Newspapers and magazines have spilled a lot of ink lately about the impending year 2000 crisis - so much that it's hard to avoid getting caught up in it. The problem, as you've probably heard by now, is that many programmers used only six digits to represent dates in their programs: two digits for the day, two digits for the month, and two digits for the year. All over the world there are thousands of computers that think today's date is 981008. Although these programs will work great today and tomorrow, many of them will mysteriously fail 449 days from now, when the calendar rolls over to 000101.
Is Y2K a real problem, or is this just another case of millennial hysteria? When I wrote about the Y2K problem on Jan. 1 of this year, I said the problem was real, but that many businesses and most home computer users would probably pull through OK. That column generated a torrent of e-mail from Y2K evangelists, who said I was being irresponsible by failing to sound the alarm.
``Yeah, it's not much of a problem for the average home PC user,'' wrote Ed Yourdon, a well-known Y2K consultant, in an e-mail message to me. ``After all, if the lights are out, the phones are dead, and the banks are closed, most people won't have time to worry about whether their PC is compliant or not.''
Of course, Yourdon isn't an unbiased observer: He's made a lot of money writing and speaking about the year 2000 issue, so he has a reason to play up the potential dangers. On the other hand, the dangers could very well be real, and Yourdon could be doing his honest best to try to bring attention to the problem.
Yourdon's analysis is pretty straightforward: If there are sporadic problems, such as a failed power plant here, or a bank that's gone belly-up over there, then things will probably be OK. But if problems are widespread or sustained, it could take the economy a while to get going again.
``If it's only one or two banks, the FDIC will have no problem handling it; if it's a couple thousand banks, then it's much more problematic,'' he says. ``Again, I look at this from a statistical perspective, and from my experience as a software engineering consultant. There are 11,000 banks in the US, and they all have a Y2K project with the same deadline. What percentage of them can we reasonably expect to be finished in time? '' (You can read more of what Yourdon has to say at www.yourdon.com.)
With this in mind, I called Boston Edison and spoke with Jim Ashkar, its Y2K coordinator. Ashkar said many of Edison's systems are ``Y2K compliant'' now - the company's IBM mainframe computer was recently replaced with a new system that can handle dates into the next century - and those systems that aren't will be up to snuff by July 1999.
Interestingly, one of the most difficult systems to upgrade was Edison's billing system, hence the need for the new mainframe. ``We started back in 1993 working on this,'' says Ashkar. ``We also started restructuring our contracts with our major business suppliers - making [Y2K compliance] a condition of maintenance contracts.''
Basically, Edison is ensuring its systems will be year 2000 compliant by requiring that its suppliers make them year 2000 compliant. It's then asking the companies to test the systems and provide the results as proof the systems will operate as expected.
Edison will have extra engineers standing by, if needed, to man the stations on the night of Dec. 31, 1999, ``in the event that we need to take some additional action,'' says Mike Monahan, a company spokesman.
``This is not instead of being ready,'' Monahan stresses. ``And I would add that, for all of our critical systems, we have made a commitment not to accept anyone's word for readiness: We will do [our own] physical testing.''
More information about the region's electrical outlook for the year 2000 can be found on the home page of the New England Power Pool, at www.nepool.com.
The devilish thing about this year 2000 business, though, is that what happens over the next 14 months will depend upon people's perceptions just as much as on the underlying physical reality. If large numbers of people think that the year 2000 problem will make banks fail, and if they withdraw their money as a preventative measure, then banks will suffer. People's reaction to the potential crisis might cause as many problems as the crisis itself.
``We are entering the era of unpredictability,'' says Peter G. Neumann, a scientist at SRI who has made a career studying the year 2000 problem and other computer-related risks. ``The folks who say `Everything is under control,' `Trust us,' `Don't worry,' are likely to have egg on their collective faces. The doomsayers also. There will be lots of problems. ... Railroads are reportedly in trouble because there is no manual backup system and all automatic control is centralized in a single system. ... Air traffic control could have problems - [there's] lots of very old legacy software.''
Neumann thinks most banks will probably survive ``because money is at stake, and money talks.'' But we are in for surprises, he says, as various Y2K projects are completed and people discover that different strategies for dealing with the problem don't work in concert. ``Things that work locally often don't work when they are linked together with other systems,'' he says.
Want to learn more about the year 2000 problem? Come to a panel discussion Tuesday, 7-9 p.m., at the main branch of the Boston Public Library. Hosted by the Center for Millennial Studies at Boston University, the panel will feature Peter Neumann and three other experts who will discuss the social implications and computer problems posed by the year 2000 bug. For more information, visit www.mille.org.
Technology writer Simson L. Garfinkel can be reached at plugged-in@simson.net.
Dow:
