Physics Simulation Forum

Hi, Thanks very much for the plugin!

I'm only starting to familiarize myself... is there a roadmap / todo / in progress / plan for Dynamica? It's already quite neat, but there is so much potential for it; I realize the first thing is going to be more robustness (I'm trying to identify some of the bugs I encountered to be fully reproduceable) and artists will want features like cache / bake, set to initial, etc. (some things can be done in MEL, I'll put some together). Perhaps more of Bullet's built-in features can be folded in? More shapes, like heightfield, constraints, motors, sleep+threshold, etc. sound nice.

Oh, as a side note, this is most likely not Dynamica, but the Bullet engine itself: I noticed sphere primitives are a little unstable together. It's funny because mesh spheres appear stable, but it seems primitives should be more efficient. Sphere primitives tend to crawl and spin when they are put together, a quick minor example with only two spheres in Dynamica:
The top sphere will crawl away from the bottom one after it lands on it, when I play it anyhow (using the 2.71 rev3 binary posted here). If this is normal, fine, I was just surprised given how stable all other shapes were.

We plan to fix the rotation issue, reported earlier, and we will add COLLADA Physics export and import for upcoming Bullet 2.72 Dynamica release.

You can check out existing issues related to Bullet and Dynamica here:
http://code.google.com/p/bullet/issues/list
There is no other roadmap yet, development is steered by feedback.

This was an embarrassing bug in latest Bullet, and has been fixed in Rev5, please download and try it out:
http://code.google.com/p/bullet/downloads/list

Thanks for the feedback,
Erwin

[...]Thanks, that will help tons, most of all for shatter effects; but also for basic object placements, even Maya's bounding box was messed up with mesh rbs.Good to know... Wow you're fast! I'll go and download it right now. I got a hold of the source code (r1316), spent a few hours looking through the daunting code, but I did stumble upon clearManifold() vs refreshContactPoints() in btSphereSphereCollisionAlgorithm.cpp, took a chance and uncommented the first one (?) and spheres seemed to work after that; luckily my PC didn't blow up.

I'd like to add my congratulations, the Bullet physics engine is very solid, I'm very impressed. Now I just wish Bullet could be fully exploited within Maya...

Hey RBD, I'm the software manager at Walt Disney Animation Studios who kicked off this project and cleared the open source release, and the developer was the excellent Nicola Candussi in my group.

Part of our interest in releasing this project open source is that we don't have enough resources at our studio to take it through all the roadmap items that there is potential for. Our internal development will only be driven by our artist needs/requests, but we intend to keep an eye on forum posts like this to extract any ideas that may be beneficial. We will certainly roll any enhancements back into the community. But, we hope that other independent developers and studios who are excited by the potential will also contribute code back so that it's a truly collaborative development project.

That said, please continue to contribute ideas for a desired roadmap on these forums, and I'll keep monitoring and providing feedback as to what items we may wish to implement on our side, as well as just to absorb requests from the community.

Thanks!
Arthur

I certainly hope to contribute what I can to this project, and would like to think that anyone else, including artists, who see the potential might as well, even if it's simply in the form of suggestions. I think it's a great move, it should pay off, assuming other's will also take and give back for art's sake.
Thank you for the explanation Arthur.