An Interview with Niklaus Wirth
Total Page：16
File Type：pdf, Size：1020Kb
Portrait Conversations on the History of Informatics (I): An Interview with Niklaus Wirth Informatics is a young science and during the first decades computing which has mutated from the military and scientists were too busy with the basics to have time for reflections. the atomic research community of the 50s Informatics seems now to come to its adolescence. It is getting more to the people today who are telephoning, and more independent from its many parents such as mathematics, playing, studying, and working. The scope systems theory, electronics and even economy, and it starts to pose of application has moved during this time questions concerning its roots, its deeper sense, and its role in society to embrace robotics, games, shopping, – in short, it starts to concern itself with its own history. However, even the opera. Advances in language de- while many agree that such a history is important, few agree on what sign have allowed for abstraction, which informatics is and how its history should be investigated. of course defines the boundary between virtual and real. And so on. Informatics herefore, the OCG (Austrian Com- engineers to a digital ‘context’ which fra- history deals with the co-evolution of the puter Society) and ÖGIG (Austrian mes the global economy and, increasingly, computer and these other dimensions. TSociety for History of Informatics) western society. If we consider history to be have decided to conduct a series of in- a study of change, then I’d say we have suf- What was the first step from the terviews with outstanding computing sci- ficient material. early computer towards this world of entists, asking for their views on both the informatics? general issue of the history of informatics But it is material that many of us I would definitely say reliability. The vacu- and for their personal reflections on both have lived. Would a formal study um tube-based computer had a meantime the past and the future role of informatics bring anything new to the table? between failure in the order of hours: each in our society.1 If you are living within (high quality) tube had a MTBF of about The following interview a particular ‘context’, 10’000 hours. Hence, in a computer with, with Niklaus Wirth is you may not be able say 5.000, tubes, some tube would go out the first of this series. to see this context wi- every two hours and with it the entire com- thout going outside puter. This meant that a program had to Is Informatics old of it. Much of today’s have a runtime of less than two hours if enough to start dis- world lives outside you were to have a reasonable chance of cussing its history? the umbrella of the its completing before the next breakdown. The word informatics computer, networks, The introduction of the transistor, around is itself rather young, or even the telepho- 1960, changed all this. dating back to about ne. We tend to forget 1970 and created by this because we are But the transistorized computer Dreyfuss. It’s a con- so firmly positioned remained the plaything of a small po- traction of “traitement within the paradigm pulation of specialists for quite some d’information automa- of informatics as a time. What was happening? tique”, an expression way of life. In looking For one, many were learning to program. which emphasizes that at the history of infor- A new kind of (written) language was something automatic matics, we step out- needed and this has taken a long time is going on. Using this sense of the word, side this context and look at ourselves. and is still in development. The computer informatics would exclude anything before was a tool which demands great attention automatic machinery. For me ‘automation’ What would you include within the from the tool-bearer. To the degree that the implies electronics, so I would leave out the scope of a history of informatics? programmer has to focus on the underly- mechanical calculating machines of the The computer itself is important but the ing computer’s details, he is distracted 19th and early 20th centuries. With this de- computer alone does not define infor- from concentrating on the task at hand. finition, a history of informatics would begin matics. There are many other dimensions. A good language provides a better plat- with electronics, or perhaps with the elec- The way the computer is used also tells a form, computing model, abstraction. For tro-mechanical computing machines of the story; in the 50s the computer sat in a labo- a long time computer science remained 1940s. We are looking at only 60 years, but ratory and a scientist would sign up days in the domain of the engineers who built we move during that time from an unwieldy advance to have this huge machine all for the machines, and of the mathematicians tool for a handful of mathematicians and himself. Then there is the user community who used them. Programming languages 1 Thanks go to Prof. Laszlo Böszörmenyi, former student of Wirth, for the idea and impetus to set this series of articles in motion. JOURNAL 22 Portrait were the first topic that came from neither is one of the most influential scientists in the area of infor- of these two camps. Among languages, Niklaus Wirth matics. He made essential contributions to Algol-60, and thus to the establishment Algol 60 was the first with some meas- of informatics as a science. He designed a series of programming languages. Early ure of mathematical formalism and rigor. works on Euler, PL/360 and Algol-W were followed by the extremely successful and This layer of abstraction was a catalyst for significant Pascal language. After Pascal and a longer visit at Xerox PARC in Califor- programming methods, language design, nia, he devoted himself to the efficient coupling of hardware, language, compiler and and algorithms. These topics attracted operating system design. This led to the simple and clean programming languages new interests, users and ideas about what Modula(-2) and Oberon(-2), the Lilith and the Ceres computers and the Oberon ope- to do with a computer. rating system. Wirth is one of the most quoted authors in computing science. Many of his books and papers belong to the evergreen classics, one of the best known Who was driving the research being Algorithms + Data Structures = Programs, which helped generations of stu- agenda? dents to understand how algorithms really work. He was honored with numerous Initially it was the military and then busi- awards, among others, in 1984, with the ACM’s Turing Award, the highest distinction a computing scientist can achieve. Wirth has retired from his professorship at the ness. Despite the interest and involvement ETH Zürich, lives in Zürich, actively follows the development of informatics, and of a new community, computer science reflects on it with a critical view – not without his fine sense of humor. remained for some time an academic stepchild; it was in the house, it was use- ful, but it had no status. The programmer were driving forces in the development of was homeless, commuting, as I did at networks. Now it has become difficult to Berkeley, between the engineering and find a research project in computing that mathematics buildings. But in1965 Carne- does not involve networks. These deve- gie-Mellon (Pittsburgh), MIT (Cambridge), lopments profited from advances in lan- and Stanford (Palo Alto) each conferred guage design, but also contributed to it. upon computer science the status of an For example, I introduced Pascal in 1970, academic discipline in its own right. Each but Pascal attained its real acceptance af- of these universities established an inde- ter 1977. This was due at least in part to pendent computer science department, the personal computer. Along came a ge- thus giving recognition, official status, neration who had ready access to compu- attention and financing to a sector much ters and, most importantly, who had not to broader than computer technology per se. unlearn old habits, whose first interface to This meant a tremendous shift of focus. Niklaus Wirth was interviewed by the computer was a high level and structu- Ann Dünki, a former Ph.D. student of him. red language. This was the first generation But it took many universities, in- which was free to focus on what you could physics. But a sound education in natural cluding the ETH in Zurich, another do with the computer, rather than on the sciences is essential to an understanding fifteen years to grant full recognition. computer itself. of our world; how does a gear, a pump, How were you impacted? an electric motor, a wing, a gas turbine, a The chain reaction had started. With or Where will informatics take mankind? computer work? What are its underlying without a computer science division, the Are there important (historical) les- principles and laws of nature? The com- informatics community at ETH and many sons that we should keep in mind? puter can provide capabilities and infor- other universities and research institutes People believe that they can no longer live mation, not more. But it seems that many coalesced around the changing agenda. without the mobile telephone, the PDA, the people believe that the computer makes internet, computer-based entertainment. understanding superfluous, “because you Somewhere we moved from a closed The computer has simply become a part can look it up in the Internet”, if needed. community of computer specialists of the fabric of life, just as mobility has. But to a global community of informatics how many people understand the relati- Would you pose a question to the aficionados.