Harbingers of Y2K

Software Magazine, June, 1999 by Ian S. Hayes

Everyone always asks me my predictions for the Year 2000 transition. I wish I could give a definitive answer. I feel like a weatherman trying to predict the impact of an approaching storm many days before its arrival. Will the storm reach land at maximum intensity, causing a wide swath of destruction? Will it turn offshore, merely brushing us with some winds and heavy rain? Or will it peter out entirely? The answers depend on the efforts and preparations of many companies, governments, and individuals. Have they taken the problem seriously enough? Did they have enough time to complete their critical projects?

It is hard to know what reality is, and to predict what will be, given the hype and conflicting reports from the media, companies, and government circles. The "official" story is always, "We're right on schedule and anticipate no problems." The rumor mill always says, "Armageddon is coming." Consultants who are working on the problem say nothing, silenced by nondisclosure agreements.

I wish I could accurately predict the future. The information would be extremely valuable. It would help determine corporate contingency plans and the amount of money to invest in preparing for those contingencies. It would help me determine what my personal Year 2000 preparations should be, and would certainly give me a wonderful investment strategy.

If I consider the situation, what I hear from my clients gives me hope. On the other hand, what I hear from callers and emails scares me. How can I cut through these conflicting messages to find out the true state of affairs? If I could, I'd be rich.

But, maybe there is a way that I can get a hint of what's to come.

Watching the Signs

Our predictions of a storm's track and intensity become increasingly accurate as its arrival date draws closer. Barometric pressures begin to change, the seas get rougher, birds and animals start to behave strangely, and winds slowly increase. One factor alone is not enough for an accurate prediction, but a series of factors together gives us a sense of a trend. Watching the rate of change in each factor over time predicts intensity. For example, we expect winds to increase before a storm, but we have a very different situation if wind speeds double hourly while the storm is still well offshore than if the winds build gradually.

What factors should we watch for our Y2K predictions? I suggest the following:

* the rate of publicly reported errors;

* SEC data on company progress;

* audit data; and

* general public reaction to Y2Khype.

The rate of publicly reported errors is our anemometer; it measures the force and velocity of our Y2K wind. We can think of the Year 2000 problem as resembling a bell curve. Software and hardware that process future dates encounter 21st century dates before the Year 2000. Remember those notorious credit card failures. Credit cards with three-year expiration dates began failing as early as 1997 because systems weren't prepared to handle new cards with "00" expirations. These system failures represent the upward curve. Conversely, systems that look to the past don't start to fall until they actually cross into the next century. The top of the curve occurs on January 1, 2000. But it takes time for that bell curve to build up and may take even longer for it to wind down. Watching the rise and fall of publicly reported failures should let us follow the progress of this curve.

These failures are easy to find and follow. They are visible because they usually involve direct public impact. Public failures cannot be hidden, covered up, or quickly resolved behind the scenes.

Companies are forced to acknowledge these failures, and the media thrives on them. These failures are the tip of the iceberg and represent hundreds of other, nonpublic failures that are quietly buried or diffused. Companies do not exactly relish negative press coverage and can hardly be expected to alert the press whenever they encounter Year 2000 problems. Think about it this way: Do you know of any Year 2000 failures that percolated to the surface in January 1997 or January 1998?

If the number of public failures remains low and builds slowly before the Year 2000, we can infer that:

* Companies have worked on the Year 2000 problem;

* They performed a thorough analysis, paid attention to "time horizons to failure," knew which systems would fail in advance of 2000, and fixed those systems in the tight order;

* They fixed their systems correctly; and

* They performed adequate testing.

Or:

* Companies are very skillful at problem recovery.

If the contrary proves true; that is, the number of public failures rises dramatically in 1999, then we can only imagine the number of failures that are occurring behind the scenes. If this happens, we will know that many companies are inadequately prepared, and that little time remains to fix the situation. We will speculate that an inability to fix and avoid these larger problems means that hundreds of smaller problems may be bungled or simply not found. We will have to ramp up our contingency planning efforts, and be prepared for some major upheavals in our personal lives.