[r6rs-discuss] Implicitly Concurrent Scheme

[Aaron W. Hsu]

Hello Aubrey,

--On Thursday, November 12, 2009 02:21:50 PM -0500 Aubrey Jaffer 
<agj at alum.mit.edu> wrote:

> I am suggesting the following restriction be *removed* from the
> "Procedure calls" section of RnRS:
>
>     _Note:_ Although the order of evaluation is otherwise unspecified,
>     the effect of any concurrent evaluation of the operator and
>     operand expressions is constrained to be consistent with some
>     sequential order of evaluation.  The order of evaluation may be
>     chosen differently for each procedure call.

It has been argued why this is a bad idea. In my limited knowledge,
there is not a proof of this concept. Maybe I'm missing it or have
missed it, but is there some implementation or interpreter which has
this particular optimization (obviously, this can't be done in the
general case right now and still comply with Scheme specifications, but
that's beside the point here) that demonstrates consistent general
improvements in execution time on programs for which it would be
completely intractible or far too slow to do analysis of the programs?

  Aaron W. Hsu