The Paul Bray perspective: Haunted by the past

About 10 years ago, I wrote a Cobol program which included a routine for handling six-figure dates. I can still recall adding a comment line (I was a conscientious young programmer), which said: 'This routine will not work after 1999 - but, of course, all computer systems will die in 2000 anyway.'

I wish I could remember which client the system was written for. If I were an entrepreneurial type, this would be because I wanted to be paid #1,000 a day for correcting the mess which I helped to create. If I were a crook, I might simply want to blackmail the system's hapless owners.

Actually, my motive is guilt. I can't help feeling that in some way, be it ever so small, the millennium problem is my fault. The least I could do is warn the people whose computer systems I helped screw up about the innocuous-looking PIC 9(6) variables which threaten to blow the legs off their businesses, like landmines lurking beneath a Bosnian playing field.

To assuage my feelings of culpability, I have been searching for good things which might come out of the millennium debacle. I suppose the obvious one is that there will be plenty of work for Cobol programmers, especially the old-timers who ageism has thrust onto the digital scrap heap long before the big iron they used to program. But since it was their fault in the first place, this seems a tawdry silver lining for this particularly ominous cloud.

The millennium bug has reminded the IT profession and the general public about the inherent fallibility of IT systems, though our memories hardly needed jogging. It reminds us that IT people can be a bit dim, too. Did we really assume that the minicomputer systems we were writing in the mid-1980s would be pushing up the daisies before the century was out - even though we knew that the mainframes which managed our pensions and sent out our bank statements had probably been in operation for over 20 years already? Were we, despite being the builders of the future, living in the past, assuming that because things had always been done that way, it must be all right?

Or was it simply that new millennia always sound such a long way off? Even today, the phrase 'the year 2000' carries an air of almost mythological remoteness for much of the population - though not for the average IT manager. More significant is the way it brings home to us how dependent we are on IT - not just to make traffic lights change and pay our pensions and prevent us buying rotten food in the supermarket, but the fact that an increasing amount of our society has no existence at all outside the strings of ones and zeros stored on some oxide-coated platter.

The most extreme example is wealth. Although few of us keep our lolly in a pot under the floorboards like some latter-day Silas Marner, we are still reassured that, if we wanted, we could convert our bank balances and unit trusts into crinkly mauve notes proudly promising to pay the bearer on demand the sum of #20.

The prospect of our life savings and pensions ceasing to exist simply because, in circa 1985, Paul Bray coded a date routine and used only two digits for the year is quite incredible, yet theoretically possible. We could not even guard against it by all withdrawing our cash and stashing it under the mattress. The effects of this would apparently be worse than those of the millennium bug itself, causing a crash in the world economy.

I don't believe either eventuality will actually happen, at any rate in the great majority of cases. But it should at least make us remember how totally technology-bound our society is - even if it is too late to do much about it.