A soldier of IBM's engineer army

When he first began working for IBM in 1967, Mr Nicholas Donofrio designed computer chips and processors with less power than those used in the simplest digital watch today.

Times have certainly changed. Nowadays, the 33year-old veteran of IBM, also known as Big Blue, chairs the company's Corporate Technology Council and reports directly to chief executive and chairman Mr Lou Gerstner on quantum computing, holographic storage, and other cutting edge innovations produced in the company's eight international laboratories.

Mr Donofrio, IBM's senior vice-president of technology and manufacturing, stopped off in Dublin this week on his way to present the third annual Turing Lecture in London. Sponsored by the Institution of Electrical Engineers and The British Computer Society, the prestigious talk is named in honour of British mathematics pioneer Alan Turing, who cracked the Nazis' secret Enigma-coded messaging system during World War II.

Being part of IBM's engineer army in the 1960s when the computing industry was in its infancy had its own haphazard charm, Mr Donofrio says. "It was fun because nobody knew what they were doing and we could make up the rules," he says with a grin. Now research at the computing giant involves some 3,000 people, has resulted in 2,800 patents, and occasionally hovers on the edge of the truly bizarre.

For example, IBM has produced a working quantum computer composed of five atoms, invented a process that enables data to be stored holographically, and designed a tiny atomic "corral" into which individual xenon atoms can be herded in a specific order.

Such pure research innovations - which are not expected to end up on store shelves any time soon - eat up about 20 per cent of IBM's multi-billion dollar total research and development budget. Mr Donofrio says the rest results in products that "get out the door". IBM wasn't always so efficient, though. "It wasn't that way 10 years ago," he admits. "Ten years ago we lost one out of every four dollars we spent."

Unsurprisingly, that time period coincided with IBM's worst performance as a company, a bleak era when many in the industry wondered if Big Blue would survive. "We were inept. We became irrelevant. Well, we were relevant to ourselves but not to anyone else," says Mr Donofrio, because IBM wasn't delivering the technologies customers wanted at the right time.

It was a difficult period, after years in which a seemingly invincible IBM set the computing agenda for businesses, and businesses followed it or risked becoming irrelevant themselves. But the company successfully re-engineered itself under Mr Gerstner's direction to become, once again, a dominant force in the technology industry.

Mr Donofrio, whose Turing lecture centred on the topic of technology innovation and the new economy, expects the computing industry to drive progress along a curve "on the order of six magnitudes of improvement" over the next 30 years, just as he claims it has done during the past 30 years.

"Most of that improvement is going to come to the front of the house, where end-users interface with it," he says, a development he calls "natural computing" . Terms such as "computer literate" will no longer have meaning, according to Mr Donofrio, because they won't matter. With future computing devices, "you talk, it listens; you look, it sees; you feel, it reacts," he says.

Mr Donofrio says he believes the technological development that will have the most far-reaching consequence in the next decade will be voice-recognition. "I think voice is a big deal," he says. "But voice now is going to be voice in the network and not voice in the device." IBM is calling these future developments of voice recognition "e-voice".

"You'll literally be able to pick up a phone and talk downline to the server. You'll talk to the software that's running the infrastructure and you won't even know it. Voice has been limited by the power you hold in your hand; it has to be at the end of the line," he says.

Atomic-level computing is a more distant innovation that will transform computing perhaps 20 years on, when circuits are expected to be on a molecular and atomic scale. IBM has created a quantum computer that harnesses together five fluorine atoms and sets them to work together as quantum bits, or qubits, to form the computer's processor and memory.

The experimental qubit-powered computer proved that quantum computers could solve certain enormously complex equations exponentially faster than today's silicon-powered machines.

But Mr Donofrio admits that all these innovations still have problems today. The qubit computer needs an entire room of technology to support it. Voice recognition is, if not in its infancy, then still in early adolescence.

The predicted skills shortage in computing may seriously affect how quickly such innovations come about. "People are a critical growth element for the future," he says. "We're not doing what we need to do to tap into that worldwide pool" of potential engineers and other technology industry employees.

Women should be the leading solution to the skills shortage problem, says Mr Donofrio, pointing out that despite being more than half of the world's population, women make up only 10-15 per cent of engineers and computer scientists worldwide.

"It would be wonderful if every other engineer was a woman," he says. Judging by the audiences he has spoken to at IBM this week and three years ago on another visit, he says the Republic appears to have a better ratio of women engineers than the US.

He is particularly concerned about the quality of mathematics and science education in the US. At the fourth-grade level, American students test at about the same level as other students internationally, he says. But by high school, they test at the bottom of international performance tables.

"We have an incredibly severe problem," he says. "You have to conclude that we have built an un-educational system in the US." Mr Donofrio, who has a special interest in education, believes the problems begin with the fact that teachers are not paid well enough and that the US lacks any national educational standards.

But he's excited about the promise of quantum computing and the other technologies of the future that can be seen in the IBM laboratory environment today. "It gives you goosebumps to think of it," he says.

Klillington@irish-times.ie