By Jack Schofield

The government has a lot more to worry about than passports when it comes to the public's perception of their computer systems

The big projects NHS Direct: public access to online healthcare advice Inland Revenue: tax returns and Vat registration over the internet Northern Ireland: online booking for vehicle testing, licensing and insurance Education: National Grid for learning; information network connecting all schools Social Security: ACCORD project uniting all personal data supplied by public in single accounts Post Office: cross-the-counter services go online Jack Schofield

There are three tests of a project's success: whether it's on time, whether it's under budget and whether it works. Most companies are happy if they can get two out of three. In computing, however, one out of three is common and a significant number of projects - there are no hard numbers, but some think it's more than a third - are simply abandoned.

If this is the situation in commercial computing, is it any wonder that so many government computer projects seem to have problems? In fact, rather than belabouring the passport agency for what is being seen as another computer disaster, the technology installed by Siemens Business Systems may well prove to be a success. The agency did win a government charter mark earlier this year, partly for customer service.

Rather than blaming computers, perhaps we should be blaming managers who shed staff too soon. What they should have done is run old and new systems in parallel until they were confident that the new system could take over. However, it costs more to do it that way, and that's hard to justify if a system has been installed to save money rather than increase the level and quality of service. Even commercial companies sometimes make the same mistake, as Anita Roddick confessed at the Body Shop.

Whether government computer projects are actually more prone to disaster than commercial ones is hard to say. We know about the gov ernmental failures because government is publicly accountable. MPs, the national audit office, trade magazines and others are keen to expose them, mindful of the fact that failures affect the public, as happened at the driver and vehicle licensing centre, some council tax systems, the London ambulance service despatch system and the passport agency. By contrast, commercial operators and technology suppliers are only interested in trumpeting their successes, and usually bury their failures as quietly as possible.

There are reasons, however, why governments and local authorities could find it harder to succeed than other types of organisation.

The main problem with projects of any sort is that big ones - power stations, tunnels, bridges, tube extensions - are harder to manage than small ones. By their nature, governments are more likely to attempt big-bang projects that affect whole populations. It's hard for them to test tax reforms in local areas the way companies test new soap powders or chocolate bars.

Large computing projects are even more difficult than other types because programming - creating the software that actually does the work - is not an engineering discipline, though the industry likes to pretend that it is. Computer programs are mostly still written the same way they have been for 40 years - by hand - and individual programmers are much more like creative writers than like brickies. The civil service has no way of supporting or rewarding superstar coders who can become millionaires in the commercial world.

Government computer projects also seem to suffer more than commercial projects from cultural problems, including what's known as the `optimist effect'. Given that most computer programming projects are going to go wrong most of the time, the most important thing is to find and fix errors as soon as possible. The longer an error is left, the more expensive and time-consuming it is to fix, which is how $10m systems end up as $100m runaways.

Programmers must therefore keep drawing attention to their failures, and be rewarded for doing so. This is a tough proposition for commercial companies, and almost unthinkable in adversarial (them and us) or fear-based cultures. Management gets the impression that the project is going well and has no idea what's actually happening at the grassroots level. By the time they find out, it's too late.

A third problem with government is that its operations are just too slow: the computer industry moves at such a breakneck pace that government systems may well be out of date or based on obsolete equipment be fore they are even delivered. Michael Dell, founder of Dell Computer, used to have a problem with government orders for computers that the company had already stopped making. (`I just used to send them a faster PC and a refund,' he told me.)

Commercial companies have the same problem, and have learnt that it's pointless to embark on five-year programming projects when they don't even know what products they'll be offering in six or nine months' time.

Governments, however, can still spend two years discussing whether to do something, and another two years on planning and reviewing the project, before there's a change of government and it's cancelled. By the time they get round to implementing something, the rest of the world has usually moved on. Known problem, so why not have an `intercept strategy' that will aim at where the technology will be in five years time? Great idea, if you think organisations that can't cope with today's technology can implement `blue sky' ideas based on guesswork.

Government's final problem is that it can't just ignore technology, because it's not going to go away. It needs to be able to deal with commercial organisations that are already trading electronically, and it needs to deal with other countries - particularly the US - that are starting to do as much as possible online.

Government also needs to be able to deal with citizens electronically, because that's what its customers are coming to expect as a result of experience with the best commercial organisations. Why can't you collect your social security payments or child benefits from a cash machine? (You already can in parts of Spain.) Why can't you pay your council tax at the nearest supermarket? Or file your tax return or vote over the internet? And if your government can't provide a convenient and efficient 24-hour service, why should you pay it half your earnings?