2D Physics Engine
I made the first release of my 2d physics engine : http://inertia2d.sf.net or http://www.sf.net/projects/inertia2d
This engine currently supports RigidBodies, CollisionEngine and Impulse Response. ( For a full list of feature go see the wiki ).
I someone could test the engine and give some feedbacks, it would be nice :)
ps : excuse for my english and for the english on the website but it's not my motherlanguage.
The links on your wiki are hosed (at least they seem to be):
http://inertia2d.sourceforge.net/wiki/index.php/Overview
Any of the links on this page set the action variable equal to "edit". For example:
http://inertia2d.sourceforge.net/wiki/index.php?title=Collision_Engine&action=edit
If I try "view" there is no page. Am I misunderstanding something or did I manage to get to a page I should not have?
Thanks,
#dth-0
http://inertia2d.sourceforge.net/wiki/index.php/Overview
Any of the links on this page set the action variable equal to "edit". For example:
http://inertia2d.sourceforge.net/wiki/index.php?title=Collision_Engine&action=edit
If I try "view" there is no page. Am I misunderstanding something or did I manage to get to a page I should not have?
Thanks,
#dth-0
Nice release,
Im sure this will help a lot of people looking to get into this stuff,
Some examples would be nice so that we can see the capabilities of your code without having to build a project around it and compile
Im sure this will help a lot of people looking to get into this stuff,
Some examples would be nice so that we can see the capabilities of your code without having to build a project around it and compile
xiuhcoatl > I know this is normal. It's because I haven't written anything in this section. I'll try to fill it as soon aspossible
Structure > Yes I know, I planned to release a binary version with some more examples ( using ClanLib maybe )
Structure > Yes I know, I planned to release a binary version with some more examples ( using ClanLib maybe )
Ive spent a bit more time looking over your code to see what you implemented and I must congratulate you for writing a flexible system with very readable source code.
I especially like the template system you’ve implemented for numerical precision changing, something I think ill put in my own code.
As a quick constructive criticism, I noticed you had a lot of variables that could be cashed rather than calculated each step like 1/mass etc, this could help speed things, but im sure you have your reason for that.
If possible could you explain where you gained the knowledge to write the LCP solver?, as im interested in learning how this works.
[edit] ive just realised you used Dantzig Algorithm, as suggested by Baraff,
Have you had any luck incorporating friction into this cos I looked at the paper and couldn’t work it out.
I especially like the template system you’ve implemented for numerical precision changing, something I think ill put in my own code.
As a quick constructive criticism, I noticed you had a lot of variables that could be cashed rather than calculated each step like 1/mass etc, this could help speed things, but im sure you have your reason for that.
If possible could you explain where you gained the knowledge to write the LCP solver?, as im interested in learning how this works.
[edit] ive just realised you used Dantzig Algorithm, as suggested by Baraff,
Have you had any luck incorporating friction into this cos I looked at the paper and couldn’t work it out.
Friction for the impulse is quite easy to implement
But I must say it's quite difficult for resting contacts. For the moment resting contacts aren't implemented at all. But I plan to add it for the next release.
For the 1/mass, I know, I should really store de inv_mass instead of the mass. There is a little bug though in the diff equation solver, the mass isn't correctly used.
But I must say it's quite difficult for resting contacts. For the moment resting contacts aren't implemented at all. But I plan to add it for the next release.
For the 1/mass, I know, I should really store de inv_mass instead of the mass. There is a little bug though in the diff equation solver, the mass isn't correctly used.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement