Polymorphic type inference
Daniel M. Leivant · ACM SIGACT-SIGPLAN Symposium on Principles of Programming Languages (POPL) · 1983
The benefits of strong typing to disciplined programming, to compile-time error detection and to program verification are well known. Strong typing is especially natural for functional (applicative) languages, in which function application is the central construct, and type matching is therefore a principal program correctness check. In practice, however, assigning a type to each and every expression in a functional program can be prohibitively cumbersome. As expressions are compounded, the task of assigning a type to each expression and subexpression becomes practically impossible, even more so because the type-expressions themselves grow longer. It becomes imperative therefore to design friendly programming environments that permit the user type-free programming, but that generate fully typed programs in which the types of all expressions are inferred by the system from the program. For interactive functional programming environments of the kind implemented for the Edinburgh functional programming language ML, a type-inference system is an invaluable tool for on-line parse-time error detection and debugging.
Cited by 179 works. doi:10.1145/567067.567077 Open-access PDF
Authors
References
- A Machine-Oriented Logic Based on the Resolution Principle
- A theory of type polymorphism in programming
- The Lambda Calculus - Its Syntax and Semantics
- Principal type-schemes for functional programs
- The principle type-scheme of an object in combinatory logic
- Towards a theory of type structure
- Constructive mathematics and computer programming
- An extended polymorphic type system for applicative languages
- A filter lambda model and the completeness of type assignment
- An investigation of a programming language with a polymorphic type structure.