Colloquium Details
Theorems, theories, and things: Computing is more than a science
| Author: | Ted Kirkpatrick Simon Fraser University |
|---|---|
| Date: | May 15, 2008 |
| Time: | 15:30 |
| Location: | 220 Deschutes |
| Host: | Sarah Douglas |
Abstract
As researchers in computing, we claim our field is most definitely a science---and we have the gargoyles to prove it! Yet many of the touchstones of science are conspicuously absent from computing "science", or present only on the periphery. Experiments are rare and empirically-supported quantitative laws still rarer. Such empirical results that we do posess seem far more context-sensitive than the universal laws forming the common currency of physical science.
I will argue that these problems arise not because we are immature scientists but because many principles of computing should not be studied scientifically. Instead of a scientific dialogue between abstract theories and concrete experiments, we need to develop a new dialogue between abstract designs and concrete instances. From this viewpoint, computing is a three-fold field, built from elements of mathematics, science, and a third component I call "empirically-supported design". I will describe the paradoxes resulting from our attempts to force-fit a scientific model on computing research, highlight some fundamental properties of our field that may make it intractable to scientific study, and sketch some approaches that might be more productive.
Biography
Ted Kirkpatrick is an Assistant (soon to be Associate) Professor of Computing Science at Simon Fraser University in Burnaby, British Columbia, Canada. He received his Ph.D. in 2000 from the CIS Dept., University of Oregon. He has published papers in haptic interfaces, visualization, and interaction techniques. A former positivist, he has taken the first step of recovery and admitted that he cannot control his urge to experiment. This talk is his effort at the second step, channeling that urge to productive ends.