As we approach the new millennium, the Year 2000 problem attracts more and more attention. The bug's potential devastating effect on our technological infrastructure has been documented only too well. Fear over the collapse of electronic systems on January 1, 2000 may be justified--but the solution could turn out to be a boon for hardware spending in the near-term.
The Y2K problem stems from the days when programmers, in an effort to conserve scarce memory space, decided to program the internal date function on computers to store just the last two digits of a year, assuming the first two digits to be "19." The date function on older computers, like mainframes and even some Pentium processor-based PCs, suffer from this problem.
Since software applications frequently need to access the date function in a computer system, we are presented with a significant problem--the ramifications of which could be enormous. For example, if there is one non-compliant system in a network, even while other systems are "good" systems, the problem could still potentially cause havoc.
The pervasiveness of computer networks today means that there is no scope for even one non-compliant system or a piece of non-compliant code. The dimension of this problem is clearly huge.
The question, therefore, is how to make these systems compliant--and what will happen to technology spending in 1999 as a result. Here the answers are varied.
There is the scenario that the most of the money in 1999 will be spent on consultants to fix the Y2K bug, with the funds coming from hardware budgets. But I believe that hardware buying is actually likely to accelerate in 1999.
Although the United States is the furthest along in terms of Y2K compliance, it has mostly been focused on the enterprise level, and not desktop level. Europe, so far, has been preoccupied with its currency unification--which as a byproduct would involve Y2K fixes--but next year should solely focus on millennium issues.
Meanwhile, Asia has been focused on staying solvent, period. With some economic stability returning to Japan, however, corporations will be in a mad rush to fix the problem that had been basically ignored during the currency crisis.
The quickest fix--given the limited time frame--and also the most cost-effective fix, is to replace older 486 and Pentium-based systems with PII/K6 generation hardware at a price of $500. The replacement not only solves the problem quickly, but also upgrades the hardware as well.
I think the Y2K problem will lead, however, to a slowdown in large, time-consuming project-based implementations, such as ERP. There is anecdotal evidence to suggest that both these trends are already taking place--as suggested by PeopleSoft on one side, and Intel on the other.
All in all, the basic human tendency of procrastination, combined with late-in-the-game panic, should fuel hardware spending in 1999, leading to the first increase in PC unit demand since 1995 for the new year.