[r6rs-discuss] [OT] Should compilers preserve termination?

From: Paul Schlie <schlie>
Date: Fri Mar 9 20:22:04 2007

> Thomas Lord wrote:
> ...
> "Returning from the loop" means invoking one of its continuations,
> passing it updated values of implicit parameters like the "store" and
> the "ports". What you are thinking of as "temporal and value state
> side effects" is still, when we talk about the formal meaning of Scheme,
> part of the "return values".
> ...
> That "other kind of math," generally speaking, involves creating
> axiomatic systems that characterize the classes of constructible
> values, including both ordinary parameters and implicit ones
> like "store" or "ports". You still wind up talking about and modeling
> things like the temporal properties of observable side effects of
> a program -- but in doing so, you're still talking about the "value
> of an expression".

- agreed (as long as it's clear that a returned value includes explicit
  as well as implicit program state side-effects inclusive of their
  sequential ordering as may itself be potentially dependent on and/or
  synchronized with externally controlled event state); and thereby all
  aspects of an expression's returned value must be preserved which may
  affect a program's observable behavior, but not otherwise.
Received on Fri Mar 09 2007 - 20:20:48 UTC