Networked Physics

[polylux]

A while ago I came to the point to decide how to calculate the physics in my project. Since it's laid out as a multiplayer game I've had quite some thinking going on about how to implement the physics. Should the client compute the physics, the server, a mixture of both? Either way there were a lot pros and cons.
Just now I stumbled upon this article by Glenn Fiedler, introducing a hybrid approach allowing the clients to compute their own physics (makes it all smoother) as long as the position difference of the client- to the server object lies within defined boundaries.
Long story short, >>here<< is the article.
Hope it helps some fellow programmers!

Cheers,
p.

[ent1ty]

I think this is a (relatively)very old and widely used approach, isn't it?

[polylux]

Well I guess so. It's just an article that describes it well, that's why I wanted to share it.