Real-Time Physics Simulation Forum

Hi

I'm a student in computer science and I'm very interesting in physics engines. I'm trying to code a basic physics engine starting from a simple implementation of collision detection and response for rigid bodies in 3D. I'm gonna improve it while I'm learning new things about physics simulation. My goal is not to develop the best physics engine but I'm doing that to improve my skills in physics simulation.

I'm reading the book "Game Physics" from David H Eberly. I'm using this book in order to help me coding my basic physics engine. I've seen a post on this forum where someone said that this book is good but doesn't teach you how to write a physics engine.

I wanted to know why this book doesn't help you to write a physics engine ?
Is it a good things to start with this book ?

Thank you very much to help me.

Well, I think I wrote this. Basically the book covers a lot on physics, but e.g. the constraint solver works on the acceleration level as described by Baraff and this is not how current engine work. Still there is plenty of good stuff in the book, but more physics in general and not so much engine development in particular. The good news is that there is a lot of stuff available for free and Erwin hosts a lot of these references here on his side. Personally I would look into these sources:

- Erin Catto Box2D Lite and GDC presentations
- Kenny Erleben PhD
- Garstenauer Master thesis
- Bullet and ODE source code

This should get you started. I would recommend to take Box2D Lite and port it to 3D - which is difficult enough. This should give you a solid foundation from where you can experiment and extend. I found it helpful to take an existing solution as orientation. When I started digging into this I started reimplementing the ODE and to improve on it. In the beginning I had the quickstep solver running in the background to verify my results. Once you catch up you can start to extend your engine and there is plenty of room for improvement.


HTH,
-Dirk

Ok Thank you very much for your answer.

thanks I will take a look a those books.

dellsmaith wrote: 3. Methods of Mathematical Physics (Paperback) :by R. Courant, D. Hilbert

Hmmm.. I know this book. I dug this one out to study Fredholm integral equations of the second kind in grad school. Maybe a bit deep for game physics though? A classic book, nonetheless.

I should comment about 'Game Physics' by Eberly, as well. Agreed it doesn't have the constraint formulation that seems most relevant, but I like the book a lot for the breadth of topics and the in depth treatment of a number of topics. And the code is great (pedagogically speaking) too (but maybe available on the web?).

Just want to throw in that dellsmaith probably is a spam bot (or at least a spammer). I saw members with similiar names and especially similiar signatures on other boards.

Thanks! He is now banned. Note that you can use the report post function (the red exclamation point) on such posts

I'll take a look a those books to. I think it may be useful to me.