One real problem that must be addressed is the "year 2000 bug." This bug is not a virus, nor some kind of time bomb. It is, rather, a deficiency in the computer code of many systems designed early in the computer age. In the 1970s, when mainframe computers reigned supreme, programmers realized they were in a rapidly changing field. Systems were designed in the anticipation that they would be upgraded or completely replaced within 7 to 10 years.

What these programmers did not account for, however, was the institutional resistance that developed to phasing out the mainframe. Although many organizations are now based on more decentralized, networked systems, mainframe computers continue to remain at the heart of many organizations, including large banks and government departments.

These old systems were never designed to keep running until the year 2000. The old computer programs running on them recognize dates in a two digit format. For example, the two digits "98" represent the year 1998. When calendars change to January 1, 2000, these computers will move the date to 01/01/00 which is, as far as the computer is concerned, January 1, 1900. Since most of what computers do involves executing mathematical formulae in sequence, throwing out these sequences will, in most cases, cause a computer system to crash.

Even if, somehow, a system doesn't crash, the math errors that it will produce would be problematic. For example, interest payment calculations would be in error by 100 years. Or, to cite a more personal example, you may find that your driver's license had prematurely expired. Combined, these bureaucratic problems could cause many business and government functions to grind to a halt.

Governments are having to spend a lot to fix the year 2000 problem. In the US, the Clinton administration recently announced that more than 12 federal agencies will be redirecting hundreds of millions of dollars in their technology budgets to solve the problem. A recent review of 24 US government agencies indicated that 7 had made "insufficient progress" in fixing the problem, while 9 remained "of concern," and only 8 had the problem "under control."

Representative Stephen Horn, a California Republican, is chairman of a Congressional subcommittee that is closely following the issue. His independent analysis of agency progress indicates that at their current pace, 10 agencies will finish the year 2000 conversion of their most critical computer systems on time, while 14 will be late. The worst offenders on the list were the departments of Labor and Energy. At last quarter's pace, Horn concludes those agencies won't finish their year 2000 conversion until 2019.

The financial sector in the United States is also concerned. The President of the New York Federal Reserve, William McDonough, has already asked his colleagues to cancel all vacations during January 2000 and avoid any travel during the month, illustrating the depth of his concern over the potential disruption. In addition, concern that the bug's impact on banks could disrupt financial markets has prompted proposals that markets should remain closed on Monday, January 3, 2000, to give computer specialists a chance to solve any unforeseen troubles.

All the while, the deadline is approaching. To properly implement a year 2000 solution, it is prudent to run a system through an entire calendar year of testing. This means that the latest a solution should begin testing is January 1, 1999, which is now just 12 months away.

Computers are involved in many more aspects of our lives than just banking. For example, computer systems are embedded in elevators operating in high-rise buildings, and in security systems. Will you be able to turn off the alarm on the first Monday in January, 2000, when you go to work? Other areas of concern include telephone systems, Internet routers, and international air control.

Canada is no exception. The recent TRAC survey of the Canadian insurance industry indicated that only 35 percent of all responding property and casualty companies are year 2000 compliant, and only 18 percent of interviewed life insurance firms say they are ready. However, that this group sees the year 2000 problem as significant enough to commission a survey should perhaps provide some solace. Industry in Canada is largely aware of the problem and is taking steps to remedy it. For example, almost all of Canada's financial institutions are well on the way to being year 2000 compliant. In contrast, all indications are that government remains well behind industry.

Some computer consultants are preparing for the potential meltdown by isolating the government systems they are responsible for from the systems that aren't year 2000 compliant. In other words, to minimize their own liability, programmers are setting up their systems assuming that the government mainframes will have problems. Although federal officials have been tight-lipped about the issue, anecdotal information suggests that departments such as Supply & Services, Canada Post, and even the Department of Defence have a great deal of work still to do. These problems are also shared by Canada's provincial, territorial, and municipal governments.

One potential policy solution that should be considered is privatization. Critical government operations that have legacy systems with year 2000 problems could have these functions sub-contracted to private firms that have compliant systems already running. Not only could this prove to be much more efficient than the current delivery of service, but it would circumvent the huge cost of getting old systems up to standard. The need for a rapid response to this situation means that out-sourcing must be considered seriously. Governments' inability to respond to a long-known computer bug until it is threatening their operations is evidence enough that the existing structure is not adequate to manage important systems implementation.

As for private industry, both the buyer and seller should beware. Many firms now include their year 2000 compliance progress in their annual report, and it is only prudent to make sure that your bank, credit union, or trust company is fully prepared. There are solutions to the year 2000 problem, and as long as a firm takes prudent steps to assess the potential impact, and search for solutions, January 1, 2000 should pass without incident.

As with so many past doomsday predictions, it is possible that fears about the year 2000 bug will not be realized. However, mounting evidence suggests that if any institution is going to falter in the wake of this bug, it is most likely to be big government.

References

Enright, Greg. "Are Insurers in Hot Water?" Computing Canada, vol. 23, issue 24. November 24, 1997.

"Government to spend big to fix year 2000 computer problem." Associated Press. CNN Interactive. Decem<%-2>ber 11, 1997. http://www.cnn.com/<%0> <%8>TECH/9712/11/y2k.bug.ap/<%0><R>index.html

Nielsen, Peter. "Fed's McDonough braces for January 2000." Reuters. TechWeb News. December 11, 1997. http://www.techweb.com/wire/story/TWB19971211S0004

Smith, Jerry. Phone interview. Conversions 2000 Software Ltd., Calgary, Alberta. December 15, 1997.