Towards More Natural Functional Programming Languages

By Brad A. Myers
Human Computer Interaction Institute
School of Computer Science
Carnegie Mellon University

Keynote talk given at the The ACM International Conference on Functional Programming, ICFP 2002. October 4-6, 2002, Pittsburgh, PA.


Programming languages are the way for a person to express a mental plan in a way that the computer can understand. Therefore, it is appropriate to consider properties of people when designing new programming languages. In our research, we are investigating how people think about algorithms, and how programming languages can be made easier to learn and more effective for people to use. By taking human-productivity aspects of programming languages seriously, designers can more effectively match programming language features with human capabilities and problem solving methods. Human factors methods can be used to measure the effects, so unsubstantiated claims can be avoided.

This talk will present a quick summary of new and old results in what is known about people and programming, from areas that are sometimes called "empirical studies of programmers" and "psychology of programming." Much is known about what people find difficult, and what syntax and language features are especially tricky and bug-prone. Our new research has discovered how people naturally think about algorithms and data structures, which can help with making programming languages more closely match people’s problem solving techniques.