Schock prize School of Computer Science, Carnegie Mellon University "Good Luck, a Good Education and Fortunate Accidents" Net Schock Prize for Hillman University Professor Dana Scott

Dana Scott began his education nearly 60 years ago in a one-room schoolhouse

in northern California. At that time, no one could have dreamed he would become

a world-famous scientist, honored by royalty for fundamental advances in an

as-yet-unborn discipline called computer science.

Yet, there he was last month--Carnegie Mellon~s Hillman University Professor

of Computer Science Mathematical Logic and Philosophy--in Stockholm to receive

the 1997 Rolf Schock Prize in logic and philosophy from the Royal Swedish

Academy of Sciences.

Scott was being honored for his fundamental contributions in contemporary

logic and most especially for creating domain theory, a mathematical theory used

for explaining the meaning of advanced computer programming languages. Initiated

in the early 1970s, the work has given scholars the mathematical tools to model

computer programs and predict and analyze what they will do.

"Dana built a bridge between traditional mathematics and the semantics of

programming languages," says Jim Morris, head of the Computer Science Department

in SCS. "He identified a mathematical structure that would correspond to

computer programs and produced a mathematical model for a simple language called

Lambda Calculus. Until he did that, the Lambda Calculus (and all other

programming languages) were understood only through mechanical means. It was as

if our only understanding of conventional calculus was through the formal rules

for differentiation and integration. Dana's models supply the analogues of real

numbers, continuous functions, and the other apparatus that allow one to justify

the formal rules."

"Scott's work completely changed people's thinking about how you'd use math

and logic to reason about computer programs," says SCS associate professor Peter

Lee. "Before domain theory, research into programming languages was pretty

informal--something of a black art. But his work changed the field. The

obligations of researchers to prove their work became mandatory."

[...]

From 1963-69, while Scott was at Stanford as an associate professor and

professor of logic and mathematics, the university was just creating its famed

computer science department. "Many people at Stanford were concerned with formal

logic in mathematics, philosophy and linguistics," he recalls. "It was natural

to bring logic to computer science. I was involved with several faculty and

students there in thinking about applications in computer science."

A dramatic turning point in Scott's research career came in the summer of

1969. He attended a workshop in Vienna about a new field of programming language

design and definition. There, he met the late Christopher Strachey, one of the

pioneers in programming languages. A man with outspoken opinions, Strachey's

influence on Scott's thinking was immediate and profound. During a sabbatical

visit to Oxford that fall, Scott began to formulate a precise definition of

denotational semantics within the beginnings of domain theory to carry out

Strachey's conceptually formulated program. "This intense period of work," he

says, "influenced the whole rest of my academic life."

But Scott insists that he himself was not a pioneer. "I'm an organizer of

ideas and solver of mathematical problems," he says. "Other people were chopping

down trees, and I was busy building them into log cabins."

In late 1969, Scott delivered three public lectures on his new approach.

According to Peter Lee, they stunned the theoretical computer science community.

People spent years trying to comprehend them and their consequences fully. Today

Scott is revisiting domain theory and trying to make the material more

accessible.

SCS professor John Reynolds remembers being in the audience at Oxford for

Scott's second lecture. "It had a life-long impact on my own research," he

remarks. "Dana introduced a new way to describe programming languages that was

far more expressive and flexible than anything known before. In a deep sense, he

showed how to say what a language meant, rather than just how it might be

implemented on a computer."

Scott says Reynolds and many other researchers have made many contributions

to and applications of the theory. And the work continues at many centers of

research. Even today, Scott himself is revisiting the ideas and trying to make

the material more accessible and usable.