JOHN THE
OBSCURE ™
©
2005
Effecto 2000: The Y2K Bug Really Did Almost
Destroy the World
Children born in the year 2000 are
now in kindergarten. I’m sure that as they grow up and learn about the nicely rounded
year of their births, most of them will be told the Y2K bug that briefly
colored that year was a myth, a hoax and/or an exaggeration.
Kids, don’t trust anyone born before
the Internet.
Because the Y2K bug was very real
and very dangerous. Untreated, it would certainly have crashed American and
international business and killed lots of people. Treated, it still caused
scores of problems that went underreported and were quickly forgotten.
Denying it is like saying the last
major California earthquake was no big deal because it didn’t kill 5,000
people, ignoring the years of building code and construction technique changes
it took to ameliorate the problem—and also ignoring the handful of people who
died anyway.
This column was born out of five
years of frustration at hearing people downplay or laugh at the Y2K bug. But
the really amazing legacy of the Y2K bug is silence. It’s virtually a forgotten
subject now. Google it and your top hits are fossil links. A recent article on Wired.com, a geek ashram, about the 10
worst computer bugs of all time didn’t even mention it.1 In the
ultimate insult, it’s gone down the memory hole like a forgotten war or a
faddish toy.
How could something world-shaking be
so easily dismissed—even forgotten? It was thanks to a complex nexus of
cultural attitudes, some of which are virtually unimaginable in today’s
climate.
One was utter computer illiteracy.
In 2000, even most middle-class people still weren’t online. They had no
personal context or perspective on what the Y2K bug meant. Descriptions of the
problem sounded simultaneously trivial and opaque: put most crudely, the
1960s-’80s practice of coding years into calendar-sensitive software by only
the last two digits, resulting in 2000 being read as 1900 and causing crashes
or misdating.2
The only context for the average
person was their personal computer at work or home; even more then than now, it
was viewed as a black box with mysterious inner workings. The Y2K bug didn’t
affect most people’s personal computers; there wasn’t much to affect in any
case, and what problems there were could be easily patched at the last minute.
This further trivialized the problem for most people; they didn’t realize that
a lot of those quick fixes were due to debugging software developed by the
massive Y2K immunization effort.
The main groups proactively
concerned about the Y2K bug were computer geeks and fundamentalist Christian
survivalists—two despised subcultures the mainstream enjoys laughing at.
The Y2K bug was also eventually
overhyped, but the mainstream public had no way of gauging when requisite hype
ended and overhype began. Dire warnings issued by authorities from the federal
government on down were certainly necessary to spur action to fix the bug. But
they continued after major repair programs were developed. And some of the
hysteria was certainly created for financial reasons by programmers and
consultants who made big money debugging things—the one truth the public was
quick to suss out. Anything less than total global
destruction would be considered a letdown after that hype, and was by the
mainstream.
(Other countries such as Spain—where the problem was known
as “Effecto 2000”—did much less repair work and
reported fewer problems, leading many of them to call Y2K a money-making fraud
hatched by Bill Gates or some other archfiend. The quality of the criticism
reflects the quality of their comprehension of software, which in turn reflects
their relative lack of computerization at the time; fewer problems in
There was also millennial anxiety tied into all of it, an
end-of-the-world mentality (or desire, certainly in my case) that folded neatly
into Y2K for many people, religious or secular. That anxiety was obviously
irrational, but somehow saying so became a dismissal of the Y2K bug, too. And
for those who did prepare for—even anticipate—disaster, the eventual perception
dovetailed with that of the mainstream: it was a letdown, a non-event.
At the same time, in the Gilded Age
of the Clintonian, SUV-inventing, dot-com bubbling,
“Titanic”-sinking 1990s, mainstream
The press was no help, either,
providing thin, scattered, sensationalist coverage that pre-Jan. 1, 2000
flirted with apocalypticism, then settled into a New
Year of triumphalism, perfectly matching the national mood. “Computers Prevail
in First Hours of ’00,” said a story on the front page of the Jan. 1 New York
Times, in a headline that sounded dazzlingly positive while being literally
meaningless. (Wouldn’t computers have prevailed either way?)
That headline also encapsulates the misunderstanding that
midnight, Jan. 1 was the problem time; in fact, Y2K bug problems were expected
to (and did) unreel over days, weeks, months and even years that followed as
various programs churned away at their appointed hours, which in most cases was
not midnight on a Saturday.
A final influence is that an even
bigger concern than the bug itself was the legal liability for it. Private
businesses were especially hit with Y2K problems, and naturally kept a lot of
them secret to this day. The Gartner Group, a consulting firm that tackled Y2K
bug problems, received about 400 confidential reports of problems in private
companies by Jan. 4, according to the Times.
“There were some things that were
critical that popped up months or years later, but you just fixed it,” I was
told by an employee at a large financial-sector company whose tenure included
Y2K and a massive internal debugging regimen. Those lingering problems included
programs crashing or throwing up error messages—fixed easily enough when the
employee would be called in to do so after-hours, but that would have been
debilitating without the overall debugging years before.
“I’m sure if I poked around now I
could find things that said it was 1905. It’s just not stuff that’s in
reports,” said the employee. “I don’t really have anything dramatic. I
certainly heard rumors that there were places…that had more serious problems
that were caught.”
“If there hadn’t been a massive
effort [to fix the Y2K bug]…large segments of the economy and infrastructure
would have stopped,” the employee said. “It wouldn’t have been the whole ‘at
the stroke of
Lingering glitches—that has been the
pale ghost of the Y2K bug, and part of what made it so easy to dismiss. Somehow
we forgot that enough glitches equals a disaster; after all, the Y2K bug was
not a problem with one brand of software, but rather a critical mass of
glitches in all software written before a certain era. “‘Critical’ is obviously
relative,” the financial-sector employee reminded me.
Let’s return to our national paper
of record, which eagerly brushed aside “minor glitches” in that Jan. 1 issue.
One of those “minor glitches” was the failure of a “monitoring system” at a
Japanese nuclear power plant.
Is there such a thing as a minor
glitch at a nuke plant? In fact, as the Times later reported, there were Y2K
bug “glitches” at an unspecified number of other nuke plants in
We’re told none of the glitches were
safety-related; those at
Obviously, none of the reactors
melted down or released radioactive steam. But what if the Y2K debugging hadn’t
happened at all? (Remember, there was no guarantee that debugging programs
caught everything at the time.)
Or what if there had been more
“glitches” like the one that caused windshear alert
systems at airports in Chicago (O’Hare), Tampa, Denver, Orlando and St. Louis
to display error messages? Windshear is a major cause
of plane crashes. We’re told the systems were fixed “within two hours.” We’re
also told there were similar problems, no duration given, at airports in
Meanwhile, computers at airports in
If SimCity-style
disaster doesn’t spook you, how about problems with security? A military
spy satellite was “blinded” for at least two to three hours, and then operated
for a couple days “at a reduced level,” after a ground station that made its
transmissions legible crashed.
Meanwhile, the power went out at
Diego Garcia, a British atoll in the
A field office for what was then
called the federal Bureau of Alcohol, Tobacco and Firearms (ATF) had an
unspecified problem with its security system that required guards to be posted
outside. The ATF refused to identify the office.
And an unnamed building in
Remember, this was after years of
Y2K debugging. What if 25 bases had gone dark instead of one? What if all the
spy satellites had been blinded? What if doors were unlocked around the
country?
A Swedish company’s kidney dialysis
equipment failed to go through its automatic cleaning cycle due to a Y2K bug.
That was caught and reportedly caused no human harm. But what if the problem
had been multiplied a thousand-fold in the medical field?
It’s amusing to read about a video
store computer charging a customer a $91,250 late fee for a 100-year-late
rental, or an unnamed company’s delivery software starting to generate phantom
orders that were caught prior to automatic shipping. Of course, it’s not
amusing if those were your companies. Imagine if those sorts of things had
happened to even 25 percent of US businesses.
The cutest stories were about dated
forms that popped out reading “1900”—a New Year’s party favor rather than a
disaster. But how funny is it that arraignments were delayed in Queens, New York courts because 50 to 100 police complaints
were dated either 1900 or 1980 by the Complaints Arrest Processing System
software?
Unidentified “stock exchanges and
other trading support organizations” reported computers misdating forms and
even miscalculating mutual funds assets. How cute would that have been on a
larger, out-of-control scale?
The same could be said for the other
glitches reported in the Times: ticketing machines jamming on public transit in
Sydney and Adelaide, Australia; “timing devices” and internal clocks going
haywire at various, largely unnamed, power utilities; a tide gauge in
Portsmouth, England failing; Sweden’s SkandiaBanken
online banking crashing; Greek cash registers malfunctioning; WordPerfect’s
word processor failing to delete files; Web pages and a reading service for
blind people giving wrong dates; untold numbers of small-business computers
crashing.
The very existence of these
relatively modest glitches somehow turned into proof that the bug didn’t exist.
And then, with the passage of time,
the equation was inverted; the Y2K bug was nothing, so there were no problems
at all. I’ve met people who literally remember it that way.
It must be said that the Times was
truthful enough at the time. It editorialized on Jan. 3 about the real threat
the Y2K bug had posed and the success of the fix. A similar point was made in
the Jan. 1 news coverage—albeit about 20 paragraphs in and on the jump page.
But the form and style of the paper
of record’s coverage belied the rationality. The Jan. 1 coverage emphasized the
funny/trivial glitches in a way that essentially downplayed the entire problem
rather than explaining the magnitude of the success in fixing it.
Over the next few days, coverage of
Y2K bug problems was scattered throughout several articles or columns in the
same issues rather than concentrated and analyzed, watering it down all the
more.3
By Jan. 5, coverage was essentially
over as the nation turned its attention to something it really enjoyed worrying
about—Cuban waif Elián González.
What’s
really intriguing is the similarities to the paper’s coverage of millennium
celebration terrorist plots, another significant fear at the time.
There were no terrorist attacks, resulting in similar triumphalist, “what, me worry?” type coverage—another
millennial mirage, it was dubbed. Then there were a few stories about curious
arrests and property seizures—explosives, men with Muslim extremist
connections. As we now know, there were major terrorist attacks planned for
several locations on Jan. 1 that were foiled by police.
I can’t blame anyone for remembering Sept. 11 forever, but
it was the anti-Y2K bug—a well-exploited series of airline security glitches
that could have been repaired easily with fixes suggested dozens of times
decades ago, but wasn’t. The Y2K bug could have easily matched the Sept. 11
death toll, but didn’t. Shouldn’t that be a cause for rejoicing, not derision
or dismissal, as a letter-writer to the Times pointed out in early 2000?
After all, the Y2K bug’s effects aren’t over—and I don’t
just mean those lurking glitches. One reason it is unthinkable as a problem
today is because of the massive infrastructure upgrades the debugging sparked.
Incredible as it seems, in the 1990s many major companies were still running
systems with Radio Shack computers speaking Old FORTRAN. The Y2K bug forced the
world into something close to 21st century technology.
If you think that’s not a big deal, here’s a little story
about Federal Aviation Administration (FAA) computer systems. At
Then on Jan. 4, an FAA computer in
And on Jan. 7, an FAA system in
At first, these problems were considered possibly Y2K
bug-related, but supposedly proved not to be. (I allow for liability-avoidance
and good PR here.) Instead, they were age-related crashes in FAA systems dating
to the 1960s that had not been upgraded and replaced as a result of Y2K
debugging.
Using outdated tech is a natural habit—and potentially a
deadly one. The Y2K bug was powerful enough to break that habit in most places.
“It was definitely a huge thing,” said the financial-sector
employee. “And I think it probably had a very wide-ranging impact even beyond
just avoiding that problem, because I think for a lot of organizations it was
the biggest project they ever did.” The massive review and rebuilding of every
corporate corpuscle changed companies, project planning and individual
careers—something the employee saw first-hand.
Kids, if anybody tries to tell you the Y2K bug was a fake,
ask if them if Salk’s vaccine meant polio was a hoax. I know you kids have no
idea what I’m talking about, but trust me, you’ll love the look on grown-ups’
faces.
1
The article’s author later explained the Y2K bug was in the original list, but he
removed it because it wasn’t “a single bug” but rather a widespread industry
problem, which he likened to thoughtless coding in general. To me, this is like
removing World War II from a list of history’s greatest conflicts because it
involved more than two countries.
2
There is dispute over whether the Y2K problem should be considered a “bug” per
se since the term usually refers to an unintentional glitch or oversight. The
two-digit year coding began as a deliberate act to save memory space, which in
those days was extremely tight and expensive. However, even as memory grew, the
two-digit coding practice remained for about a decade, which most programmers
agree was because of sheer habit. In any case, the two-digit coding was
recognized as a built-in problem from the start; its wide-ranging effects were
not wholly anticipated. Some programmers began writing debugging programs as
early as the 1980s.
3 I recently discovered a cache of my personal
Y2K archives, which include further glitch reports. On
Computers choked again on
Update: On Dec. 31, 2008, we saw what really happens with
unpatched clock bugs in computer drives. Thousands of
Microsoft’s Zune music players suddenly froze and crashed for reasons that were
initially mysterious. It turned out to be an improperly coded internal clock
driver that did not account for the calendar change away from a leap year.
Complaining users called this “Y2K9” or “Z2K9.” A stupid little clock updater
crashed thousands of machines simultaneously as Jan. 1 dawned. How can anyone
possibly argue that the Y2K bug, unchecked, would not have wreaked havoc in a
similar way, but at a far greater scale? (Source: Representative coverage in
“Leap-year glitch freezes Zune MP3 players” by Brandon Griggs, CNN.com, Dec.
31, 2008, at www.cnn.com/2008/TECH/12/31/zune.player.failures/index.html.)
The bulk of
the facts in this column were taken from all Y2K-related articles in the