[fluxus] every-frame functions

[nik gaffney]

> that's a neat solution... I think the multiple interpreters approach would
> be the best fix, as the scripts wouldn't be able to mess with each other -
> but I'm worried about the cpu overhead. Not that you'd have to use them, I
> suppose...

how much overhead is involved with each new interpreter? im not so
familiar with scheme, but looking at R5RS gives the impression that it
might be possible to use enviroments to separate evaluation
contexts. there may be issues in how guile handles this, or if in fact
there is any sense in it whatsoever. ..