01.01.00

This could well be the scenario if computer users in the public and private sectors do not wake up in time to overcome a major information technology system glitch deadlier than a virus. For, you can protect your computer against a virus coming in from outside, but, here, the disease was built into the computer when you installed it. Like a genetic flaw.

Explains Ashok Jain, CEO, HCL James Martin Inc: "To save disc space in computer hardware and thus cut down on expenses, software engineers in the '60s and '70s had allocated only a two-digit code for the year, without prefixing the century." Thus in a computer data bank, a typical date would read, say, 15/05/96. On the stroke of midnight on December 31, 1999, this date would change from 31/12/99 to 01/01/00. A computer will—naturally—see 00 as less than 99, sparking off colossal and unimaginable blunders in computing. Suppose a bank lent money to Mr X to be paid back in eight years, sometime in 1996. In 2000, when the bank checks to see how many instalments are still left to be collected, the bank's computer will neatly subtract 96 from 00 and tell the bank that the loan was not given four years ago, but will be given 96 years from now.

Well, a computer is a machine, after all, and it functions according to the programmes fed into it. Were software engineers so dumb then? Not really. They had anticipated the problem, but imagined that the old software would be replaced with a four-digit code. However, the software continued into the '80s and '90s, with 2000 AD seeming distant.

But by 1995, the problem was staring people in the face. The question that gave sleepless nights to investors, both big and small: What happens to instruments with a five-year validity at the turn of the century? Would your life-long savings put into long-term investments become negative in 2000 because the computer cannot think beyond 99, thanks to its human master?

Says Saurabh Srivastava, managing director, IIS Infotech: "The Gartner Group, a US-based market research agency, has estimated that 90 per cent of business will be hit in developed countries." But in India, where computerisation is not so widespread, between 10 and 20 per cent of business would be affected by the 2000 AD syndrome, he says. Unfortunately, Indian companies still seem blissfully oblivious. "Not a single Indian bank or financial institution has booked contracts for rectifying their systems to overcome the 2000 AD glitch," adds Srivastava.

One major reason for this could be that the job is frighteningly expensive. Every line of code that was ever keyed into an organisation's computers will have to be painstakingly checked to pinpoint every instance of a date occurring, and then corrected from two-digit to four-digit. Behemoths like the Life Insurance Corporation of India (LIC) could easily have 25 million lines of code that need to be analysed and corrected. For scanning, correction or cleaning up (conversion of 99 into 2000), then testing the systems to see if everything is okay, the charge is between $1 and $2 per line of code—that is, 10 words/figures. That's Rs 90 crore of investment for the LIC. And if conversion/cleaning up is postponed, the charges increase—$3.80 in 1997 and $8.40 in 1998, as the work pressure intensifies, with the year 2000 getting closer.

Between now and 2000, the conversion industry will earn up to $600 billion, estimates the Gartner Group. The world software industry is $350 billion currently.

Indian companies, which are competent and inexpensive in software services, have jumped into the fray. IIS Infotech which has been in the business for the last 14 months has executed orders worth Rs 20 crore, mainly in the US and UK.

The HCL James Martin joint venture, formed on April 5, has already bagged orders from a dozen overseas firms. Wipro Systems, Bangalore, claims to be the first Indian company to venture into the complicated business two years ago, in a joint tie-up with the US-based Nezup Software Inc. E.R. Chandramouli, manager, Wipro Systems, says his company has undertaken impact analysis—which gauges the seriousness of the problem before initiating clean-up operations—for half a dozen clients in US, Europe and Japan. The company also claims to be executing a big conversion operation for a major US company. On May 2, HCL James Martin bagged an order worth $450,000 for impact analysis for New Hampshire.

When then will Indian banks and institutions wake up? Past experience has shown that Indian firms are generally slower to react to a problem than their global counterparts, points out Srivastava. But Chandramouli says conversion problems mainly concern IBM mainframes or older and complex systems developed on hierarchy or network databases. Most of India went in for computerisation after the '80s when mainframes were becoming history and barring public sector undertakings, organisations moved to Unix-based, and later, to client server systems where the conversion problem was taken care of. Computers built after 1990 already had a four-digit code for the year, adds Srivastava. 

But half the computers in India would still be based on the two-digit code, industry sources estimate. "We are not even looking at the Indian market as we are already experiencing a shortage of people to address demands from abroad," admits Chandramouli. Is it because the Indian market is not lucrative enough? HCL's Jain hits the nail on the head when he says: "Which Indian company would be willing to pay $1.50 for cleaning up a line of code?"

But companies may be willing to scale down prices for the Indian market. Charges may be brought down by 30 per cent if demand is forthcoming, says Srivastava. This may induce Indian banks and institutions to clean up their systems before it is too late.

 So there is still hope that Indian banks and your life long savings and insurance policies may not crash in the year 2000. Amen.