A world expert in software engineering who helped discredit former US President Ronald Reagan's controversial "Star Wars" defence plan will head a €6.4 million research programme at the University of Limerick.

The five-year programme, led by Dr David Parnas and grant-funded by Science Foundation Ireland, is to begin in October.

It will focus on improving software design to make it less "buggy".

One of the first big computing initiatives funded by Science Foundation Ireland, the project will employ five further researchers and fund four PhD students.

It is intended to establish the State as a stronghold in good software design.

"An investment in Dr Parnas is really an investment in Ireland's future," said Dr Bill Harris, director of the foundation, at the programme's launch yesterday in Dublin. He is a person of world-renowned reputation for his accomplishments."

Currently director of the software engineering programme at McMaster University in Canada, Dr Parnas is author of a widely used university-level textbook on software design and has received several industry awards.

In a long-standing battle against bug-ridden software, he has pioneered approaches to programming and to standardising the way in which software design is documented. His concern with the general unreliability of software led him to resign from a committee advising President Reagan on the "Star Wars" initiative to place anti-missile devices in outer space.

Dr Parnas's emphasis on the potential dangers of such a system due to software glitches is considered one of the key reasons public sentiment turned against the programme.

"Dr Parnas has had an influence on how software engineers think and work and act, and how they're perceived by the public," said president of UL, Prof Roger Downer.

Dr Parnas said he intended to work closely with the software industry, particularly smaller companies, and was buoyed by the positive response he had already received from Irish industry.

UL was especially attractive to him because it was supportive of giving him the right environment to further his research and had many researchers in complementary areas, he said.

"The next frontier in software is not more features or greater speed. What we most need is improved reliability," Dr Parnas said yesterday.

He said the future belonged to those who can develop "reliable, consistent" software programs - and he hopes to make the State a centre for such design.

"My goal is to leave this institute as something that runs without me," he quipped.