which physics book?

I personally wouldn't recommend it. I got it a while back. If you know nothing about physics whatsoever it might be a good introductory read but otherwise its quite lightweight. I know for example that robust collision detection isn't strictly speaking a physics problem but it should be required content for a book on game physics. The O-Reilly book only covers bounding box collision. Theres a book by a guy called 'Dave Eberly' called 'game physics' which is quite well regarded as far as i know.

D.

 

If you just want 2D physics/collision, check this article out. I found it very good.

 

I got it on Amazon after much consideration, but I totally missed it with this book. Way too heavy for me.
As you're not aware of my level of physics knowledge, my comment may not be of any value to you, but let me put it this way:
It's LOADED with heavy math, so if you're not on an advanced level with math, it will be very hard to follow.

 

Haven't seen Eberly's physics book but his book on 3D game engines reads like a math PhD thesis, which is to say extremely unnecessarily complex. If you want concise, formal definitions of things, give it a try. If you're looking for practical help or tutorial information, skip it completely, assuming it is anything like the 3D engines book, which (despite the way it is marketed) is really only useful as a reference when you want to look up some specific equation.

 

This book:

http://www.amazon.com/exec/obidos/A...5/sr=2-1/ref=pd_bbs_b_2_1/104-3621457-3995135

3D Math Primer and Graphics for Game Development.
ISBN: 1-55622-911-9
Fletcher Dunn and Ian Parberry

It took me forever to get my copy since at the time they had sold all the copies that were published and I had to wait on a used one. BUT, it was *very* well worth it! This book assumes you know nothing about 3D math, and devotes whole chapters to things like coordinate spaces, vectors, matrix math, etc. instead of a few pages like other books. Look at the table of contents there on Amazon and you will see what I mean.

Every other book I've looked at fly through the basics in a chapter (maybe), and are then way over the head of people just starting out. That's where this book fills the gap. Anyway, I'm starting to get redundant. This is a great book and I highly recommend it for anyone who wants to learn the necessary basics and foundations of 3D as applied to computers and gaming.

Oops, I just realized I'm off topic. This book does not really cover physics except for things like intersections and such.

Matthew

 

Baraff's course notes are very good, and free . Chris Hecker's articles are also decent.

I'm seriously thinking about picking up Real-time Collision detection right now. Although it doesn't cover physics collision detection is a big part of a decent physics engine, and this book looks pretty good.

 

I got a few some time ago and none was really helpful, the most useful was the thesys accompaining the free physics DynaMo library.

 

Yeah, I have several of them. I totally recommend the 3D Math Primer book mentioned above as a starting point for the maths (dead sheep as a measurement? cool!), as well as the Physics for Game Developers book, as it's the most sane of the bunch. The Eberly book is an awful starting point. Once you grasp what's going on in the first 2 books, with some further reading into the math notations, you may be able to understand what crack Eberly is on. To compliment the first 2 books, the Jakobson verlet physics article is gold. To truely appreciate the awesomeness of verlet, you need a grasp of the standard stuffs which the physics for game developers book does a great (well as good as I've seen) job of teaching it.

Oh, and the collision detection book in the Eberly series is weak, so watch out. 3D Math Primer has a great chapter on, plus it covers the basics, so go with tha instead.

 

Well, when you talk physics, talking vector math essential, and that topic is best documented for game programmers in the 3D space. Dealing with 3D games does involve 2D here and there, and when you talk about the math, it is as you suggested 1 less axis, or that axis always aligned to zero. Pure 2D physics isn't well explicitly documented, but stuff like dot products are mostly the same and crazy useful, and instead of the cross product, you can get a lot out of tangent vectors (a 2D vector (x,y)'s tangents are (y,-x) or (-y,x) depending on which tangent you want, and which way positive x and y are). Fun fun.

 

I'd still suggest both. Take a look at the article Gnat mentioned (here). If you vector-fu is strong, and you can implement that, then do whatever you want. If you need guidance to do it, then the 3D primer is probably the place to start.

 

It looks like you know your maths and physics. To understand how this math is actually incorporated into a game, you can check out a number of books. Some of them have already been mentioned.

There is one book, however, that is very lightweight and a no-nonsense type, that I would recommend. Keep in mind that this book is based on Opengl. Anyways, the book is 3D Computer Graphics: A Mathematical Introduction with OpenGL.

I mean, at 400 pages, this book is packed with information. It also includes a section on inverse kinematics/kinematics at the end of the book. A quick but very dense read!