2000 doesn't add up for computers

British businesses may have to find pounds 10bn to overhaul old program s that won't be able to cope with the new century

WHEN David Atkinson asked the Prime Minister in December last year if he was aware of the problem of computers crashing come the year 2000, John Major could perhaps be excused for thinking the MP for Bourne- mouth East was jumping on the bandwagon for making doom-laden Millennium predictions.

The problem Mr Atkinson was referring to is the inability of computer programs that work on a six-digit date system to recognise the change in centuries. When 2000 arrives, many computers will record the year as "00", or the start of the 20th century, rather than the 21st. This could possibly throw out of joint everything from cash machines reading the expiry dates on cash cards to billing systems for telephone calls.

US-based industry expert Peter de Jager says that upwards of 80 per cent of PCs could prove unreliable come the Millennium. Some information technology experts have gone as far as to paint an apocalyptic picture of computerised weapon systems being triggered off by failed date-recognition programs.

The possibility of this catastrophic scenario becoming real suddenly became dramatically clear last December, when a message originating from the office of the US Secretary of Defense was published on the Internet.

The message warned the US navy: "To ensure weapon, command and control, and information system proponents, sponsors, designers, developers, maintainers and users are aware of potential software, firmware and hardware failures associated with the arrival of the year 2000. Accordingly, corrective actions, if not currently underway, should start now."

The UK government's information technology advisory body, the CCTA, is also sufficiently concerned to be organising a big conference in April to bring together IT managers from government departments, local authorities and other public bodies to discuss the problem.

"We would like to worry the people who are not already worried," says Dr John Stewart, in charge of the project for the CCTA. The group's recommended approach to the problem will be set out in a briefing paper to be issued this March and to be made freely available on the Internet. "This will point people to other sources of information on how to deal with the issue," added Stewart.

IT consultancy firm Millennium UK, is one of several companies, including IBM, to be involved in organising the London conference with the CCTA. Les Thomas, Millennium's managing director, is greatly encouraged by this step by the Government, but remains dismayed at the general attitude of private companies.

"We believe that as many as 80 per cent of the top 500 UK firms have yet to seriously address the problem, by which we mean appoint a full- time manager to deal with it," says Thomas. At the root of this blindness is the cost.

Ian Baker, who manages IBM's European Year 2000 Services group, estimates British firms will be stumping up in the region of pounds 10bn to put their systems right. The worldwide figure for dealing with the problem has been calculated at $400bn (pounds 265bn).

Mr Stewart refuses to be drawn on cost figures for individual government bodies, but agrees that: "It is a substantial problem, which has already attracted attention at the highest levels of government." He reckons the total bill could be over pounds 100m.

Financial services companies are the main firms to be hit by this "Millennium bug". Banks, including Abbey National, Lloyds and NatWest claim to be on top of the situation and Barclays' David Scott has said the cost to the group should be no more than pounds 3m.

Others in the industry feel that this reveals Barclays either know a lot about the problem, or very little. A programming manager for another large bank says that his company will be lucky to escape a final bill of pounds 25m.

Most of that cost is accounted for by the number of staff who need to be hired to deal with the problem. Mr Baker explains: "Let's say a bank has a typical 40,000 programs to check and that each check takes a day. That's 40,000 days' work which realistically has be undertaken in the next two years or 650 working days."

While firms may be able to farm out the program writing and a little of the testing to countries lsuch as India that specialise in cheap computer programming, the actual testing - accounting for half the work - needs to be done in-house.

Most of the work will be focused on the "legacy systems", or application programs written for mainframe computers, some as long ago as the 1960s.

Keith Valentine, at NatWest, points out that these have evolved over time and that all of the changes made inside the programs will have to be checked in the context of the Millennium.

No wonder the insurance industry, itself busily checking its own systems for Millennium glitches, is turning its thoughts to the implications of firms claiming compensation on their business insurance for computer failure.

Phil Ward, at the Association of British Insurers, says that insurers will be instructing their inspectors to check up what bus- inesses are doing to address the problem during annual risk assessment visits prior to the renewal of policies.

As if all this was not enough to be dealing with, financial services firms are wising up to another pre-Millennium computer problem: European Monetary Union.

If a 1999 timetable is adhered to for conversion to a single European currency, even by a few countries in the EU, banks are going to have to start preparing new computer systems now.

Already working flat out to check revised 2000 programs, overstretched IT managers, just like a previous Prime Minister, could end up saying "No, No, No" to the Euro.