Lambda the Ultimate

Extending the Scope of Syntactic Abstraction by Oscar Waddell and R. Kent Dybvig, POPL '99. (Also: Waddell's thesis with the same title.)

The benefits of module systems and lexically scoped syntactic abstraction (macro) facilities are well-established in the literature. This paper presents a system that seamlessly integrates modules and lexically scoped macros. The system is fully static, permits mutually recursive modules, and supports separate compilation. We show that more dynamic module facilities are easily implemented at the source level in the extended language supported by the system.

This paper is probably known to many LtUers, but it's never been posted, and I find it very enjoyable.

It introduces two very simple forms, (module name exports-list body) for naming and enclosing a body of code containing definitions and expressions, and (import module-name) for importing the definitions from such a module into the current scope.

Module names are lexically scoped just like variables, and modules can appear wherever definitions can occur, so one can define modules (say) inside a lambda. Furthermore, modules and import forms may also be generated by macros.

They show how more advanced features (such as qualified references ("module::var"), anonymous modules, selective importing and renaming, and mutually recursive modules, among others) can be built on top of this simple base using a hygienic macro system, cleverly and without much fuss.

Side note: such a "syntactic" module system can be contrasted with R6RS's "static" library system. There is currently some discussion to this effect in the Scheme community.

What advantages does purity offer in a programming language? I often hear (read) about these advantages, but never with concrete examples. I figured here is the best place in the whole www to ask this question :)