“If there were only professors like me, the department would be a disaster”
After two and a half years as the Head of the Department of Computer Science, Professor Emo Welzl is stepping down. In a wide-ranging interview, the outgoing Department Head talks about the future of computer science, the balance between theory and practice, and the importance of appointing new professors.
Professor Welzl, you handed over the position of Department Head to David Basin at the beginning of 2019. Was it difficult for you to leave this post?
Not really, but perhaps it has been an epsilon more difficult than I would have expected. It surprises me that I get the odd pang of nostalgia from time to time. But I'm not worried, as the department is in very good hands.
Did you volunteer for this position?
I don't believe that anyone really volunteers. However, there is the realisation that someone has to do this job. Not everyone accepts this position. You can’t do something like that on the side; it occupies you 100 percent – at least it did for me. If I had tried to do a lot of research at the same time, I would have spread myself too thin. But I have had such a good life in science that I wanted to give something back after all these years.
Did the amount of work involved surprise you?
I was active on the department committee for 16 or 17 years, so I knew that the running of the department was very labour-intensive. What did surprise me was that the demands keep growing. Running the department was much less work ten years ago than it is today. The number of obligations is rising, the administrative apparatus is constantly expanding and interest in computer science is increasing. This results in a large volume of the many small things that arise in daily operations. At the same time, more and more new professors need to be appointed, and this is the most time-consuming task of all.
Which areas of the department's management did you enjoy the most?
Hiring new professors is the most exciting part. The market is very competitive; the people we want are also in demand both at other universities and within industry. You meet at eye level; the candidates evaluate us and we evaluate them. This process takes time and has great potential for frustration if a suitable person decides in favour of something else in the end. But when that person accepts, you really feel that you have achieved something. I missed that feeling when carrying out everyday business, where after hours of answering emails you don't get the impression that you have done anything tangible.
What tips can you give your successor, David Basin?
I have already given him some advice. We have had a very intensive exchange over the past few months. I do not want to present the obvious as a recommendation, since it is already perfectly clear to the other side. David Basin, for example, knows very well that the appointment process is the most important thing and that a good balance between core science and applications is essential. Because, especially in a field as relevant as computer science, there is a tendency to focus too much on applications.
"Scientists are experts in their field, but sometimes they run the risk of believing that they are experts in other areas too."Professor Emo Welzl
You're a theorist yourself. Do you call for more theory in the department?
A balance is important. The fundamentals and theoretical subjects should continue to evolve in parallel with applications. Theory is extensively covered at all good universities, including ours. I believe that the very well-founded education we offer is a major factor in our success. Not just because we teach theory, but because we focus on both sides. If the department contained only theoreticians like me, it would be a disaster. Similarly, it would be just as bad if there weren't any theoreticians at all. Finding this balance for the future is what makes the appointment process so exciting.
Do you think the department will continue to grow at the current rate?
It's like the stock market: it grows until it overshoots the mark. When stock prices rise, everyone thinks they know exactly why this cycle is different from all the previous ones and why there will be no downhill this time. Reality can be different. We already had the dotcom bubble, which burst. The difference between that period and the present is that the dotcom boom was the only big internet-driven fad at the time. Today there are different areas of computer science that are booming relatively independently of each other: security, data science, machine learning, medical informatics and more. In ten years, some of them may be less relevant and new ones may emerge, but it's hard to imagine that none of them will matter any more. Of course, there are all kinds of explanations as to why it is impossible for computer science to decline in importance, but we cannot predict the future with accuracy.
You joined ETH in 1996. Did you have any idea at the time that computer science would experience such strong growth?
It was clear that computer science had the potential to dramatically change society. What is never clear in advance, however, is the scale of this development. For example, you could build something like a smartphone 30 years ago; however, it would have been extremely large and very expensive. The amazing thing is that this device counts my steps, takes high-quality photographs, tells me when the next tram is coming, can even make a phone call, and it does all that while taking up such a small amount of space. If someone had told me 30 years ago that such a device would exist, I would have thought that I wouldn’t live to see it.
What challenges does the rapid growth of computer science pose for the department?
There are many, although every challenge can also be seen as an opportunity. One of the biggest challenges for us is increased demand from every quarter: from students, from politicians, from industry and from the public. Computer science changes society, in a positive way, but also in a negative way to a certain extent – provided that one is able to clearly distinguish what is positive and what is negative. The idea that computer science makes us all happy is illusory. It also brings great problems, which we must deal with in order for computer science to really serve humanity.
"The idea that computer science makes us all happy is illusory. It also brings great problems, which we must deal with."Emo Welzl
Is the department well equipped to meet these challenges?
We are forced – and this is a good thing – to constantly consider new strategies for the future. We are very well equipped to do this in the core areas. The challenge is to find enough good talents. We are not as well positioned when it comes to the social and ethical aspects that exceed our core competencies. Here we have to decide whether we want to address these issues ourselves or whether we want to involve others who can do it better. Scientists are experts in their field, but sometimes they run the risk of believing that they are experts in other areas too, especially with regard to ethical and societal issues. When we talk of deep mathematics, hardly anybody is tempted to believe that they can do it just like that. When it comes to the social sciences, however, this occurs all too frequently. It makes a big difference whether someone is studying the subject intensively or just occasionally reading the newspaper. Fortunately, at ETH Zürich we have the Department of Humanities, Social and Political Sciences (D-GESS) with whom we can approach such questions in a coordinated manner. I believe that D-GESS will play an increasingly important role within ETH going forward.
Does the system of a rotating Department Head have a future when there are so many strategic decisions to be made? Wouldn't a permanent dean be better?
I think our system is good, although both types have pros and cons. I believe that systems that are not focused too strongly on only one direction have better chances of survival. Our system has its price: it paralyses at least one professor in the department for the duration of their term of office. For me, it's worth it. If someone were to do this indefinitely, it would, in political terms, be comparable to a dictatorship. If you take over the leadership from within the collective for two years and then return to the same collective, it’s a completely different premise.
What's next for you now?
First, I'm going on a six-month sabbatical to Vienna and Berlin. I want to do some proper research again. I also want to think about what I'll do afterwards. I have five years left until retirement. It may well be that in these five years I will continue with research and teaching as before. However, I might also consider becoming more involved in the further development of teaching, which is an exciting area that both D-INFK and ETH want to address more thoroughly. We have more and more students, but at the same time, it’s becoming harder to get good teaching staff. In this situation, computer science can provide useful tools. I’m also interested in the tension between education and training. Especially when an area is so relevant and exposed to strong social and political pressure, there is a danger of placing too much emphasis on training, i.e. on the pure acquisition of skills. My hope is that we will remain a higher education school, not just a training school.
Will you continue to be active in the department committee?
Definitely not! Even if I wanted to, it's time for other people to take the helm. And there are many people at the department who can do this job very well; I have no concerns there.