some suggestions

[delerium10101]

Well, interesting catch here. Sure, we could say \"corpse is here\" but the precise arrangement is up to the client, however, collisions would get very screwy with the corpse (you will not be able to walk through corpses like it was nothing). And other things may collide with the corpse, like shrapnel. If the arrangement of the various limbs differs from computer to computer, we can\'t do things like that. The server says one thing, client A sees something different, and client B sees something different still. The server might say that shrapnel did hit client A, while client A sees the shrapnel blocked and client B sees some shrapnel hit and some miss. Personally, it\'s little details like using your fallen comrades as meat shields that add that extra immersive factor to a game, and unless client/server are in agreement, we can\'t do that without making people frustrated. And yes, collisions are in fact that accurate. No bounding box collisions for us, no no. If we do ragdoll, we will probably use the cheap cheat known as \"precomputed ragdoll\" which really means \"ooh, spiffy animation!\". Much *much* cheaper. A single byte transmitted once to indicate the animation, and no additional overhead beyond that.

And two X1800XTs? I am officially obligated to hate you.