From the estimable Dorothy Mallonee:

My husband, Bill, works for a company that sells solutions to Y2K problems, so I am familiar with some of the issues, and have a couple comments.

First, Y2K is so big and scary that companies are in absolute panics. Some of them are so overwhelmed that they will listen to anyone, buy anything, and pay any amount of money to someone who promises to fix their problems. I have heard about companies that will fix your code for, say, $1.50 per line; some organizations have BILLIONS of lines of code. There’s lots of opportunity out there for Y2K charlatans; we should not suspend common sense when trying to solve the problem.

Second, Bill agrees with today’s correspondent that ’98 and ’99 will see the beginning of some big troubles. He suggests that, starting TODAY, your readers keep every scrap of paper dealing with financial transactions: statements, canceled checks, buy/sell confirmations.

Third, don’t be too confident that, because people know about the problem, they have it in hand. I heard a few weeks ago that some federal regulatory body fined some bank because it had not even started to deal with its Y2K issues.

And from the equally estimable Dan Hachigian (though how does one estimate these things, anyway?):

The most astounding thing about the Y2K bug is that we haven’t learned anything from it. Our children’s children or perhaps one or two generations further out are going to face a crisis that makes the Y2K problem look trivial. I refer here to the “social security number bug” about which I haven’t heard anybody else saying a thing.

Essentially the problem is that social security numbers are “hard coded” to be nine digits long. Virtually every program that uniquely identifies individuals in the US makes use of this fact. Hence virtually all of these programs limit themselves to 1 billion unique id’s. Imagine the scenario that US population doubles every 60 years. That gives us less than 120 years until the name space is not large enough to accommodate a unique id per person. Recoding this field is going to be several orders of magnitude more work than Y2K.

Note the second poor assumption here, that SS #’s can be recycled, which to my knowledge is not currently done. Every year we retire something like 1/75th of the population’s id’s. We can buy some extra time by recycling them, but it will be at the expense of potentially getting ourselves confused with dead people. On the other hand, maybe it wouldn’t be so bad to end up with Bill Gates pension benefits! But, of course, I’d hate to end up with the id # of a wanted criminal!

Well, I guess it’s kind of like the environmental problem, the energy crisis, or the social security cash flow question. Most of us probably won’t be alive when it hits the fan. Wouldn’t it be astounding if the real crisis in social security isn’t running out of money but rather running out of id numbers?

Reason enough to join ZPG! Well, not really. This is only slightly less worrisome to me (and I suspect to Dan, whose tongue was poking around his cheek) than this one, from the estimable Monty Goolsby. I think even I can solve this one:

So you think the Year 2000 problem is bad, huh. Just wait ’til 2100. Most computer programmers I know determine a leap year by dividing the year by 4 and checking for a remainder of 0. If it is 0 they assume that it’s a leap year.

But 2100 isn’t a leap year and is evenly divisible by 4.

The rule for determining leap year is: Every fourth year is a leap year except that every hundredth year is not except that every four hundredth year is. Whew.

So, the year 2000 will be a leap year and 2100 will not. On March 1, 2100 the whole world will be one day later than the computers think it is. This is going to be so much fun, I think I’ll have a party. Of course, if all the enterprises begin now and fix this problem on their legacy systems while they’re working on Y2K, then their synergy will avoid a paradigm that they won’t like.

Again, we have to assume Monty’s tongue is in his cheek. But apart from the near certainty that technology will have advanced to the point that problems like this have become trivial, there’s this: the world could simply agree to make 2100 a leap year.

(But wait: maybe Monty’s right. What about the legacy software that had already been designed to do this right? Then it would be goofy. Still, I have to think that getting the date wrong by a day will be less catastrophic than getting it wrong by 100 years — the current problem.)

Tomorrow: A (not so) Cheery Year 2000 War Story from S. McGrath

 

 

Comments are closed.