Martijn Warnier: Systems are Tools and Mathematics is Aesthetics

Amsterdam, 27 April 2009: At the Intelligent Interactive Distributed Systems group at the Free University of Amsterdam, Martijn is one of the spiders in the web. When I entered the group as visiting fellow he made me feel at home. While arranging practicalities he would, in his own special shorthand way, make remarks about the research I was about to do. It seems that Martijn does not appreciate socially accepted thinking or questions, as becomes apparent in the interview hereunder as well. For a start he will defend his hard position in current debates, not at all willing to allow any ‘soft’ reasoning of the social scientist I am. Surprisingly, after he has cornered my thinking, he starts to be really interested in what is actually at stake. In the exchanges I have with him, the concept of ‘interdisciplinary collaboration’ does not describe what happened. It has been more a query into how two planets communicate. In such conversations the scientific, social and political world all connect and little by little space for thinking evolves.

----------------------
Read the interview here

Expand selection Contract selection

Martijn Warnier

Dr. Martijn Warnier graduated with a Masters of Science in Cognitive Artificial Intelligence from Utrecht University, the Netherlands, in the beginning of 2002. He did his PhD in the Security of Systems group at the Radboud University in Nijmegen. His research focused on Language based Security and the mathematical formalization of properties such as non-interference, confidentiality and integrity. For the last three years he has worked as a Postdoctoral Researcher in the Intelligent Interactive Distributed Systems group at the VU University Amsterdam. Since September 2009 he is appointed the position of assistant professor at the TU Delft. His current research interests, besides security, include the interdisciplinary research field of Computer Science and (Computer) Law, and Self-organizing and Autonomic systems. In his free time Martijn Warnier loves to be an actor in the theatre.

The ideal situation

Computer words are very different from the words human beings use, because computer words have very clear, specified semantics. There is no ambiguity according to Warnier.

Computer words are very different from the words human beings use, because computer words have very clear, specified semantics. There is no ambiguity according to Warnier.

If you type the command he’ll do exactly the same every time. While with human words, depending on the context, or the intonation, or the body language, they can mean a lot of different things.
Logic words function in a structure and have a very precise meaning. In mathematics there is no ambiguity. If you give a computer a command, a word, you can trace back all the steps to see what actual gates are flipping at the processing. Computer systems can talk to each other and this can lead to negotiation, but in principle it is still the same complete reduction back to the CPU, the central processing unit, or the underlying network infrastructure. You can explain everything that is happening before it happens basically. In principle it is a closed system; there is no influence from outside. A user can influence the system, and doing so brings in the randomness, but what the system will do will be completely predictable. This would be the ideal situation.

Expand selection Contract selection

What goes wrong

Because of the current business models in software, we end up with very crappy software that we don’t really understand how it works, argues Warnier.

Because of the current business models in software, we end up with very crappy software that we don’t really understand how it works, argues Warnier.

The current business model is “release first and bash later”. So first make sure that you’re the first to market and then you sent out updates on and on and eventually you get something that works, or maybe not. That is what Microsoft is doing. And there is no real demand by users for reliable software. Maybe in niche markets, if you want to assemble a rocket you do not want it to explode because a computer divided by ‘0’ does not work or so. But there’s no real market for it; that is why it is not happening. Many side effects happen as a result of this. For example users get frustrated when the hard disc fails or the operating system crashes and they have to reboot and loose half a days of work.

Expand selection Contract selection

Specifications

One of the reasons for the muddy software world we live in is, according to Warnier, that we don’t even know what the real specifications are.

One of the reasons for the muddy software world we live in is, according to Warnier, that we don’t even know what the real specifications are.

By formulating specifications, which are about function and independent of users, quality control is supposed to be guaranteed. One tries to write down all the possible inputs and all the possible outputs and these are written down in real words; we want the system to be secure or we want the system to be reliable. But those words don’t any longer have an exact precise meaning; they can mean lots of different things. We don’t really know what we want our system to do, says Warnier. Nobody takes the effort to really find out. It’s very complicated to get these things right, and as long as there are no people demanding these things and are willing to pay for it, it is not happening. There’s no need for it because people will buy the crappy software anyway because there is no real alternative.

Nevejan suggests that it is an interesting concept for witnessing, to focus on conditions and specifications as two things you have to create for a system to work. There are conditions that you witness and to which you can be witness if you have some specifications. Otherwise you can’t see the conditions. Warnier may agree or not, because these are definitely not his ideas but Nevejan’s, he states.

Expand selection Contract selection

Systems are tools

For Warnier systems are tools in the first place. He would not ever call them participants. Maybe abstractions, but systems themselves don’t have any power.

For Warnier systems are tools in the first place. He would not ever call them participants. Maybe abstractions, but systems themselves don’t have any power.

Systems give people a context or a structure, but the computer system itself can’t force them to do anything, or anybody else either. If worse comes to worse systems can be closed them down and people would use a different tool to communicate. For these reasons Warnier argues that the system in itself is not an actor. Also, he argues, systems cannot witness anything because witnessing implies a form of consciousness that systems do not have. They observe, they monitor, but they do not witness. You interact with a system, but the interacting doesn’t imply witnessing.

Of course, it is possible for a user to project witnessing to a system because a user interacts with a tool and the tool is very complex and most users have no idea how it works basically. Warnier thinks it is a survival tactic of humans to try to humanize the system, to try to give it more human qualifications. This is something that designers exploit, he argues, trying to make the system more humanlike because it makes it easier for humans to interact with systems.

The systems themselves are not active, they are just tools that are exactly the same as a TV or a hammer, you can use them to do a lot of different things, but they don’t do anything themselves. Warnier himself uses systems to communicate with other people, to do simulations and to do experiments. But he could also use pen and paper to do a lot of these things, he argues. He feels that the technology does not change him, the way a potter is also changed by clay for example as Jogi Panghaal decribed. He makes the technology do what he wants because he understands the technology he is able to do so. Most people don’t want to understand how everything exactly works and why should they? They are using a tool, Warnier argues.

Expand selection Contract selection

To better understand systems

Nevejan replies that we need to better understand systems because we live in a world where they have become very important; they interfere in more personal spheres than any tool ever before, even in the body.

Nevejan replies that we need to better understand systems because we live in a world where they have become very important; they interfere in more personal spheres than any tool ever before, even in the body.

Witnessing each other between human beings sets the terms for interaction. Human beings witness and are witnessed by systems and these systems influence possible survival more and more. Systems operate with a speed and scale human being’s cannot compete with. It’s more micro than ever and more macro than ever. Nevejan argues that the basic way we witness, that we perceive what we see including our understanding of this, is changing profoundly because of systems. So it is of vital importance to understand the relation between human beings and systems better than we do, she argues.

Warnier answers that systems are also very stupid. Systems themselves are not the real danger; it is in the people who are using the systems. Also he argues, a lot of changes we now attribute to computers and distributed systems for example, are changes and feelings of uncomfortability that were also triggered by earlier media and technologies. Our grandmothers have seen huge changes happening in technology. He agrees though that the invasion of the personal sphere is much larger than ever. But for Warnier this is an issue of political concern for him personally and less an issue of scientific research.

Expand selection Contract selection

YUTPA in systems

Computer infrastructure is not much more then a couple of bites in one way or another. So there’s nothing there that lives. Time, rhythm, place and culture are all human concepts that people can perceive but actually they are human projections onto the system and they are not there, Warnier argues.

Computer infrastructure is not much more then a couple of bites in one way or another. So there’s nothing there that lives. Time, rhythm, place and culture are all human concepts that people can perceive but actually they are human projections onto the system and they are not there, Warnier argues.

If you take humans out of the equation, there is just the system and the system is a tool. If one uses a very high-tech fast computer-network, or one uses smoke signals, at both one can do confidentiality or transparency for example, because both are basically just carriers of human information.

There are a lot of practical reasons why you need to think about the time and space of your system: for example the universal time between computers so they can synchronize and physical structures deteriorate after some time. Also, mathematics as action can be understood as transformation. Or in other mathematics relations are defined and for example operating systems will be able to recognize each other or not. It is not possible to compare human beings with systems; human beings are too complex. The best you can do is to make an abstraction of a human being, and compare that to a system or equation or what ever, since we don’t understand how human beings really work, argues Warnier

Expand selection Contract selection

The platonic heaven of mathematics

Warniers work is basically thinking and publishing these thoughts, which is why he only needs pen and paper to do his work. Technology does not affect his way of thinking because he usually thinks in higher abstraction levels, in algorithms, information theories and mathematics.

Warniers work is basically thinking and publishing these thoughts, which is why he only needs pen and paper to do his work. Technology does not affect his way of thinking because he usually thinks in higher abstraction levels, in algorithms, information theories and mathematics.

For Warnier the most interesting part is formalizing ideas and abstractions into mathematics. He tries to make ideas crystal-clear by formalizing them completely and enjoys the deep insight in how information processing works for example. There are beautiful mathematics and very ugly mathematics. Warnier appreciates beautiful mathematics in the same way as he can appreciate art for example. Beauty and elegance are important in trying to understand how things work is about.

Of course Warnier has to think about how his abstract ideas can be applied in some specific context. Currently for example Warnier and colleagues are trying to get spikes out of energy production in larger energy systems. However, a lot of the systems are mainly crappy realisations of beautiful mathematic concepts. Technology is independent of these; it is this realization of some of these ideas but mathematics is basically a separate world, a platonic heaven. In Warniers work users are not important; they are way too concrete. Warnier and colleagues think about the algorithm to do some simulations and than somebody else will use it in the real world. It’s a whole other process.

In mathematics there are no values as there are values in communities of people. Values like trying to be polite to each other, smile if somebody makes a joke, listen to other people, respect them etcetera, are completely independent of the system. Of course science can be used for good and for evil and as a scientist one has a certain role in this. According to Warnier one should try to at least envision the things you come up with can’t be used for evil. However, to have moral values in a system does not make sense to Warnier. Moral values are for people. The only value he cherishes in his work is to create good quality and one can argue whether this is a moral value or not. In the end people will do with the tools whatever they want to do with the tools. So if they can abuse it for something, they will do it. And it can help other people that will also do it. It is independent of the system, I think. There are always people using systems for abusing others or helping others, and for anything in between.

Expand selection Contract selection

Not wanting to know

Even though it is not his ambition, Warnier agrees if you want to design a good system that people are going to use, you really have to think about how people want to use their systems.

Even though it is not his ambition, Warnier agrees if you want to design a good system that people are going to use, you really have to think about how people want to use their systems.

A lot of people don’t understand how a system works in the first place, the tool is very complex, and people do not want to know how it works. Like they do not want to know how their car works as long as it drives. Designers and all others who make the system, try to make their system more humanlike because it makes it easier for humans to interact with systems. Doing so though, they are actually blurring the issue of using a system that is a tool.

This blurring between tool and intelligent system is more and more a problem argues Warnier. Systems cannot witness, but systems become more powerful and people connect differences to each other and find new ways to exploit data that are already in the system. The data become more accurate and more things are interconnected. It can lead to some good things but can also leads to abuse. There are the European laws, like the European Data Directive, that defines a lot about what companies and governments actually are allowed to store, which kind of information, for how long, and that they have to inform people. Nevertheless, very few people understand what is at stake. Warnier thinks though that changing the research agenda will not make a difference. He does not believe that current society can inform research to what it should and should not work on. One cannot prevent people from studying and inventing new stuff. Also, researchers themselves do not know what the results of their research or the implications of their research will be in say in fifty years time. When discussing the example of the child dossier that is recently introduced in the Netherlands, Warnier argues that the systems engineer or the computer scientist, or the mathematician, can not be responsible for how a couple of civil servants decide to use a system.

Expand selection Contract selection

Transcript Warnier

View full transcript including film fragments here

Hereunder the transcript in text.

Expand selection Contract selection