The Evaluator

The evaluator’s role is to infer some meaning from the reader-supplied lists of lists.

Evaluation is a slight misnomer. The evaluator has a hidden task which confuses and exasperates in equal measure. If a template is identified then the template is expanded and the result is (re-)evaluated. This is problematic because Idio code is going to be run during evaluation. Ducks in a row, people, ducks in a row.

You get used to it! -ish!

If this was an interpreter then as soon as the evaluator has decided that something is, say, a function call then it can immediately invoke the function value with the argument values (having evaluated everything).

Idio, following in the style of LiSP ([Que94]) instead generates code for the virtual machine and then asks the virtual machine to run it.

Ostensibly, that doesn’t sound terribly different but we’re advancing two causes in this approach:

1. we’ve stopped using variable names (mostly) – although that’s possible for a interpreter too. See LiSP for the details of a fast interpreter.

2. we’re generating lean and mean byte code

   Sort of. Generating byte code at all is a huge leap forward in terms of processing speed.

There are further opportunities in and around the evaluator as well, primarily in the form of optimisation.

• Evaluating
  • Context
  • Fundamental Meaning
  • Specific Meaning
• Evaluating Functions
  • Function Calls
• Expansion
  • Template Definition
  • Quasiquotation
  • Re-Evaluation
  • Expansion Space
  • Timing
  • Mechanisms

Last built at 2024-10-18T06:11:17Z+0000 from 463152b (dev)