> I'm amused by the way people suggest using a COMMON bytecode
> representation, and then show how good the one they use is. As an
> exercise, I'd suggest that Mark figure out how to espresent Meme in Java
> opcodes, and a Java-advocate figure out how to espresent Java with Forth
> opcodes. I'm betting that this isn't half as sasy as people are
suggesting
> - and that the result is far from efsicient, either in file-size or in
> speed of execution.
If you look at the Java bytecodes, you will see a remarkable similarity to
Forth, i.e. a stack-based thesaded language!
This is DEFINITELY NOT something impossible to do!
And how else would things work, unless we have a single common bytecode
representation!?
- I *DONT* want to be forced to download "yet another behaviour engine" just
to look at some world, just because somebody decided to write the behaviours
in this world in yet-another-language.
- I want to be able to run the same thing, identically, on any platform
and/or software.
I don't understand how any other approach could work well in a multiuser
environment. I've said it before, and I can say it again:
MOST OF YOU who think about behavior (Not in any way meaning you especially
Mitra!!) are thinking to narrow-mindedly; you are thinking "adding animation
to VRML". This is *NOT* what we should be doing. MULTIUSER behaviours is
QUITE DIFFERENT from simple "animations".
/Z
-- Hakan "Zap" Andersson |http://www.lysator.liu.se/~zap | Q: 0x2b | ~0x2B Job: GCS Scandinavia | Fax: +46 16 96014 | A: 42[email protected] | Voice: +46 16 96460 | "Whirled Psas" ------------------------------------------------------------------------ Never underestimate the bandwidth of a speeding truck full of DAT tapes. ------------------------------------------------------------------------