From the Bulletin of EATCS No. 97#

A Dialogue on Theoretical and Applied Computer Science with Professor Hermann Maurer#

by Cris Calude

Professor Hermann Maurer is well known to the EATCS community of which he has been a prominent member for many years. He is Chairman of the Institute for Information Systems and Computer Media (IICM), Graz University of Technology, Vice-Chairman of the Board of Hyperwave AG and Chairman of the Advisory Board of Austria’s Competence Center for Knowledge Management (Know-Center). His webpage is www.iicm.edu/maurer.

Professor Maurer wrote more than 20 books and over 600 scientific articles, started or has been involved with a number of companies, supervised over 40 dissertations. He has obtained impressive results in both theory and applied computer science, so he is eminently qualified to compare theoryfrom both inside and outside.

Professor Maurer has also written a number of Science Fiction books, a few available in English, see eg. www.iicm.edu/Xperts[1]

Cristian Calude: Let us start this dialogue by asking you to reminiscence about your time as EATCS Bull, editor.

Hermann Maurer: I took this over from Maurice Nivat with No. 3 of the Bulletin, so we had to create the EATCS logo that still appears (after some 35 years!) on the first page of the Bulletin and also the picture of the automaton which always appears with the Council composition, When I was secretary, the work in EATCS intensified a lot, so the Bulletin grew from some 20 pages to its current volume. The fact that the volume increased rather than decreased is due to G. Rozenberg, who took over the editorship of the Bulletin after me. His imagination, the various columns and cartoons from Dardara have influenced the Bulletin up to today in a decisive manner. I am happy I was able to be an active member of the community for about ten years—after which my interests slowly changed to more applied areas—yet my allegiance to EATCS has never changed.

CC: You started your career by obtaining a doctorate in mathematics in 1965 under Professor Edmund Hlawka, with a dissertation entitled Rational Approximations of Irrational Numbers. Please comment two results from your Thesis.

HM: In the thesis I basically disproved the claim that using hyper-geometric series would give better approximations to irrational numbers than were obtainable by continued fractions or similar methods. Since it is a negative result, in the sense of lower bound, the proof was reasonably tedious.

CC: Then you worked for IBM in Vienna ...

HM: Before I worked for IBM I was an assistant to Professor John Peck, then from the University of Calgary, Canada, one of the leading experts in programming languages. He was one of the inner group developing Algol 68. I am indebted to him for showing me the beauty of exact (formal) definitions for programming languages. I then had almost a year a job as systems analyst with the Government of Saskatchewan, the first time I had a chance to apply computer solutions to large practical programs. On my return to Vienna I joined the IBM research Lab in Vienna, headed by the father of computer science in Austria, Professor Zemanek. I developed a first list processing compiler and then worked in the group that eventually developed VDL, the Vienna Definition Language. In this group I was apprentice to powerful brains like the one of Peter Lucas: under his (and Hans Bekic and Kurt Walk) the technique developed into what is today called Formal Methods Europe; it has survived four decades, something unheard of in computer science. Lest I am misunderstood: I was a tiny and unimportant part of this development, but I learnt a lot about formal methods.

CC: You have been an active—better still, hyper-active—member of the theoretical computer science community. in which areas have you worked?

HM: I continued my work in theoretical computer science when I re-joined the University of Calgary as assistant, later associate professor, in 1966. During my first years there I completed what was to be the tirst German pocket book on the theory (formal description) of programming languages: it turned into a best-seller. Nobody can still understand why this book sold some 50,000 copies at a time (1969) when there were certainly not yet 50,000 German speaking computer scientists. For a number of years I continued in formal languages and automata theory.

CC: How was theoretical computer science in the 70s?

HM: In the early 70s Formal Languages and Automata Theory and related areas (like undecidability results and such) where the central focus of theoretical research. Towards the middle of the 70s the theory of data-structures and algorithms gained more and more importance, and I started to work on this with a number of PhD students and post graduates. Some members of my group, like Professor Ottmann (now University of Freiburg), Professor Albert (now University of Würzburg), Professor Edelsbrunner (now Duke University) or Professor Welzl (now ETH) are just a few of my early students who eventually became more successful in this area than me.

CC: Please tell us your story about the famous MSW team?

HM: MSW stood for Maurer-Salomaa-Wood, with the person in the center, Salomaa, being also the central figure. However, we were really a great team; two of us would get together and try to find new proofs and results, the third one was the one who would do the proof-reading, trying to find holes in the proofs, etc. We became particularly well-known by investigating L-systems and L-Forms, combining and extending ideas of the late Professors Ginsburg and Lindenmayer. Altogether we published some 50 journal contributions together and there was never the smallest amount of friction between us. Clearly, we have remained life-long friends.

CC: How easy/pleasant was to do joint work with Karel Culik?

HM: Karel Culik II has a brilliant mind. Working with him was not easy, however, since he was always ahead, but always also very intuitive. So when I worked with him I was never sure: am I not clever enough to understand what he says, or is he just arm-waving. And it was always both: he had this incredible talent of immediately seeing the rough solution of very complicated problems, but the first idea had to be refined often dramatically to finally give a watertight proof. The story of the DOS sequence equivalence problem, where he submitted n versions (n quite large), each not correct, is typical for him: he eventually, however, managed to overcome all gaps in the original proof-idea and solved this difficult problem.

CC: When and why did you switch from theory to applied computer science?

HM: My switch was fairly gradual, and much dictated by circumstances: first, formal languages, then algorithms, a bit of cryptography, then due to demand, applying this more and more to real-life problems, eventually reducing theory to just one of the things to do. It was also a pragmatic switch: it was much easier to get funding for applied projects than for very theoretical ones.

CC: Are there different skills required in doing theory vs. applied computer science?

HM: I think to do good work in applied areas one does need a solid grounding in theory. The connection is not always clear to see, but the correlation is. There is, however, one basic difference: to be successful in applied work you have to be able to lead substantial teams, something less important for theoretical work.

CC: Can you compare your most important results in theory and applied computer science? Which gave you more satisfaction? Why?

HM: This is a very hard question. A good result in theory gives you the satisfaction of having solved a difficult puzzle that nobody has solved before, the satisfaction that you see things that were overlooked before. Applied projects give you the satisfaction that what you have done may be used by a large number of people, may lead to the formation of companies, to new jobs, and also to some recognition as entrepreneur.

CC: Last week[2] you gave a very interesting lecture in Auckland entitled Theory is important but dangerous. What was its main message?

HM: The answer is almost in the title. We do need theory, both to train students and to solve the theoretical basis for important problems. Examples are numerous: from cryptography to picture compression, from the organisation of data to computer vision, etc. At the same time, theory can be dangerous, mainly because negative results tend to scare people away from trying to solve problems whose exact solution may be very hard, but where there are very good approximate solutions.

CC: As in mathematics, in theoretical computer science beauty is a guiding principle as well as an ultimate indicator of value. Does it make sense to ask “How really useful is theoretical computer science"?

HM: No, the question does not make sense. If we did not have, did not teach, did not research in theory, we would not get anywhere in applications.

CC: What would be your principles in designing a balanced curriculum in computer science?

HM: A balanced curriculum must include a solid basis in mathematics and theoretical computer science, but must also teach more applied areas in a generic way. The emphasis is here on “‘generic". One must not study or work with concrete systems (software—packages) but one must make clear the underlying principles. Only in this way it is possible to make sure that the knowledge acquired does not obsolete too fast.

CC: Why computer science seems to have lost its appeal to the young generation?

HM: I really don’t think it has. That numbers have dropped in some places is due to the negative PR around the "bust of the internet bubble" that suddenly created the wrong impression that computer science would not be important any more, However, this is definitely not so, The world needs many more IT specialists than are currently produced.

CC: You wrote extensively on the future of technology using metaphors such as telepathy and teleportation. Tell us please more about these unusual writings?

HM: Basically, any technology that is far beyond our current one looks like a miracle. This statement is often attributed to Arthur Clarke, the SF writer, but goes actually much further back to around l850! Just consider: if you had used a digital camera a few hundred years ago you certainly would have been burnt at the stake. Things like telepathy or telekinesis that sound like impossible or miracles today will be implemented with suitable technologies before too long. Telepathy is almost a reality due to new brain-computer interfaces, and things like telekinesis will be handled (I quote Kurzweil) by swarms of nanobots. So: it is fun to write things that sound like SF but might well turn into reality at some stage, see XPERTS in www.book1ocker.com or my home-page www. iicm. edu/maurer.

CC: How dangerous is Google?

HM: Google is THE largest data—mining company we have in the world. It knows more about people and organisations than any intelligence agency, and there is no control via data-protection laws at this point in time. This gives more power into the hand of one particular company than is good for all of us. Actions will have to be taken to curtail those powers, and I know that such actions are on the way, without ruining the great advantages Google has for all of us due to its marvellous services, nor without ruining the value of the company for Google share holders.

CC: What are your next decade projects?

HM: Open access scientific journals, helping to develop quality-improved community encyclopaedias akin Wikipedia, and assuring that when we look for something on the Internet we don’t find millions of pieces of information, but the knowledge we have been looking for.

CC: Thank you very much.

[1]He recommends (for anybody interested) to start with the "Paranet".
[2]November 7, 2008