Under consideration for publication in J. Functional Programming 1 Exceptional Syntax.pdfVIP

Under consideration for publication in J. Functional Programming 1 Exceptional Syntax NICK BENTON and ANDREW KENNEDY Microsoft Research, St. George House, 1 Guildhall Street, Cambridge CB2 3NH, UK (e-mail: {nick,akenn}@) Abstract From the points of view of programming pragmatics, rewriting and operational semantics, the syntactic construct used for exception handling in ML-like programming languages, and in much theoretical work on exceptions, has subtly undesirable features. We propose and discuss a more well-behaved construct. Capsule Review The propositions-as-types principle has often been cited as an influence on the design of functional programming languages. Often the influence is seen only indirectly. In this short note the authors draw lessons from the proof theory of disjunction to suggest changes to the syntax of exception constructs that not only improve their utility in programming, but also admit simpler expression of common program transformations. 1 Introduction Many programming languages (from Mesa and PL/I to SML, Java and C#) include exceptions to provide a structured, but non-local, way of signalling and recovering from error conditions. Programmers often also use exceptions as convenient, and sometimes more efficient, way of varying control flow in code which has nothing to do with what most people would consider error-handling (for example, the parser combinators in (Paulson, 1991)). The basic idea of exceptions is simple and familiar: the evaluation of an expression may, instead of completing normally by returning a value or diverging, terminate abnormally by raising a named exception. The evaluation of any expr