The 21st century doesn’t actually begin until January 1, 2001, but for most people, January 1, 2000, seems more significant, just because of the change from writing 19 in front of the year to writing 20.
That little change is even more significant for businesses that rely heavily on computers, because computers, too, are used to assuming a 19 in front of the date: and unlike people, computers can’t change their assumptions at the drop of a New Year’s Eve noisemaker. (People aren’t very good at it, either, really—expect a lot of confusion when it comes to writing the date on cheques!)
Until recently, most computer software that required a date to be entered asked for only two digits. But a few years ago, someone finally realized that the 1900s are coming to a rapid conclusion, and asked, “Wait a minute. What happens in the year 2000?”
Computer experts considered that question, and became alarmed. What happens in the year 2000, it turns out, is that programs suddenly start spewing out garbage–or quit working altogether. Come the year 2000, for example, a company might not be able to process its 1999 year-end billing or payroll probperly; corporate credit card holders might be refused transactions because their accounts appear to be in arrears; year-end profit data might not calculate correctly, and utility companies might cut off services because of bill payments apparently decades late.
That’s because a program that uses only two digits to calculate the date is suddenly going to find itself faced with dates that are almost a century old, from its point of view. Suppose you were born in 1935. In 1999, a computer needs to calculate your age, and simply subtracts 35 from 99, getting the correct 64. In 2000, the year your pension is supposed to start, that same computer calculates your age and subtracts 35 from 00–and you end up minus-35 years old. If the sudden appearance of a negative number doesn’t cause the computer to crash, at the very least it’s going to assume your pension isn’t due for another century. You probably don’t want to wait that long.
In one corporation, Information Systems did an internal survey (which alone took 10 weeks) and found that of 104 systems critical to their operations, 18 would fail in the year 2000. Those 18 systems were made up of 8,174 individual programs and data-entry screens plus some 3,313 databases.
The computer industry calls this the “Year 2000 problem,” although that’s a bit misleading: it first cropped up as early as 1970, when software that was calculating amortization of standard 30-year mortgages started crashing. The good news is that banks and insurance companies and other businesses that deal with calculations stretching into the next century have already had to deal with this problem. But as we get closer and closer to the actual date, other corporations and businesses that never even thought about it are suddenly finding it affects them, too.
Many of the problems arise in so-called “legacy software,” software created many years ago, usually on large mainframe. Since the programmers were desperate to save data and program storage space in those relatively primitive days, the year 2000 was still a third of a century away, and they couldn’t imagine that their programs would still be in use decades later, they used only two digits to represent the date. This had the added advantage of reducing the number of keystrokes needed to enter data.
However, the problem isn’t limited to legacy software. Even some recently developed mainframe and desktop programs can run into difficulties, due to shortsightedness on the part of the programmers (or, if you want to be paranoid about it, a cynical attempt to ensure that everyone will have to buy upgraded versions of the software before the turn of the millennia).
The problems are fairly easy to correct on a database-by-database and program-by-program basis; the trouble is in the sheer number of affected programs and data files, and the time it takes to change each one.
Here in Regina, IBM Canada has just officially opened its national Year 2000 conversion centre, which already employs about 50 people, with prospects of tripling or even quadrupling that total over the next couple of years. C from across Canada send in their computer programs to the centre. Using special software, the centre’s employees examine those programs for Year 2000 problems, and rewrite them as necessary. Shelley Smidt, manager of the centre, said it’s already seeing an increase in the volume of customers, and they expect it to get busier and busier as 2000 approaches.
Year 2000 conversion work should really be done by the end of 1998, so that all of 1999 is available for testing, Smidt says. “People have waited longer than they should have.”
The fact that eventually 200 people will be working at this one conversion centre in Regina alone gives some idea of the magnitude of the problem. Putting a dollar figure on the total cost of converting systems in all the world’s businesses and corporations is impossible, but you can bet the figure is in the billions, because by one estimate, just the 50 largest corporations in the U.S. will spend between $50 and $100 million each.
And you just know that in the back of every programmer’s mind is going to be that niggling question, did we catch ALL of the Year 2000 problems?
On New Year’s Eve, 1999, watch your fellow guests. The ones holding their breaths when everyone else shouts out “Happy New Year!” are the computer programmers.
And the next morning, the whole world may have a digital hangover.
