LFS doesn't need to store the key frames in the replay -- it just needs to cache them as it reads the replay. So if you view a replay, arrive at lap 8 and decide you want to go back to lap 5. LFS will have kept a copy of the full state at the start of that lap, in memory or in a temporary file. Going back to lap 5 can be done instantly. No need to change the file format.
This brings you back-skipping. For forward-skipping, LFS could use the technique described in post #3. True rewinding (= reversing the direction of time) is much tougher, because the physics can only be calculated forward. You'd need to cache more frames -- 1 per second or so -- to get a smooth backward motion. Or you'd have to drop the physics-recalculating stuff and adopt a purely frame-based replay format.
If LFS does not discard the key frames but stores them in a permanent file, then it can re-use them next time you view the replay, so skipping forward will be faster.
Dogs have ethics too, if you define ethics as "group norms that tell you how to behave towards fellow beings". A dog is a social animal and must learn the rules of the pack when it grows up. And apes have been observed to be nice to animals of other species, such as a wounded bird.
The US economy is failing because of Iraq, not despite. Some Yanks make a good profit out of Iraq, but the nation as a whole doesn't.
It's comforting to believe there is sharp distinction between the morally good people (which includes oneself) and the bad guys, and that the good folks will never, under any circumstances, cross that line. Then you can safely push cases like these aside. Call the soldiers psychopaths, say they should be put behind bars, and carry on.
But there is much evidence to the contrary. Examples: Milgram's experiment, the Stanford prison experiment, and the complete societies of Germany, Rwanda and Cambodia at some point in history.
@ Anyone raising his voice in anger about those poor dogs: Fine. Great. Have you asked yourself how the meat in this evening's meal was treated when it was a live animal?
(And yes, I think it's sick what they did in the videos. I'm sure those soldiers were nice, decent guys back at home. That's what a war does to people. It's sick, but not half as sick as Abu Ghraib, Guantanamo Bay, or My Lai, to name a few.)
It's not so much the politicians I have problems with. It's the mindless civilians who vote for them and go on bleating "support our troops". If they really care about those boys they should get them out of the war asap.
In LRA I calculate the G forces by differentiating the position: this results in the center of the turn circle, and the lateral and longitudinal forces on the CoG. These need to be corrected for the slip angle of the car's body, using the forward vector and right vector.
To calculate the center of the turn circle at time T, LRA takes the position at <T - 0.05s> and <T + 0.05s> (formula). This also "averages out" some noise.
Your approach (calculating the G forces from the tyre forces) had crossed my mind, but I didn't follow that route, because:
1. It will give wrong results for the longitudinal force, because you need to subtract the force from air resistance -- which isn't in the RAF data. (At top speed the force on the wheels is pretty high, but acceleration is zero.) For the lateral force you should get a similar error, but smaller, because lateral drag will be very low, if LFS models this at all.
2. The tyre forces have big spikes when the wheel touches a kerb or bounces on the road surface. I expect the G forces derived from that will also be "spikey."
(I never tried it, though.)
If you want a good test for your calculations, use a RAF file from a lap at the Kyoto Oval. The banked turns there are an extra complication.
Sales rep: Our new model has a hi-res camera, 50 Gig RAM, MP3, text messaging, LFS pre-installed, a built-in microwave, aaannd... Lessee... Oh yes, you can even make the car go round the bend.
1 float 0 suspension deflect : compression from unloaded 1 float 4 steer : including Ackermann and toe 1 float 12 X force : force right 1 float 16 Y force : force forward 1 float 8 vertical load : perpendicular to surface
(Why are the fields out of sequence?)
LRA, my replay analyser, reads the data in the order
0 suspension deflect 4 steer 8 lateral force 12 longitudinal force 16 tyre load
So the lateral force should be at offset 8. At offset 12 you'll find the longitudinal force.
If you load the file into F1PerfView or LRA you can check for yourself. Or you can post your RAF file here.
When I wanted to create an account at RSC, my first attempts failed because the name already existed. I was in a hurry, so I used the name of the account I had back at the University.
It was at the dept. of maths (wiskunde) and computer science (informatica); the last two characters are my RL initials -- the admins were Un*x guys and had a fixed way of naming accounts. Not very readable, but it does give you a unique ID.
When I moved to LFS I kept the name, because it had become a net identity. Guess I'm stuck to it now.
Unfortunately, this is not true for all cars. For FWDs you get the rear wheels first (RL, RR, FL, FR). AWD cars are like RWDs.
The correct way to determine it is to read the X and Y position of each wheel from the static wheel info block. If X < 0 then it's a left wheel, if Y < 0 then it's a rear wheel.
So you expect the devs to publish the source code and all other resource files? So they should make the results of many years of work, and the source of their income, available for anyone (including competitors)?
And then you think that some smart guys in the community -- experts in car physics, games programming and 3D modeling -- will step forward, take a good look at the stuff, and solve the problems? C'mon...
Scawen, Eric and Victor are very good at what they do. If these problems were simple to solve, they'd have done so a long time ago.