Re: Is interpret/compile the wrong distinction?

On Mar 24, 12:48 am, John Doty <j...@[EMAIL PROTECTED]
> wrote:
> In LSE, the factors of the system's immediate words are, in effect, the
> metaprogramming toolkit. You don't worry about POSTPONEing anything: you
> just use its (non-immediate) factors. If you need to compute something
> mid definition, you define a word for it and make that immediate.

The way to encapsulate this in a Forth with assembled or metacompiled
primitives is to follow a discipline of never making a primitive
immediate, so that there is never an immediate word without non-
immediate factors.

In an implementation, or indeed in a standard for a family of Forth's
designed under the above discipline, this could be enforced by not
having an "IMMEDIATE" word that acts after a definition has been made,
but rather having a compiler word for immediate words.