Self-Interpretation and Reflection in a Statically Typed Language

Reflection is the ability of a system to perform a computation about itself. This ability typically includes a way of representing programs as data (“reification”) and of executing representations of programs (“selfinterpretation”). The interpreter is accessible to the interpreted program in the form of an “eval” function. Reflection is traditionally studied in untyped or dynamically typed languages such as LISP [4] [2], Smalltalk [3], or the λ-calculus [9]. By contrast, we consider self-interpretation and reflection in a statically typed metalanguage along the lines of [10] and [7]. Since the language is statically typed, the data structure used as a representation for programs is statically typed as well. Reflection in a statically typed context can be characterized as follows: