Plans for ODE integration

[Hannibal]

Hi,

I'm new to the bullet library, but quite experienced using other commercial and non-commercial physics libraries, and I love using continuous collision. I am currently working on a spare-time project where I have ODE, but I have heard about the plans of integrating Bullet into ODE, and to me that sounds like the best of both worlds.

I can see that it has been on the todo-list for a while, and I was wondering when you think it will be completed? Or when it will be started?

Also, I am curious about if the continuous collision will require any changes to the actual solver, or if it is a straight collision fix.

By the way, I must congratulate you on having an impressive rate of releasing updates. It is very impressive seing the rapid changes. Keep it up

[Erwin Coumans]

It has been lower priority, because of recent refactoring the Bullet code. You can already use ODE quickstep and ODE box-box in the Bullet framework. So I would call this the best of both worlds.

I will still add Bullet Collision to ODE, but that's still in inferior solution compared to Bullet+quickstep. What feature of ODE is missing in Bullet?

Thanks,
Erwin

Hi,

I'm new to the bullet library, but quite experienced using other commercial and non-commercial physics libraries, and I love using continuous collision. I am currently working on a spare-time project where I have ODE, but I have heard about the plans of integrating Bullet into ODE, and to me that sounds like the best of both worlds.

I can see that it has been on the todo-list for a while, and I was wondering when you think it will be completed? Or when it will be started?

Also, I am curious about if the continuous collision will require any changes to the actual solver, or if it is a straight collision fix.

By the way, I must congratulate you on having an impressive rate of releasing updates. It is very impressive seing the rapid changes. Keep it up

[Hannibal]

Since I am new to Bullet, maybe I am basing my opinion on older comments in the ODE thread - where I got the impression that this integration would be the best solution, when it was done.

If you are happy with the bullet+quickstep then I'm perfectly happy using that. I was afraid that that was a stop-gap solution that would die once Bullet was integrated into ODE.

Thanks for the quick reply

[Erwin Coumans]

Since I am new to Bullet, maybe I am basing my opinion on older comments in the ODE thread - where I got the impression that this integration would be the best solution, when it was done.

If you are happy with the bullet+quickstep then I'm perfectly happy using that. I was afraid that that was a stop-gap solution that would die once Bullet was integrated into ODE.

Thanks for the quick reply

I can recommend trying out the new Bullet 2.3x (or later). You probably don't even need to use quickstep. The only missing bits in Bullet's constraint solver is lack of constraint motors and slightly less performance as quickstep. I'm working on closing those gaps.

Benefit of using Bullet's constraint solver is more flexibility and customization (for example better friction model), and its Genetic 6 DOF Constraint can be customized similar to Ageia PhysX. This means its compatible with the COLLADA Physics specification.

Feedback of issue are welcome and will be incorporated in frequent Bullet updates.
Erwin