It is a great honour and pleasure for me to address you on the occasion of the 2014 Erasmus Medal of Academia Europaea being awarded to Kurt Mehlhorn.
Please, allow me a personal remark to begin with. I was born in Rotterdam where I spent the first 18 years of my life. In those years, Rotterdam was searching for a new post-war identity and I can assure you that Erasmus was seen as an icon of intellectual achievement connected to Rotterdam. To this day, Erasmus is seen as possibly the most visible if not the most important person ever to be born in Rotterdam.
A second personal remark pertains to my role as chair of the informatics section. Each time the Erasmus Medal is awarded, two instances of achievement come to the forefront. First, the intellectual achievement of the winner of the medal, which I will discuss shortly, and secondly the achievement of those who nominated the candidate arguing the case in a highly competitive field of candidates, all of whom merit this exclusive level of distinction.
This year’s Erasmus Medal is the first medal to be awarded to a member of the informatics section. Naming our field in a stable manner is not so easy. A classical term is computer science. Without hesitation, I may state that Kurt Mehlhorn has made his academic career within the core of computer science. IT and ICT are more recent descriptions indicative of different perspectives of the field. And now informatics serves as an umbrella for a steadily increasing family of specializations, including computer science.
Informatics as an academic discipline is relatively young, although its birthdate is hard to determine. But it won’t be much more than 64 years old, which is around the age of today’s Erasmus Medal winner.
For all members of the Informatics section, it is a clear and important sign of academic maturity of the field that the medal has been awarded to a representative of informatics. And all of us are keenly aware of the implicit achievement of my predecessor, Herman Maurer, who nominated Kurt Mehlhorn and made the case within Academia Europaea.
Without doubt, Kurt Mehlhorn represents the strength of European Informatics Research, not only by being a very effective and influential scholar in his field, but also by playing a key role in establishing a functional academic infrastructure.
The Max Planck Institut für Informatik in Saarbrücken, Germany, serves as an iconic example of Mehlhorn’s institutional contribution.
For many years informatics featured a competition, or rivalry, between three approaches to the field: (a) algorithms and complexity; (b) semantics and programming methods; and (3) networks and systems.
To some extent, the awarding of this medal settles this rivalry. Yes, we all agree that without the ability to invent significant algorithms there is no fundamental incentive for the development of new program notations, or for analysing their meaning – the key challenge of semantics.
And equally important, the conception of new algorithms constitutes a necessary driving force for the development of new computer architectures. This driving force is represented in a remarkable manner by Peter Shor’s factorization algorithm for quantum computers. That single algorithm constitutes a major inspiration for the development of practical quantum computing, a worldwide effort that catches the imagination of a steadily increasing number of researchers.
I will now briefly sketch Kurt Mehlhorn’s contribution to informatics. Appreciating his work requires a framework of computer science terminology that seems to be less known than the basics of astronomy, cosmology, genetics and medicine.
Mehlhorn’s field centres on the following notions. First of all come algorithms. An algorithm is an abstraction of a real or imagined computer program for a real or speculative or even entirely hypothetical computer. Algorithms support this abstraction level, which is needed for academic research.
Next to algorithms are their implementations, still understood within a theoretical framework. The implementation of an algorithm can be analysed in terms of a range of performance criteria: computation time (involving counting a number of steps), memory requirement, correctness (either in principle, as a matter of logic, or statistically in terms of a low probability of failure) and, very importantly, energy consumption. These forms of analysis constitute a part of theoretical computer science referred to as analysis of algorithms.
However, a mere theoretical implementation of a sophisticated algorithm will not make your supercomputer, laptop, or smartphone work. What is needed is a realization of the implementation in a software production environment able to deliver software for widely used computing platforms.
In order to develop coherent families of useful programs, specialization and focus around relevant themes is needed, in such a way that dedicated software libraries can be constructed. In this aspect Kurt Mehlhorn has been effective and influential in structuring the algorithmic part of software engineering.
Starting with the library of efficient datatypes, developed in cooperation with Stephan Mäher in the 1980s, he proceeded with the development of a software library for computational geometry, a library for the analysis of curves and surfaces, and a library for symbolic approaches to linear programming.
The programs of these libraries have found widespread use throughout technology. To use software engineering terminology: Kurt Mehlhorn stands out as an influential software architect.
However, Mehlhorn has always been working from mathematical first principles with a fine taste for analysing problems at precisely that level of abstraction where a path towards further progress can be found.
The majority of AE membership consists of academic professors and you may be happy to know that Kurt Mehlhorn has always been very active in teaching: he has supervised some 65 PhD students, many of whom have made research careers, and he has written four textbooks.
There is no sign of a slowdown of Mehlhorn’s research in recent years. On the contrary, a remarkable range of very sophisticated algorithms with attractive complexity properties have been developed by him and his team in the last 5 years, dealing with topics as diverse as isolating zeros of real-valued polynomials and computing approximations for equilibria in market economics.