For those who doubt the severity of the problem, Alan Light suggests “How To Explain Y2K To Non-Believers” — at

Krishna Kunchithapadam certainly doubts it. He writes: “I enjoyed your article on preparing for 2000. However, I do not agree with your prognosis. The best way to describe the hype surrounding the Y2K problem is through a joke:

“A man observes his neighbor spreading yellow cream on his lawn every day. He finally asks why. The neighbor says, ‘To keep away elephants.’ The man replies, ‘What do you mean? There’s no elephant around for thousands of miles.’ The neighbor winks knowingly, ‘Works like a charm, doesn’t it?’

“Y2K is a little like that. If you thought people took El Nino’s name in vain for any and every meteorological problem in sight, just wait till 1st January 2000. One of two things is going to happen:

  • “Every (and I do mean every) problem will be blamed on Y2K, and not on general programmer or human/corporate incompetence as it should be. Today, if you walk into your bank and get bad service, you can complain to the manager and, perhaps, get some remedy. Post-2000, the same incompetence will be blamed on Y2K, and damned if anyone is going to sympathize with your dissatisfaction.
  • “Or, every one of those Y2K ‘consultants’ is going to stand up and loudly proclaim that their doomsaying and ‘solutions’ worked. And they get to laugh all the way to the bank to deposit the fat checks they get for doing nothing more than drink coffee, and write silly books. ‘Works like a charm, doesn’t it?’

“There is some deep psychology in the Y2K phenomenon. The fact that company X is spending Y billions has little to do with ‘solving’ anything (or with the gravity of the situation), and much to do with avoiding culpability in lawsuits between 2000 and 2005 (say). I can bet that this cost will far exceed the paltry Y billions that a company spends (or claims to spend) now.

“Y2K has become the uber-act-of-god of the millennium, the ultimate anti-panacea, if you will. Everybody can absolve themselves of blame by pointing their finger at others, or else at some incomprehensibly vast, insidious, and pervasive abstraction. And if everybody gets ‘hit,’ no one feels like a sucker — there is comfort in numbers. It is this psychology that we see at work, not the ‘problem’ itself.

Legitimate Y2K problems are of two kinds: (1) programming ‘bugs’ and (2) user-interface problems.

“Type (2) problems are common, even rampant — my checks all say ‘____ 19__’ for the date field. Most ‘date’ entry forms have just two ‘boxes,’ not four. However, these are the easiest problems to fix. Nobody is going to refuse my check just because I strike out the ‘19’ and write ‘20xx.’ Imagine the IRS refunding your tax payment because the check/1040 has the printed ‘19’ crossed out. Sure, the IRS claims it has a problem and needs to modernize its computers — any excuse to increase its budget. Ditto for everyone else crying for a Y2K budget.

“Type (1) errors are the serious ones, but none of them are visible to the average person on the street. Most of today’s computer users were not even born when type (1) errors were caused. And, here is the important thing to know: the errors happened because they were easy to cause. Programmers either took the easy way out of a software engineering task, and/or did not plan for their systems to exist to 2000 (short-sightedness and premature optimization). Both of these approaches were facilitated by programming in COBOL (incidentally, the biggie culprit in Y2K), but the reverse is true with pretty much every other programming language in use (then or now).

“A vast majority of the systems that a person is exposed to today (Internet software, word processors, etc.) is not written in COBOL. Almost every type (1) Y2K problem is confined to programs written in COBOL. Fixing these problems is made difficult because the source code for these software systems is no longer available. Incidentally, the problem is not the lack of knowledge of COBOL (as is usually reported in the media, along with stories of grizzled old-timers coming out of retirement to save the world, a la Armageddon — more psychology). Any competent CS grad student armed with a COBOL language reference can fix all Y2K problems in a matter of days, IF the source code is available.

“Specifically, both of the supposed Y2K problems you mentioned in your article are emphatically not Y2K.

“First, embedded processes have never, to my knowledge, been programmed in COBOL. In pretty much every other candidate programming language for these systems, it is not only not easy, but actually excruciatingly difficult to allocate no more than 2 digits for the year part of a date. Although I have no illusions about the laziness and short-sightedness of most programmers (being a student of computer sciences myself, and having seen my fair share of shockingly bad code), it is orders of magnitude more difficult to cause a Y2K problem in a language like C than it is in COBOL. So much more difficult that I can confidently say that a Y2K problem does not exist in such an embedded system.

“Moreover, just ask yourself, what is an embedded processor in charge of switching lines doing managing dates at the granularity of years. These systems work at times scales of seconds, minutes, and (at most) hours. One might as well say that these processors keep track of the difference between day and night, the change of the seasons, world time zones, daylight savings time, and (shudder) the ascension of Mars and Saturn as per some new-age horoscope.

“Here’s more psychology — Y2K hype no longer has anything to do with the real world, it has become the domain of sci-fi doomsday plots. Here’s how the game is played. Friend A suggests a widget (the more obscure and removed from everyday affairs, the better). Friend B now concocts a doomsday story about how the widget will break down because of Y2K and bring an entire planet to a grinding halt. Plausibility alone matters — facts are a hindrance and will be stubbornly ignored. Pithy remarks about the inter-connectedness of the cosmos, chaos theory, and butterflies in Hong Kong working in concert with El Nino.

“Your second example, about cooling fans in a mainframe, is even more off base. No processor ever determines if ‘20 minutes have elapsed’ by subtracting the current year from 1900 (or 2000 or whatever). These systems do subtract a constantly changing value (the current ‘time’) from an epoch value — but for a number of technical reasons, a problem will occur, if at all, sometime in 2036 or 2037 or some other equally non-millennial year (computers being binary, 2000 is not as ‘special’ to them as it seems to humans; conversely, what is ‘special’ in computerese is a seemingly random date for a human). It does not matter whether the computer/fan in question was built yesterday or during the 1960s. Serious programmers are more worried about the Y2036/Y2037 problems (and have already come up with standards and transition schemes to solve the problem long before it becomes a crisis).

“Want more psychology? Y2K, being the ultimate anti-panacea, has the power to paralyze the mind, and turn humans into helpless slaves of technology run amok. Why else would anyone [A.T.: i.e., me] suggest that a power breakdown would cause gas stations to stop working (and presumably lead to long lines of cars stranded on the highways)? Come on, people, ever heard of a siphon? Ever considered that an empty juice-can can also transport gasoline? Or did your circadian clock also throw a fit because of Y2K?

“It seems as if even the most resourceful human will suddenly, on the eve of 2000, stop thinking and sit motionless before the Y2K menace.

“All Y2K hype that I have seen has this property — superficiality. The supposed disaster is based on a number of ludicrous events that even a mildly thoughtful analysis would reveal are irrelevant to the situation at hand. Or else, there are plenty of solutions, trivial ones at that, available at hand.

“Sure, there will be some Y2K problems. However, almost none of them will be as ‘visible’ as the media hype makes them out to be. Of course, media hype helps float ‘solution companies’ and ‘consultants’ and even ‘Y2K stocks’ (let’s start a Y2K index and trade options on it). Almost none of the problems will cause a global catastrophe. [A.T.: Does this mean that only a handful will cause global catastrophe?] What is more, all of the problems are easily solvable.

[A.T.: Well, this is the point addressed in the article I referred to up top — “How to Explain Y2K To Non-Believers.” The article explains that the problems are easily solvable — just very numerous. Its author, Jim Lord, likens the problem to polishing a basket of marbles by Friday. He hands you the basket, a rag — what’s the big deal? It’s boring, but that’s it. Except, he says, picture not a small basket of marbles, any one of which is easily polished, but the Grand Canyon full of them. That’s a lot of marbles to polish by Friday, he says, or even by 2000.]

“There is just one thing that needs to be learnt from the Y2K ‘problem.’ However, the lesson is for software engineers. And there is good reason to believe that such lessons were learnt by the 1980s — the rest of the world is about 20 years too late with the story.

“Still, I am sure that the Y2K frenzy will continue — after all, this is a world that predominantly reads (and believes) the astrology pages. Post-2000, ‘yellow cream’ will be proclaimed a success by one and all. As for myself, I plan to fly (simply for the sake of flying) on 1st Jan. 2000 — what is more, I am going to get my air-ticket dirt-cheap. Nyah-nyah to the chickens-little of the world.”

A.T.: Well, I am fairly certain Krishna is right about one thing: He will get that airline ticket dirt cheap. (Of course, as a precaution, the FAA might require airlines to scale back flights by 60% the first week of January, which might drive up the price of seats — nothing is certain when it comes to predicting the future.) And I certainly hope he is right about the rest. But as noted, I’d still take at least the same sensible precautions one takes in anticipation of a power outage (i.e., in hurricane season), and maybe a few more. And I wouldn’t wait until November 1999 to do it.

You have so many really smart and self-assured technical people, like Krishna, making one case or another — and they disagree. I do believe the scariest scenarios are highly unlikely, if only because there are so many very smart people in government and industry working on the problem. As long as we have the power and communications grids largely working, as I expect they will be, improvising the rest is a relative piece of cake. But I forwarded Krishna’s e-mail to a friend who is orders of magnitude smarter than me and whose opinion, given his credentials and background, must be credited at least as heavily as Krishna’s. He responded:

“Krishna is either ill-informed or in denial. An unknown fraction, ranging from about 0.2% (Gerstner estimate) to perhaps 7% (the highest knowledgeable estimate I’ve seen), of embedded processors do have a Y2K problem because of two-digit year-date fields. It has nothing to do with COBOL. Embedded processors have also caused quite a few well-documented hardware failures all over the world: no myth. Nor does the legacy code issue — which is mainly COBOL as it relates to mainframes but includes many (even very modern) apps, BIOSs, and operating systems in PCs — have anything whatever to do with preprinted checks or other forms filled out by hand. I’d suggest you ignore this guy.

“My wife, by the way, was told last Tuesday that her driver’s license had expired because although you could see by looking at it that it expired on her birthday in 2000, a computer in California just read the ‘00’ and thought it meant 1900. Fortunately, there was a human being there too.”

A.T.: I think one can fairly reasonably conclude that the problems will not be catastrophic unless the whole world panics and makes them so (unlikely). But beyond that, no one knows. Could there be business failures and a recession? Certainly. (Young readers may be interested to know there were once recessions fairly regularly even without unprecedented global technology problems.) Could there be power outages for a few days while things are getting sorted out and brownouts or pockets of blackouts for a few months while the entire grid is being brought back up to full inter-connected capacity? Certainly. Might gas lines be pretty long during any such power outages even if siphons and empty juice cans are called into service? I should think that, too.

(Might a terrorist or two have infiltrated the tens of thousands of people working to solve the problem, hoping instead to plant a few whopping problems? One shudders to think.)

So I find myself in the odd position of recoiling at and largely scoffing at the flood of gloom-saying, exploitative newsletters and newsletter solicitations I just got in the mail last week (Howard Ruff headlined his: “MILLENNIUM MELTDOWN: We Turned Our Lives Over To Hal . . . And He’s Going Nuts!”), but not entirely. If we take the problem seriously enough, it probably won’t be that much of a problem. Those who dismiss it all as hype, on the other hand, are taking unnecessary chances. How hard is to stock up on water, tuna, and a little emergency gear? These things are always worth having anyway.

More on this in a few days — many of you have sent me very interesting feedback. I’m learning a lot! (Maybe hold off on sending more until you see if I cover what you were going to add?)



Comments are closed.