Re: Mindy future

To: info-dylan@ai.mit.edu
Subject: Re: Mindy future
From: Rob <robm@h2g2.com>
Date: Mon, 22 Jan 2001 10:34:04 +0000
Organization: Pan H2G2
References: <B68F0BC2.3A7B%j.faichney@gu.edu.au> <bruce-925D67.13583520012001@news.nzl.ihugultra.co.nz>
Reply-To: robm@h2g2.com
Sender: owner-info-dylan@ai.mit.edu

Yes, Mindy's Bytecode interpretation is horrendously under-optimized.
It's very easy to read the interpreting code, but very slow to execute.
If we wanted to I'm sure we could optimize it fairly easily.
Mindy is kept working so we can bootstrap d2c on new platforms. It was
designed for UNIX but has been compiled and used on MacOS, BeOS and
Cygwin. It's no good as an embedded interpreter because the compiler and
runtime are separate, and it's no good as a playground for the same
reason and because it is a very incomplete Dylan implimentation.
Its loading of shared libraries and calling of arbitrary symbols is
cool, though.
For embedding or a playground, try Marlais. It's also incomplete, and
old, but it works well enough. 

- Rob.

Bruce Hoult wrote:
> 
> > Also is there any reference for the bytecodes that Mindy uses? Is the
> > bytecode format any good? Is it fairly high-level or low-level?
> 
> High level.  For example, I'm pretty sure that each operator (e.g. "+")
> is looked up from its symbolic name each time it is executed.

-- 
Rob Myers - http://www.robmyers.org/   H2G2 - http://www.h2g2.com/
MacOS wonderfulness for The Hitch Hiker's Guide to the Galaxy Game.
"Smash Global Capitalism! Spend less money!"

References:

Mindy future

From: Jolon Faichney <j.faichney@gu.edu.au>

Re: Mindy future

From: Bruce Hoult <bruce@hoult.org>