Flawed DSS computer system may cost pounds 50m

Click to follow
The Independent Online
THE Department of Social Security has tried to cover up the mismanagement of a computer contract that could cost the taxpayer as much as pounds 50m. The DSS and its private sector contractors are arguing about who is to blame.

A key function of the system was to give ministers a wide range of benefits statistics on which to base parliamentary statements. Inquiries by the Independent and Computer Weekly show that the DSS is considering dropping the project, known as the Analytical Services Statistical Information System (Assist), which was also to be used in policy formulation.

Leaked internal documents show that Assist is four months late and trials have found dozens of errors in the software. Civil servants who are expected to use the system have complained that it does not even meet the standards of the system already in operation.

It was commissioned in January 1993 from two of the largest private computer companies working in the public sector: ICL and Hoskyns. The cost was put between pounds 35m and pounds 50m.

During early performance trials on the system, 275,000 social security records were unintentionally altered. In final acceptance tests, there were 'some 91 incidents raised'. Software deficiences gave rise to at least a quarter of the problems.

The Government's computing agency, the CCTA, has stepped in and is anxious to keep the matter out of the public eye. A letter from an official to the DSS says it 'is essential to . . . prevent systems failures and potential bad publicity'.

The department has ordered a second trial of the system. Documents show that officials may 'terminate the project at any time'.

A spokesman for the project managers, the Information Technology Services Agency, which oversees computers in the DSS, said: 'We won't deny there have been problems. This is a very complex system.' A spokesman for ICL added: 'You will get hiccups in something of this kind; and we have had hiccups.'

Catalogue of disasters, page 3

Leading article, page 15

Comments