same here
Cyan~Fire - Sun Jun 20, 2004 8:58 pm
Post subject:
But can you imagine 20 people trigger-pushing the button like that? That's a lot of packets it sends out. Anyway, this reminds me that there is also, of course, a delay in position packet sending, so the effect would "lag" anyway.
MapMaker+ - Mon Jun 21, 2004 12:42 am
Post subject:
So there is no possible way without programming?
Cyan~Fire - Mon Jun 21, 2004 11:45 am
Post subject:
There is no possible way to make it look nice, period.
There is no possible way at all without making a module for ASSS.
Bak - Mon Jun 21, 2004 3:16 pm
Post subject:
you could adjust it so it predicts where they will be on average, when they get the packet. It'll be delayed .2 seconds (or whatever the roundtrip time for the packet will be, but I don't think they'll be that be that bad.
If 20 people are tapping all over the place it still shouldn't affect it. All it is is one extra packet per person, or as much as if there were one more person. And that's only when they do their little tappy trick, which I doubt 20 people will do it the same instance, and even if they do, at most you get the amount of packets sent as if there were one extra client.
Yeah you'll get the "lag" effect you can see in hyperspace, at least until it gets fixed in the client.
Mr Ekted - Mon Jun 21, 2004 4:45 pm
Post subject:
You can't ever solve the "lag effect" caused by accel/decel. Either the client predicts position, which makes normal movement nice and smooth, or it only draws when it gets updates, which makes the movement extremely jerky but you can start/stop arbitrarily.
Bak - Tue Jun 22, 2004 9:50 am
Post subject:
can't the client predict position unless it gets an update, in which case it goes the the update and then uses that for it's next prediction(s) ?
Smong - Thu Jun 24, 2004 2:34 pm
Post subject:
The position in the update is compared to the current predicted position and if it is within set limits the velocities are fiddled so the motion stays fluid (this is how people eat bombs).
Catid changed the time stamp in the packet so the client is supposed to use the values in the packet and not interpolate using them.