Back in the last century when it was widely believed that computers could do everything, the government of the day would embark on projects such as the holes in the road project.

It was a simple idea. Build a database of all scheduled utility works that involved digging holes in the road. Then each time a water company wanted to close off a road to upgrade a water pipe, the gas, telephone, electricity people and the local council could schedule their works to take place at the same time.

It was the kind of project that computers were meant for. It was also the kind of project beloved of suppliers with lots of expensive consultancy up front and no actual systems to deliver. It never happened. Competing utility companies refused to share information and it became known as the pouring money into the holes in the road project. It was scrapped quietly within a couple of years and £30 million (US$60 million) later.

The premise was that money could be saved by avoiding duplication of effort. Adding some structure to the information would lead to controllable and rational behavior. This sounds a little bit like the government's shared services mantra.

Another new dawn

Thankfully all that feasibility study nonsense is over. From now on we're told, public sector IT will be based on systems that are standardized and services that are shared across central and local departments and agencies. But are shared services delivering what they ought? Or are government departments being side-tracked into self promotion of their own chosen shared services in competition with other departments? Shouldn't public sector CIOs be focusing on putting the IT and other systems in place that do the job in hand, which is delivering better services to the public, not selling their services to each other?

Proponents say that shared services are about fulfilling the government transformation agenda through de-duplication of infrastructure and processes. The system works by first building a framework, letting departments, agencies and councils choose which parts of it they wish to use and then serving them with repeatable processes. The idea of shared services for transformation is anathema to some people. "Government transformation is a mirage that will have to be abandoned," believes Ross Anderson of the Cambridge Computer Laboratory, an outspoken critic. But whether Professor Andersen is correct or not is certainly not going to stop them trying.

The issue appears to be the number of shared services schemes out there and also what exactly constitutes a shared service. Two government departments using one water cooler could be an example of a shared service.

The shared services agenda is also a competitive marketplace agenda. The Department for Work and Pensions (DWP) is currently assessing its own shared services efforts because it wants to sell them on to other parts of government. David Myers, director of shared services at the Home Office, is on the record as saying that: "My team's first goal is to share services within the Home Office, and over the next two years we will turn our platform into a fully-fledged business for other parts of government." The shared services Myers is responsible for includes information technology.

And flex and stretch One of the most high-profile shared service attempts is Project Flex. Flex is a deal struck between the Cabinet Office (CO) and the Fujitsu government business division, after it put its ICT Service Acquisition and Change project out to tender in 2006. The CO also has Pygmalion, which is a shared services framework. In legal speak the CO provides overall 'client side' framework management while Fujitsu delivers the service. In effect, the CO acts as a promoter for Flex to other government departments as well as being a user of the framework itself. What Flex covers is a framework for supply of desktop hardware and software including thin and fat client devices, laptop computers, service desk provision, network management, security, disaster recovery and flexible