[plt-scheme] xxx chooses MzScheme as preferred language

From: Edmund Dengler (edmundd at eSentire.com)
Date: Thu Jun 1 15:58:51 EDT 2006


Because all projects are a multi-variate optimizations with different
trade-offs basedon the specific project: performance, ease of use,
implementation available, library, language constructs, etc.

Don't get me wrong, there are a number of features of MzScheme that I
really wish to have for out projects: modules, the event model, the thread
model, wills, decent libraries, ... . But for our projects, lack of
regional allocation and lack of optimizations just overwhelmed the
equation and forced us to not to be able to use MzScheme. Most of the
"high performance" schemes may score higher on the performance front, but
not enough to justify the lack of a good event model or lack of general
libraries (Gambit, for example), or just plain did not work (anybody
manage to get a recent Bigloo to work on OpenBSD?).

For other people, the equation plays out differently, and different
choices are made. Not that they are right or wrong in general, just that
different trade-offs were made and they made the right decision for them.

I suspect that there are many projects that do not have the high
performance requirements, do not require building a finished small
executable, and can benefit from the libary and language support of
MzScheme, which is why it is quite popular. Is there anybody out there
using it for a performance-critical project?

If I could have a dynamic language that allowed optional typing/contracts
to allow for high performance optimizations with better memory management
to avoid unnecessary garbage collection and support for distributed
computation/work in a hostile environment with better namespace control
and metacompilation along with extensive reusable libraries, then I would
be happy. But _no_ language has all of these, and most languages
(especially the research oriented languages such as Haskell and Oz) make
very deliberate choices as to what they consider important. Most of the
time, the choices are away from what we can use in large high performance
projects, and so we usually have to stick with C/C++, even if this is not
the language we want.


On Thu, 1 Jun 2006, Joe Marshall wrote:

> On 5/31/06, Prabhakar Ragde <plragde at uwaterloo.ca> wrote:
> > Obviously there are times when speed is the bottleneck and the deciding
> > factor. But I suspect we computer scientists tend to make it an issue
> > when it is not, simply because it is easier to compare hard numbers than
> > the more vague and softer qualities associated with the broader goals I
> > mentioned.
> What is the *fastest* Scheme implementation?
> What is the *most popular* Scheme implementation?
> If performance matters, why aren't these the same?
> --
> ~jrm
> _________________________________________________
>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme

Posted on the users mailing list.