January 2000

"This is not the end. This is not even the beginning of the end. But maybe it is the end of the beginning." -- Sir Winston Churchill, during WWII.

Y2K has come and gone, with surprisingly few problems.

Or has it?

As expected, most of the "big things" are working just fine. When you throw a couple hundred billion dollars at a problem, it usually gets fixed. But this is only the first phase of the year 2000 issue....

Over the past few months, many people have asked me what computers would do at midnight of January 1, 2000. They thought their PC's would just quit, or blow up, or something else as dramatic as that. However, I explained that the vast majority of PC's would do something far, far worse: they would keep right on working, but some of the numbers would be wrong.

Yes, the big things like power, water, telephones, military systems, and other such nationwide and global systems appear to have been fixed. Some of the patches were implemented at the very last minute, such as the FAA's altitude tracking system for airplanes. Humorously, the Naval Observatory's clock actually moved *forward* 17,100 years to a mythical year "19,100" before being repaired. The Internet is still up and running, of course (being mostly Unix-based). And the banks and ATMs are still functioning properly.

However, there is still the "invisible majority" of millions of small-business computers to consider. Most of these people have taken a "wait and see" attitude because they just don't have the money or the time to deal with a major computer overhaul. Over the next few months, we may hear from some frustrated people who cannot reconcile their accounting records, or who have problems generating correct tax reports. Other such issues will not crop up until April of 2001, when last-minute filers may find their data does not reflect reality.

In addition, new Y2K-related bugs can occur at any time. If a Windows-based system crashes and must be reformatted and reloaded, the technician must remember to reinstall all of the appropriate Y2K patches, too. Many of the big-system patches were temporary fixes such as hard-coded date fields, sliding windows of time, or some other alteration that specifically targeted the January 1 boundary.

In reality, passing the Y2K boundary brings us into a new era of uncertainty; it is not the time to go back to "sleep", so to speak, and once again blindly trust the computers and the software companies that everything will be just fine. We have bought a few years of time, perhaps, but sooner or later the date-related issues will reappear and threaten the peace of mind that we all desperately strive for. Did you know that almost all Microsoft products have a Year 2035 problem? Did you know that C++ has a similar date-related problem? Did you know that even Unix and Linux systems have Year 2038 problems? Not to mention the Year 2029 issue that most sliding date windows have now produced.

Hopefully, people will not go back to "business as usual," but will instead begin taking a more long-term approach to system design and software specifications. However, I doubt that will happen. People are inherently short-sighted and lazy; the vast majority of today's programmers and executives plan to be retired in the year 2029, not standing in the line of fire yet again. It is quite likely they will pass the buck, just as their forefathers did.

For those who have not yet addressed the Y2K issues in their own PCs or those of their neighbors, clients, and relatives, OS/2 Headquarters will soon be making available a Y2K Fix-O-Matic CD-ROM. This CD will contain a set of patches and utilities from several software companies, so that people will be able to make the minor vendor-approved repairs to their operating systems and applications without having to waste time finding and downloading these patches. The price will be a mere $15 per copy, which is a small price to pay to have the necessary software tools all conveniently available. E-mail me at if you are interested in obtaining a copy of this Y2K CD.

Hopefully some people have learned a lesson from this exercise: software must be designed with long-term goals of compatibility, maintainability, flexibility, and especially reliability, because unlike hardware, software does not "wear out" and has no natural "life span". Software does not automatically become "obsolete"; millions of people are still chugging along with DOS and other so-called "old" products. Design for the long haul, because a little more thought on the front side will bring huge dividends in reduced costs years later. Think about ways to make future updates, expansions, and modifications easy to perform and easy to document. Notice how well the Fixpack process works for OS/2, as well as the Feature Installer and Warp-Up products.... We should start thinking of software as an organism in constant flux, not as a block of stone that is set in concrete and never altered. Maybe we will learn something useful from Y2K after all.

Most recent revision: January 30, 2000
Copyright © 2000, Tom Nadeau
All Rights Reserved.