Lol @ the 2007 comment.... Hopefully their engine will be a hundred times more integrated than my project ever was.... and hopefully Novodex is more accurate than Karma 
The key to the non-slideshow aspect is interpolation and approximation. It only sends movement information if it moves a certian amount (though it isnt a large amount) and dosen't send physics data when stopped (stops massive net spam.)
Then I took advantage of the same interpolation inherintly built into UT2004's cars, so a single frame of updated data could be turned into a half second+ approximation of where it would be. It makes it look much more tolerable, even if the object is occasionally in the wrong place because of sparcity of net updates, at least with the new update it will be in the proper place.
It definitely took some work because of the server/client setup. I ended using existing kcar and simmilar net code for the basics, as the engine does almost all replication and physics calculations through native code... so it's hard to find the functions which are called, and even harder to make them do more stuff.
Though finals are approaching for me, I do want to release another version of this.... of course with full documentation and source code. I have come a long way with it.... and don't want to really give up (though it's been some time now)

The key to the non-slideshow aspect is interpolation and approximation. It only sends movement information if it moves a certian amount (though it isnt a large amount) and dosen't send physics data when stopped (stops massive net spam.)
Then I took advantage of the same interpolation inherintly built into UT2004's cars, so a single frame of updated data could be turned into a half second+ approximation of where it would be. It makes it look much more tolerable, even if the object is occasionally in the wrong place because of sparcity of net updates, at least with the new update it will be in the proper place.
It definitely took some work because of the server/client setup. I ended using existing kcar and simmilar net code for the basics, as the engine does almost all replication and physics calculations through native code... so it's hard to find the functions which are called, and even harder to make them do more stuff.
Though finals are approaching for me, I do want to release another version of this.... of course with full documentation and source code. I have come a long way with it.... and don't want to really give up (though it's been some time now)
Comment