[WF-General] I had a dream, or 4.

Tess Snider malkin at Radix.Net
Sat Oct 7 19:25:34 PDT 2000


On Sat, 7 Oct 2000, Richard Wilkins wrote:

> Probably already been mentioned in the past, but I don't know that. It
> may be impractical in programming and, in particular, a program as
> lag-aware as a game. Using principles of energy & work, movement could
> be variable, depending on a number of factors. As an example a simple
> collision will be used. Object A colides with object B.

Believe it or not, lag is a non-issue when it comes to physics simulation.
There is a little magic trick that you do, called "prediction," to mask
the effects of lag.  Basically, the client program needs to know the same
physics rules and algorithms as the server.  The client uses its best
guess of what should be happening, based on its knowledge of the
situation, regardless of how mmuch lag is going on.  The server need only
squeeze through a few packets, so that the client can correct its
worldview, if it has gotten out of synch with the server's idea of what is
going on.  The server is always authoritative, so you meddle with the
client's physics model at your own peril.  :)

That said, there are still problems with intensive physics simulation.
In particular, it can be extremely computationally expensive on the
server-side to track all of the statics and dynamics of a world at once.

Ultimately, I believe thast there are folks who are very interested in
doing some physics modelling in WF, so it wouldn't surprise me if there is
some dialogue about this in the future!

Tess





More information about the General mailing list