify him as a "Christian Reconstructionist." For those who aren't familiar with it, Christian Reconstructionism is an ultra-right-wing sect that believes that the second coming of Christ is imminent. The Reconstructionists (and many, many evangelicals) went BIG for the idea that "since God created the world in 4,000 B.C., the world would end on 2,000 A.D." (Each millennium being one of the six days of creation... get it?)
North didn't broadcast these whackdoodle beliefs to the reporters who interviewed him, but he successfully folded apocalyptic imagery and concepts into the Y2K discussion through his "clearing house" of Y2K online links. This intensified of the sense of impending doom was helped along by his commentary, which always put the most dire spin possible everything.
So, when you peel away the onion layers of the Y2K disaster scenario, you discover at its core nothing more than that "ol' time religion."
Why Fixing the Y2K Bug Didn't Really Avert Disaster
So, then, what are we left with? Not much. A bug that needed to be fixed, but which never posed a world-shaking threat.
Which leads us to second complaint about my original post: "There weren't any disasters because we fixed all the bugs."
The IT folk claiming this aren't thinking straight. They need to look at every large IT project that they've worked on. How many of those projects have been completed, perfectly on time, running near flawlessly, the moment they're switched on?
I'll answer that: exactly zero.
If you work in IT or have ever worked in IT, you know I'm right about that. If there'd been the possibility of huge disasters, at least one, and probably many, would have fallen through the cracks. Murphy's law was not repealed for Y2K and Y2K bug fixing was not an IT miracle.
Beyond that, the potential severity of unfixed Y2K bugs can be assessed simply by looking at countries that didn't do much or any Y2K bug fixing. This, from the January 7, 2000 edition of Australia's Financial Review:
"By looking at the impact of Y2K in countries like 'Russia, Bulgaria, and Vietnam, we are able to get a reasonably good fix on what would have happened if we hadn't prepared as much in countries like the US, UK and Sweden', said Mr John Gantz, IDC's chief research officer." (IDC is International Data Corporation, the second largest computer industry analyst firm at the time, rivaling Gartner.)
While some piddling Y2K problems occurred in those countries (as in the US) there were no disasters, even though their computer systems were more likely to contain obsolete software.
Ipso facto, there wasn't a substantial threat of an "end of the world as we know it" disaster due to the Y2K switchover.
Now, as a baby-boomer myself I can empathize with the boomer IT folk who want to tell their grandchildren they did their part to save the world. Very superhero. Yay, us! But, between you and me, it's just not true. I realize it's not nearly as fun to tell your grandchildren that you fixed a bug that kept some of your company's older programs from erroring out.
But that's what you did.
Why Y2K Is Still Important
Why is this important today? Here's why: the failure of any Y2K disaster to occur distorts how some people view climate change.
On the one hand, some people compare climate change to Y2K and believe it's a similar hoax. That's dumb because climate change, unlike Y2K which had a deadline, is something that we're experiencing on a daily basis. It's not something where we'll hit a certain date and then we'll find out whether or not the disaster scenario is real.
On the other hand, some people figure that Y2K was a huge threat that we averted by throwing some money at it--that any problem caused by technology can be solved by more technology. The logic is, "hey, since we prevented the Y2K disaster we can do the same with climate change." And that's REALLY dumb because it vastly minimizes the difficulty of addressing climate change, which isn't going to be fixed by an intrepid band of IT geeks spending a couple of years patching systems.
In other words, Y2K and climate change are completely different but they're treated as parallel cases because they're both associated with apocalyptic predictions. Y2K was the "experts" crying wolf. As a result, the "experts" now have a credibility problem. They're not the same experts but the damage, alas, has already been done.