Microsoft's British research arm is looking into what kind of software scientists will require in the future.

The company has brought together 40 leading scientists to a meeting in Venice to discuss their needs.

The researchers have been telling Microsoft about their growing problems of information overload.

The aim of the meeting is to investigate ways of creating new tools to cope with the vast amounts of data generated by scientific investigation.

Proposed sub-atomic physics experiments will generate hundreds of billions of pieces of information.

And there is a similar problem in biology as scientists struggle to decide what the mass of data generated by the human genome project really means.

Many in the scientific community believe that they will need new specialist software to help them sift through and interpret this bewildering array of information.

New science

According to scientists, versatile data bases and intelligent search engines are likely to take over from the DNA sequencer and the mass spectrometer as the main tools of scientific investigation in the 21st century.

The challenges facing scientists have been outlined by the man behind the initiative, Stephen Emmott of Microsoft Research.

"By 2020, science will, I claim, be in the process of a profound transformation as a consequence of the emergence of 'new kinds' of science'," he wrote in a paper entitled Towards 2020 Science.

"For example, advances in areas such as computational systems biology could re-shape the health and pharmaceutical sectors as a result of a fundamentally greater understanding of biological processes, and therefore of disease.

"Advances in artificial chemistry and nanoscience could create entirely new technology.

"It is highly likely that 2020 science can create a new economic era of science-based innovation, whose importance could dwarf that of the technology-based innovation era of the 'IT revolution'," he wrote.

The discussions are still are a very early stage and any new software designs from Microsoft are still a long way off.