[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Independent optimizing compilation

This page is part of the web mail archives of SRFI 83 from before July 7th, 2015. The new archives for SRFI 83 contain all messages, not just those from before July 7th, 2015.

On Wed, 28 Dec 2005, Michael Sperber wrote:

> As to modularity, the whole "linking" issue is certainly very
> important, and the one that's currently receiving (I think) the most
> attention research-wise.  But that one's even harder, and it *also*
> has interconnections with the issues surrounding namespace management.

My requirement for separate compilation may be more stringent than
yours, but if you need to compile modules separately with no order
dependencies, and one module can import both procedures and macros
from other modules, I think that there is, really, exactly one way
to solve the linking problem.  And that way changes the rules of
evaluation so completely that the resulting dialect, unless
artificially restricted, is no longer scheme.  And, unfortunately,
even given such restrictions, its performance is likely never to
match that of a scheme whose macros are all resolved before runtime.