Racket vs Scheme macros

23
votes

Do racket macros have more advanced functionality than those found in Scheme or Common Lisp? I suspect so, especially regarding modules, namespaces and scoping, etc, but I'd appreciate a simple rundown of what Racket macros can do, if anything, that other lisps can not.

Additionally, Scheme/CL also expose the reader layer to the developer, and using this, is it not possible in those lisps to create entirely new languages (not just s-expr macros), such as Scribble, the way you can in Racket?

In other words, is Racket simply a philosophy/convention around "language oriented programming" with convenient syntactic wrappers to this end, or does it more fundamentally extend Scheme to do things technically impossible in that language?

Thanks a lot.

3
macrosschemeracket

3 Answers

17
votes

Two key features that make "language-oriented programming" work in Racket are:

  • the module system allows the specification of a base language
  • context-sensitive macros

The first bullet is basically what #lang does. Having the module system be in charge of setting up the reader and the bindings available in a given module allows the easy use of different languages. Without this, it'd be more cumbersome to work with languages like Scribble or at-exp.

The second is that Racket provides overrideable context-sensitive macros such as #%app, #%module-begin, and so on that allow customization of, say, every function application within a module or of the entire module. This allows adding, for example, a type-checking pass over the entire module.

This isn't an exhaustive list of features that help build languages in Racket. If you're interested in reading more about it, the "Languages as Libraries" paper is worth reading.

14
votes

I'll recommend reading the paper "Composable and Compilable Macros - You Want it When?" by Matthew Flatt.

It explains how macros and modules work together in Racket.

Then implement a small language in Racket - say TinyBasic. Then compare with the other languages.

UPDATE

http://www.cs.utah.edu/plt/publications/jfp12-draft-fcdf.pdf

6
votes

See also Matthew's reply for the same question after @Scott posted it on the list, with a more recent overview paper.