[ODE] PSCL implemented in Rocket

[John Ratcliff]

I am glad people like the Rocket demos and are interested in the PSCL
language.  Based on the feedback I have received here, I have decided to
create an ODE implementation of the physics virtual interface for PSCL.
I will release this open source even prior to the full source release.

I am heading out of town for a week to attend Meltdown and shortly after
that attending Siggraph.  While I am out of town next week James Dolan
(lead on Fatal Traction, www.fatal-traction.com, that used ODE for its
physics) is going to create the ODEPSCL implementation.  We will then
upload a new copy of Rocket.  If you edit the 'rocket.ini' file you will
be able to have it run using ODE instead of Novodex.  It may not be a
complete implementation, but you will at least be able to run some basic
scripts with it.  

Once I make the source available to the ODE community I hope somebody
will step up and complete the implementation.  It would be great if we
could use Rocket to run physics scripts on multiple engines.  I am not
going to support the split-screen option though, that was cool but
maintenance was a pain in the butt.

I want to reiterate that my goal is that you should ultimately be able
to write highly complex physics demos using a combination of PSCL and
another scripting language.  The current scripting language we offer if
Lua, but I am interested in also adding support for something else, like
Small C, C-Sharp, etc.  

I will have James give the ODE mail list updates as development of the
ODE implementation for Rocket/PSCL progresses.


-----Original Message-----
From: Jon Watte [mailto:hplus-ode at mindcontrol.org] 
Sent: Thursday, July 22, 2004 8:30 AM
To: John Ratcliff; ode at q12.org
Subject: RE: [ODE] PSCL implemented in Rocket


> To be quite honest, this was documented in the documentation for 
> dWorldStep(), with the suggestion that dWorldStepFast didn't suffer 
> from the problem :-)

Oops, non-sequiteur ;-)

STACK size being the problem, this is still documented in the 
documentation. Bump up the default Windows stack size, or use 
one of the alloca replacements.

I really think ODE should make those replacements the default, 
and make dWorldQuickStep() the default stepper, to make most 
of the common complaints about ODE behavior go away. Document 
the others in an appendix somewhere, so newcomers don't 
accidentally run into them.

Cheers,

			/ h+