Bringing knowledge back home
Scientific research in the USA has been attracting the best brains in the world for many years. Other countries of course both produce and retain equally brilliant people but, for the scientist who has worked in the eye of the storm in the USA, it can be hard to return home where their work may never be noticed as much again outside their own national boundaries, as they get sucked into their own national research and education systems.
Efstratios Gallopoulos - Stratis to his friends - is trying to change that in his homeland of Greece. After nearly 20 years of training and researching abroad, he has achieved his ambition of a full professorship at a major Greek university and is trying to bridge the gap between his country's own environment and the US-dominated global scene. He has brought back with him not only his knowledge and skills but also some ideas from the US system of education and research, which he feels can supplement and improve the traditions of his country.
As a former student of Dan Slotnick, and having spent many years at the centre of things in the University of Illinois at Urbana-Champaign, he is well connected and is bringing a lot of attention to the University of Patras and making it one of Europe's best-known centres of computational science.
Professor Ahmed Sameh, head of computer science at Purdue University, who has known Gallopoulos since he was a research student said: 'He is a very pleasant person to work with, amiable, extremely bright and witty. He is also quite energetic and I valued him as a colleague, always thinking, always coming up with original ideas both about his research and the research of others. He is highly regarded in parallel computing and he has tried to carry this through to Problem Solving Environments. He has met with success.
'He has not exiled himself by going to Patras; it resolves a family situation. He could have easily obtained good positions in the US but he did not try very hard. Stratis is in touch with me and others who he worked with, so he has not lost touch with researchers in the US and Europe. Greece may be a little bit removed from the centre of activities, but he is starting to wake them up a little bit.
'He has worked with the pioneers of massively parallel computing and he is a prolific researcher and publisher. He is what I call a true computational science and engineering person. Not only does he concern himself with the classical computer science issues, but he also likes to look at whole applications across a variety of component disciplines. He has made significant contributions in numerical algorithms, systems software and implementation on an architecture, and what he has developed is being used. He is one of the people who have pushed the whole area of problem solving environments.
'I and several of my colleagues miss him but I hope he can now travel more and I assure you we will continue to hear about Stratis Gallopoulos.'
Gallopoulos was born in Thessalonika in northern Greece. His father was a chemical engineer who had studied in England earlier in his career. He started learning French at the age of four and then English at the age of nine. He went to an American school in Greece, so he was well set for an international career. He said that three of the 'better ones' went on to study at MIT. As he finished high school, Greece was thrown into turmoil by the collapse of the notorious 'Colonels' regime. Gallopoulos said he never really felt any repression from the regime except he was not allowed to have long hair or wear flared jeans. As a child he enjoyed reading science fiction and, although his parents were against television, they allowed him to watch the Apollo moon landing. He had no idea at that stage that he would later work at NASA himself.
He started applying for university places abroad and settled on Imperial College, London, as much for the cultural attractions of London as the course itself. He had always dreamed of doing research but there were few opportunities to do pure maths research so he had to look around. He had taken a course in numerical analysis, which interested him greatly, but he didn't gain much appreciation of computing as his course-work was based on punch cards. England itself was having its own political problems with the 'Winter of Discontent' so he started looking for a graduate place in the USA.
He said: 'The idea in Greece is that if you do mathematics you become an engineer. I had done a course in control and figured that this was the most mathematical part of engineering. I was offered a place on a control programme, but it was offered by the computer science department and in Greece there is a danger that, if you change your subject, your degree might not be recognised. I applied to the Maths department of the University of Illinois at Urbana-Champaign and they came back with the offer of a fellowship that would fund my studies so I took it. When I arrived there it was a cultural shock because it was in the middle of nowhere. Eventually I found that it had a lot to offer, but not as much variety as London.'
Gallopoulos said he was always interested in learning new things and, although he was on a maths Ph.D. programme, he could not resist short courses offered by other departments. One day he started attending a course in the computer science department given by Daniel Slotnick, who had worked with Von Neuman at Princeton and was a famous pioneer in parallel processing. He found Slotnick inspirational and so he immediately arranged a transfer to the Computer Science Department with Slotnick as his supervisor.
Gallopoulos said: 'What he said to me was very attractive. He asked me to come and work in his group on his new project and in the summer he would send me to the Goddard Space Flight Centre. It was my dream ever since I had watched the Apollo Moon landing.'
The contract he worked on at NASA was designing a computer system to process Landsat pictures. It was a system called the Massively Parallel Processor with 16,000 processors in an array and Slotnick's team was developing software to run on it. He decided to make his Ph.D. project about writing software that could use this system for economic modelling as well as for applications in physics such as meteorology.
After gaining his Ph.D. Gallopoulos was offered a job at the University of California, Santa Barbara. Although this would take him away from his hero Slotnick, it had many attractions. Not least of these being the weather. Growing up in Greece with long hot summers and mostly mild winters he was deeply fed up with the Mid-Western winter. He said: 'As soon as I arrived for my interview I felt at home. It was like paradise.'
Another thing he had to leave behind was his new wife, a classical philology graduate student. Very long distance commuting was to become a feature of his life.
He enjoyed his time at Santa Barbara, but his heart was still in Urbana. Slotnick had won a contract to build a new massively parallel machine and Larry Smarr had founded the Center for Supercomputing Research. Slotnick's sudden death while out jogging in 1985 came as a bitter blow to Gallopoulos. A year later Ahmed Sameh tempted him back.
Gallopoulos said: 'I kept getting these messages saying that everything interesting in parallel computing was happening in Urbana. I was approached by Ahmed Sameh who invited me to join a group developing software for the Cedar project. The weather in Santa Barbara was good and the people there did not want me to leave but after a year and a half I went back to Urbana.'
The work on Cedar continued until 1993 when the machine that was envisaged was built and the funding from the Department of Energy dried up. Gallopoulos starting looking for new projects. He was interested in an idea that Dave Cook and Ahmed Sameh had at Urbana for setting up a masters or Ph.D. programme in computational science, which had not taken off because of political factors at the university. Another project was being proposed by John Rice and Elias Houstis who were working on environments for solving partial differential equations at a higher level than writing programmes. This emerged into Problem Solving Environments.
Gallopoulos helped organise a conference on the subject in Washington DC, which attracted interest from the National Science Foundation and Cleve Moler, one of the founders of Mathworks. A report on the conference became widely publicised and a number of important projects in problem solving environments (PSEs) started to take shape. These PSEs became a major thread of Gallopoulos's research and he has developed a number of programmes that are widely used.
Gallopoulos said: 'These environments allow scientists to specify their problem in a way that is closer to their natural language, just like the high level CAD tools that engineers use. You can think of them as boxes containing very complicated software that requires a great deal of expertise to build, but the user doesn't need that expertise to use the box. There are many things, which are quite generic and do not depend on the particular application that you have in mind. For example the software bus on which you have various activities being connected and exchanging data across this bus.'
Gallopoulos believes that the field can best be defined by its products and so he describes problem-solving environments as the products of computational science and engineering. He likes to focus on practical outcomes but admits that this could simply be a product of his constant desire to learn new things. He accepts that the practical focus is not enough, and that others need to concentrate on more abstract problems in order that new discoveries can be made.
He said: 'There needs to be a balance and you could argue that nowadays we are overdoing it with the practical side of things, as opposed to the type of research that leads to great new ideas. If we had a very educated public that really appreciated science, then you wouldn't need any practical outcome. The politicians would vote for budgets to support research. It is very important to educate the public about the advances that are made by science and technology. The outlook is positive because people are a lot more educated than they were 20 years ago, but I think that every scientific body has a responsibility to show what the products of their research are.'
As the field took off Gallopoulos started to gain more exposure, either organising or chairing conferences. He was not short of job offers but there was one that contained the irresistible draw of returning to his homeland as a full professor. A position had arisen at the University of Patras, home to one of Greece's most prestigious computer science departments. He was approached at a conference and persuaded to apply. Some US colleagues may have thought that Patras was a bit provincial after the heady world of US science, but to Gallopoulos it was the best job he could imagine ever getting. Gallopoulos was expecting a struggle to get the position, as local applicants were often favoured over those educated in the US, but he got the position.
He said: 'A lot of the obstacles that are sometimes put in the way of Greek scientists returning home simply did not happen in my case.'
He admits he has stirred things up a bit since arriving at Patras in 1996. It was in fact his first exposure to the Greek system of Higher Education and it took some time to adapt. His priority has been to raise the profile of computational science in the Greek scientific community. He has had to translate a lot of his writings into Greek, and in some cases has had to create new vocabulary because there were no standard terms for many aspects of his work. In particular he has pioneered a more US model of a graduate programme in computational science. His approach is supported by the Greek Government because, as a member of the European Union, it is committed to convergence in the recognition of qualifications. He admits there is some resistance in Greece, as it is perceived that degrees from some other countries, such as the UK where first degrees are obtained in just three years, may not be seen as equal to degrees granted in Greece.
He continues to travel to the US frequently and not just to keep in touch will colleagues. After returning to Urbana to be with his wife, they were separated again when she had to move to Pennsylvania to take up a job as a professor at Penn State University where she raises their five-year-old daughter. They have lived in different cities for ten years and different continents for the last six. He said: 'For me, travelling from Patras to the US is like someone else travelling from London to Reading. I make the journey about once a month or so.'
But at last they will be together as, from this summer, she has taken up a position in Patras.
'What was extremely easy for me was difficult for her,' said Gallopoulos, who now has a good excuse to stay home. But having spent so long on the world stage nobody is expecting to see the last of him.
Imperial College, London. B.Sc. Mathematics
University of Illinois Ph.D. Computer Science
University of Illinois Teaching Assistant Department of Mathematics
NASA Goddard Space Flight Center, Summer Research Fellow
University of California, Santa Barbara, Department of Electrical and Computer Engineering, Assistant Professor.
University of Illinois, Assistant Professor of Computer Science.
University of Patras, Greece, Department of Computer Engineering and Informatics, Full Professor.