Digging for information

Soon after the first computers were built it became obvious that they would become useful tools for storing information. Ever since, people have been working on ways to actually retrieve that information - and the solution to the problem is nowhere near as obvious. In recent years, the sheer volume of information that is available has actually made it harder for people to find what they really want.

Professor Keith van Rijsbergen, head of the information retrieval group at the University of Glasgow, has been one of the pioneers of developing information retrieval into a science in itself. And the importance of this to research information has been recognised by the UK Online User Group of the Chartered Institute of Library and Information Professionals (CILIP), which has just awarded him the 2004 Tony Kent Strix Award, an annual award for individuals or groups who have made an outstanding contribution in the field of information retrieval.

For more than 30 years, van Rijsbergen's research has centred around using mathematical methods to model flows of information and devise ways of retrieving significant information from the mass that is out there.

For many people Google is the only search engine out there, and much of van Rijsbergen's early work has fed into the process that created such tools. But van Rijsbergen believes that search engines can do better. In particular, they could include elements of feedback from each individual user to improve the quality of retrieval.

At one time, he was virtually alone in exploring these issues but today commercial publishing companies are starting to wake up to the fact that there is much more that can be done beyond simply dumping the contents on their journals onto a website. This means he is much in demand for consultancy work, but he turns much of it down because his main interests are doing his own research and supervising PhD students.

Van Rijsbergen was born in the Netherlands in 1943 but from the age of about five he became a world traveller because his father was an engineer specialising in dredging and land reclamation. He began his education in Indonesia where he went to a Dutch-speaking school, but spoke Indonesian with his brother and sister at home. Between trips back to the Netherlands he also lived in Australia and finished his secondary schooling in Namibia, where everyone spoke German but his schooling was in English. He is fluent in English and Dutch and can get by in Africaans, French and German. His Indonesian is a bit rusty.

'I changed schools about 17 times but there was no residual trauma. I just took it in my stride,' he said. 'What would happen typically is that the local school would give me a pile of books to study and get up to speed. I liked that. At the end of my schooling in Namibia I was top of the class. I liked languages, literature and science and it took a long time to choose science and mathematics.'

At 18, he moved with his family to Western Australia and enrolled at the university there. He started with the intention of doing theoretical physics but eventually settled on applied mathematics.

After completing his degree he decided he was fed up with science, so he went back to university to do a year of an arts degree. 'I loved it but I found that in my spare time I was doing mathematics. I decided there was a message here and that I should stick to mathematics,' he explained. 'I had a job in the mathematics department tutoring but computing was just starting then so I got myself a 'meal ticket' in the form of a postgraduate diploma in computing.'

By this time he had met and married his English wife, and they decided to move to England. Just before completing his diploma he was offered a job as a programmer with Cambridge University's Computing Service. While there, he got himself a part-time job at the Kings College Research Centre, helping with the computing aspects of a research project. He said: 'After a while I got bored with programming and they suggested I do a PhD. I asked what that was and they said I just had to spend three years doing research and write it up. So I started a PhD in the area of automatic classification and applying it to information retrieval.'

While at Cambridge he was offered a lecturing job by the head of computer science at Monash University in Melbourne, who was working in the same area so, once he finished his PhD, he was back on the plane to Australia.

He said: 'That was the best thing and the worst thing. I really had nobody to talk to at Monash about my field. I said to my wife that the only way I was going to survive is if I talk to myself so I wrote a book on information retrieval, which became quite well known. I actually wrote it as a research monograph rather than a text book but, because the field was so new, it became a text book.'

He wanted to return to Cambridge, so he applied for a Royal Society Research Fellowship and went back to the Cambridge Computer Lab and Kings College. 'That was a fantastic time because I could spend 100 per cent of my time on research and found my feet in my field,' he remembered.

The next stop was University College Dublin, where he was head-hunted again to lead a push into computer science. 'The department was a mess. It was not really functioning well for various reasons but it had some good people, so I started to rebuild the department,' he said.

Alan Smeaton, professor of computing at Dublin City University, was a student of van Rijsbergen's in the early 1980s and describes him as a grandfather figure of information retrieval with a huge legacy to the field in terms of the students he has supervised. He also said van Rijsbergen was one of the two greatest gurus of the field in the world over the last 20 years. 'When he came to University College the computer science department was in a terrible state. What he left was probably the top department in the country,' said Smeaton.

In Dublin, van Rijsbergen concentrated on building up the teaching side but, after six years and one new building, he wanted to do the same on the research side. He approached various sources of funding but they decided they could not afford it. His wife did not like living in Dublin much so, when the chance came up to move to his present base in Glasgow, he leapt at it.

Glasgow University was looking to build up its computer science department and wanted a focus on research, so it fitted with his aspirations. The department flourished, getting a five research rating. He rose in the university hierarchy to become head of department but eventually decided that his first love was research and that is what he does now. He has no intention of moving as he has received a great deal of backing and has built the information retrieval group into one of the top three in the world. Recently, he published another research monograph representing the results of 10 years of his own work, which he is hoping will become as significant a milestone in the field as his first.

Van Rijsbergen has never got bored with information retrieval as a subject for research and is as enthusiastic today as he was 30 years ago. He has followed a path that has taken him into a very diverse range of subjects and believes there is always more to learn. 'Although I liked physics I was always more excited by applying mathematics to non-physical problems, like linguistics or psychology or information management,' he explained.

'The great thing about the field is that you see things eventually being applied and people using them. For example, the search engines that people use today are really an outgrowth of the work that many of us have done over the last 20 years.'

Life is not so lonely now. With so many of his students staying in the field either in the academic or commercial world, the field has become more interesting as it has grown. 'There is an increasing number of people that I can talk to. I am constantly learning things, especially from PhD students. They really make me think hard, which is quite exciting.'

Van Rijsbergen's group at Glasgow has been involved in several large research programmes, including EU projects. Its latest development is to design its own search engine called Terrier. This is on the verge of being released into the internet to see what people will do with it. 'Terrier is a better engine than Google, if you measure it, but it not so much better that people are going to say forget about Google and start using Terrier,' he predicted.

In the future, van Rijsbergen hopes that search systems will go beyond the rules of inference that the likes of Google are based on. 'You have an incomplete set of rules that make intelligent decisions and present you with the information. You then make some assessment of that information, and then the rules should be flexible so that they can adapt to your reaction to that information and then improve its performance,' he commented. 'Google gives a kind of universal feedback from the connectivity of the web. But you don't always want the most popular site. Some very important work is often very badly connected to the outside world, and people want to find it. The feedback I am talking about is personalised. What guidance does Google give you on changing your query? One of our aims is to build systems that incorporate relevance feedback.'

Research is not the only thing that van Rijsbergen has demonstrated his endurance at. One of his hobbies is long-distance swimming, which he does with his wife and daughter. This summer they travelled to Greece to swim between the islands. He also spends a lot of time visiting Scotland's Western Islands but he admits that the water is probably a bit cold in Scotland for long-distance swimming. He likes living in Glasgow and his work there so, despite having spent so much of his early life moving around, there is every chance that he will be staying put for a while yet.

Curriculum Vitae

Education
1965 University of Western Australia, BSc mathematics
1968 University of Western Australia, Dip NAAC computing
1972 University of Cambridge PhD

Employment
1966-68 Tutor, Mathematics Department, University of Western Australia
1969-72 Senior research officer, King's College Research Centre, Cambridge, UK
1973-75 Lecturer, Department of Computer Science, Monash University, Australia
1975-79 Information research fellow, Royal Society Scientific Computer Laboratory & King's College, University of Cambridge, UK
Spring 1978 Visiting professor, School of Library and Information Studies, University of California at Berkeley, USA
1980-86 Professor and head of department, Computer Science Department, University College Dublin, Ireland
1986 to date Professor, Department of Computing Science, University of Glasgow, UK (including time as head of department)