Home

Awesome

Rebol (Ren-C) Whitespace Interpreter

This is an interpreter for the Whitespace programming language, using the Ren-C branch of Rebol3.

Although Whitespace was invented as a joke (or at least, released on April Fool's Day), making a working implementation has much of the spirit of a "real" programming task. The existence of implementations in Haskell, Ruby, Perl, C++, and Python mean it is possible to take a real look at the contrast in how you can approach the problem (see the Whitespace Corpus for other implementations).

Specified in Terms of a "Whitespace Dialect"

Any sensible implementation of a task like this will be somewhat table-driven. Otherwise you will end up with unmaintainable spaghetti code.

But the Ren-C implementation attempts to do something novel by first bending the Rebol language itself so that it becomes customized to seem like a language designed for implementing whitespace interpreters.

A dialect is made as a library which defines a CATEGORY and OPERATION. This allows the whitespace.reb code to be written in a way that naturally mirrors the specification.

Here is a small piece of that from the "Stack Manipulation" category:

Stack-Manipulation: category [
    IMP: [space]

    description: {
        Stack manipulation is one of the more common operations, hence the
        shortness of the IMP [space].
    }

    push: operation [
        {Push the number onto the stack}
        space [value: Number]
    ][
        insert stack value
    ]

    ...
]

This shows a novel way of defining the PUSH function, where the category supplies an IMP prefix (space). The operation definition says PUSH is signified by the pattern of a second space, after which it will extract something matching the whitespace pattern for a number into a variable called value. The body of the function is then able to use the decoded value in the runtime behavior when this instruction is seen.

It's certainly possible to imagine creating such a language from scratch. But the promise (pipe dream?) of Rebol-based languages is to make this bend well within reach of a layperson. And since code-is-data and data-is-code (as in Lisp), the aspiration is that none of the specification need go to waste...the strings in the operation could be reflected as help. In Ren-C the experience is often characterized as trying to be "The Minecraft of Programming".

The devil is in the details, of course. There are still plenty of open questions that have been nagging Rebol derivatives due to its pariah status as "the most freeform programming language ever invented". Hard problems persist in variable binding and datatype semantics. Yet it does show some promise.

PARSE-Based Virtual Machine

As an added gimmick for this implementation, it uses Rebol's PARSE dialect for more than just analyzing the whitespace sequences. It's also the virtual machine!!!

It uses the fact that when you give the parser input to process, you can programmatically move the parser position--back to something you've already parsed, or forward to something you haven't seen yet. Consequently it can be used as a program counter! This happens to work for the whitespace VM, though it's probably not the best solution. It's kind of a pun, and only done to show an axis of flexibility in PARSE.

Historical R3-Alpha Version Included

An early exercise for @hostilefork within a couple months of knowing Rebol's "R3-Alpha" was to try to write a whitespace interpreter. That code is not particularly notable when compared with the current version. But the experiment is being kept alive in this repository as a test of R3-Alpha/Rebol2/Red emulation.

Discuss on the Forum

Discussions on this and other topics are conducted on the Ren-C Forum.

Useful Links