🚀 go-pugleaf

In article <28750:Sep2402:58:2290@kramden.acf.nyu.edu>
 brnstnd@kramden.acf.nyu.edu (Dan Bernstein) writes:
>
>I'd like to focus on this point. Upward extensibility through macros and
>functions is pretty well understood. Its syntax may not be as well
>understood, but that's not a serious problem. What would it take to make
>a realistic language downward-extensible?

You'd want the components of an abstract machine model to be described
in the high-level notation, with an obvious, efficient implementation
(e.g. RegisterA := RegisterB + RegisterC
        would be implemented as a single low-level instruction);
and you'd want parameters of the abstract machine to be constants
in the environment
(e.g. TimeForAddition
        would return 12 nanoseconds, or whatever).
Then, each of the operations might be implemented on each of the
machines, either directly or as sequences of other operations;
and the conditionals would be in terms of
        If TimeForAddition(AbstractStackMachine)
                < TimeForAddition(AbstractVectorMachine) Then
            Push(Pop1+Pop2)
        Else
            VectorR1 := VectorR2 + VectorR3.
This obviously isn't quite right, and there would have to be
relationships defined among the abstract machines; but I think
this approach might be the sort of thing you're looking for.
[Clearly, when the compiler can figure out which branch is to
be taken, only the code for that branch need be generated.]

-----------------------------------------------------------------------------
 Chris.Holt@newcastle.ac.uk      Computing Lab, U of Newcastle upon Tyne, UK
-----------------------------------------------------------------------------
 "What two ideas are more inseparable than Beer and Virtual Reality?"