wiki:ModuleSystems

Introduction

As the charter states, "the purpose of this work is to facilitate sharing of Scheme code." To this effect, a module system is explicitly given as a requirement in the charter. We need to consider all the different types of module systems supported by R5RS and R6RS implementations and come up with a proposal that can gain widespread use.

Module systems serve the purpose of encapsulating code and managing namespaces. Although lambda lets us manage local identifiers, it doesn't allow us to encapsulate macros, and it wants for a friendlier interface for module-like uses.

There are a number of fundamentally different ways in which modules can be provided. The major types of module systems include:

First-class environments that can be passed as a second argument to eval (or the equivalent) can be used as a form of encapsulation, where a module is the set of bindings defined in the environment. If macros are not first-class, the environments would need some interface for passing back and forth macro bindings between environments. The disadvantages of this approach are that many implementations do not support such first-class environments, separate compilation and phasing becomes more complicated, and static analysis is almost impossible.

Another natural approach to creating modules is to build them on top of, and allow them to be composed with, macros, and allowing import forms to be expanded from macros. It is relatively easy to implement this for non-syntactic bindings only, and one sample portable implementation can be found in lexmod. Allowing importing and exporting syntax requires non-portable extensions. Chez Scheme uses this approach.

An alternative approach to extending a module system is to provide one or more extensible languages in which to compose the modules. Generally, these languages are described within the module itself. PLT takes this approach with multiple languages, which can also modify the read syntax and other features such as the semantics of function application. Scheme48 takes this approach with a single meta module called the config module, which is globally extended when you want to add new module syntax. In either of these cases, you obtain the flexibility to extend the module syntax with new features, but at the same time provide a separation between the module meta-info and the actual body of the module itself, allowing a certain amount of static analysis without actually expanding the module.

R6RS provides a simple static module syntax which provides just the basics needed for importing and exporting (optionally renamed) identifiers from other modules. Neither the module itself nor the import or export forms may be expanded from macros, so the only way to extend the syntax is with future standardization efforts. The primary reason a static module system is desirable is because by specifying only the syntax it can be implemented by any of the above module systems. Implementations can then extend the syntax as needed when not concerned with portability, but a common ground exists when no advanced features are needed.

Additional Issues

Existing Proposals

Last modified 6 years ago Last modified on 10/10/10 21:08:59