Head to Head / Emet Prize Winner David Harel, How Do You Know if There Is an Elephant or a Worm in Your Computer?

David Harel, a professor of computer science at the Weizmann Institute of Science, was one of seven recipients yesterday of the 2010 Emet Prizes, a $1 million prize for achievement in science, art and culture. An Israel Prize laureate, Harel has earned international distinction for his research in the fields of logic, computability and software engineering.

David Harel (Daniel Bar-On)
Daniel Bar-On

Congratulations, Professor Harel. The prize committee cited as one of your greatest research achievements the identification of the key problems in the development of complex systems, in particular software systems, and the invention of a language of Statecharts. What does this mean?

This has been my preoccupation for nearly 30 years - complex systems and their dynamics. It can be your cellular telephone, an F16 fighter jet, a nuclear reactor or any other complicated system. What I look at is not the system's parts or how they're put together, but rather their behavior. How do you get what you want from the system and ensure that it does what you want it to do and so on.

I'm not an expert on fighter planes and large systems. My expertise is in the languages and methods that engineers who build these systems use. One thing I've been doing for many years now is developing and inventing languages. But along with that, you need to devise and develop tools and methods so that these systems develop as they should, so that they do their work and continue to do what we want with the least amount of effort. What characterizes my work is that I've chosen languages which are visual diagram languages. Just as you would find it unthinkable for a staff of architects and engineers to build a house by talking to each other without using diagrams, but only corresponding in writing, for me it would be unthinkable to build a complex system with a code that's only symbolic and textual. What makes work difficult in this field is that in contrast to a diagram of a bridge or a house - static systems - in a dynamic system, the diagrams need to express movement and they have to be programs which can be run, like animation.

Could you describe in detail how this applies to systems we're familiar with?

Almost every complex system we know, like an ATM, a cell phone, passenger planes, fighter planes, the systems that explored Mars and many interactive software systems use positioning diagrams or something similar. They are in very wide use.

A few years ago you suggested creating a virtual life system which would be a kind of complete model of real life. Your first example was that of the c.elegans worm, and you even offered a new version of the Turing Test which would examine the success of the double. How is this research going?

It's going well. But I'm not going to make a model of this worm. In order to get there, we'll have to work very hard for years with many researchers. I threw out the idea as a meta-challenge. I suggested that the worm would be an object worth modeling, because even though it has only 1,000 cells, it's amazing. Like far more sophisticated living things, it has a nervous system, digestive system, reproductive system and a system that allows it to move around.

What fascinates me is building a model in order to understand an entire system and not only one of its processes, which would be a little like lifting the hood over the motor of a car and only looking at one thing inside.

Of course, the question that comes up immediately - it's a scientific and a philosophical question - is how do you know when you've completed the model and whether you have an elephant, a heart or a worm in your computer. I suggested a version of the classic Turing Test of artificial intelligence, which tries to determine whether the system you've built in a computer is similar to a real one. The objective is to see whether an expert can tell the difference between the two systems, and there are technical ways to do this. You separate the human and the machine being tested. And then you can ask what happens when I stand on an elephant's leg. The computer system has to come up with results just like it would in the laboratory.

It took us eight years to make a model of 80 to 90 percent of seven cells. Progress is not linear, but we'll continue to make advances. The advantage of doing something like this is enormous. If you can make a model of an entire multi-celled biological system so that it can pass the Turing Test, the amount of knowledge you acquire is unthinkable.

Professor Uri Sivan of the Technion said recently at a press conference that if the Technion still used the admission tests it used 19 years ago, the situation would be terrible. It appears that his remarks were not only aimed at the institution where he works.

Education in science and technology, and also in culture and tradition, are our only chances of surviving here. What makes a nation stand tall, move forward and take its place is not excellence but professionalism - that everyone is good at what they do. I'm involved in enrichment initiatives for young people, like the establishment of the Anthroposophic high school. These are genuine, good and important things. I think that's a small contribution each of us can make - to educate the young.

You recently signed a petition, along with many Israel Prize laureates and other intellectuals, against the bill that would allow committees in small communities the right to reject candidates who don't meet certain criteria. Can you discuss your opposition to the bill?

I'm not the type of person who demonstrates on street corners, but I will lend my support, despite my time constraints, to causes that seem genuine and vital to me. This law smacks of racism. You can't make a law that discriminates against Arabs. I read the petition and identified with every word, and so I signed it. But it's sad that in order for your voice to be heard, what you've done has to be written next to your name.