No on first thought because I already have a dashboard on the screen.
On second thought I might sympathize with a small LCD screen that displays some car stats (e.g. speed, whatever I choose). Though I wouldn't spend much money on it because as said, the info is already available. Also I wouldn't know where to mount it because it'd obstruct my screen if mounted on top of the MOMO's body.
He also could have said "Look into your LFS folder, under /data/language/English.txt and look for lines ..." - everyone of us has those strings in his LFS folder.
I think it's good to know that there are at least intentions to tackle some of the more important problems with LFS within the next few years after they weren't fixed for years, so I'm thankfull for this thread.
That's cool.
I think I'll wait a bit with going crazy until there are a few complete new rallies. If I learn them one at a time I'll know them within 3 hours after their release. So I better wait until there are more stages than I can memorize.
Great to see that making stages finally works. Can't wait for more.
[Edit] Does anyone know the artist of the onboard video's background music?
I strongly disagree. There is no other justification for motorsports other than technology.
If it wasn't for the technological point of motorsports I'd immediately support closing down all race tracks.
(That's why I'd support the end of NASCAR.)
I think it's not about being a flat-6, it's about being rear engined.
Real rear engined cars have very interesting exhausts. LFS cars don't. The sound of a 997 GT3 RS is a very far away - just looking at the exhaust tells you that it will take a while until we can simulate that properly in realtime.
In other words: Scawen knows the engine sound isn't perfect. It will propably get some more attention when the engine simulation is improved.
@duke_toaster:
Actually that's the same reason why I don't like things like TC in F1.
F1 is top-technology racing. And the technology devised for F1 should be applicable to normal cars.
TC helps to control the car - the engineers should think of different ways to stabilize the car. They should be thinking about suspension-improvements instead of building a virtually uncontrolable car and then fitting it with electronic gadgets until it can lap consistantly.
The F1's TC-development has no relevancy to electronic safety-systems used in road cars, thus I'd like it to disapear.
On a similar note I think F1 should run turbocharged engines for a while again. It's an important development-area that can help to improve the efficiency of road cars.
@Jakg: There're downsides to everything in IT. There's nothing easier than stating the shortcomings of a system.
@al heeley: Could you perhaps put up a picture of the system running, please?
I've been thinking about getting a 22" widescreen, but I always wondered wether it'd be worth it coming from a 17" TFT that does it's job well. What is your experience with other software? Not all programs appreciate the non-4:3 format.
That is exactly what we're talking about when saying the LFS clutch pack doesn't work well.
You go through a corner, about 0g longitudinal acceleration, the car tries to oversteer away. In LFS it's a good strategy to apply 30-40% throttle now, because as soon as you do that the rear wheels find grip and the car turns properly.
That is a typical LFS driving technique. What you do there is you force the diff into locking under power, which encourages understeer.
The same thing happens in slower corners while lifting from the brakes. Under coast the diff is locked, you approach the apex and as soon as you leave the brakes the car tries to oversteer, but as soon as you apply throttle the oversteer is gone, because the diff is under power-locking.
It is a very tedious behaviour and promotes all kinds of weird driving techniques. The basic thought when driving a clutch pack diff in LFS always is "Never, ever, operate the diff between coast and power or you'll oversteer."
To my understanding that is the case when preload is either non-existant or very very low.
I tried to circumvent this so very often by trying out everything with the setups and I can't avoid that problem with the clutch pack diff.
Not a challenge, but a nuissance.
See, when I'm taking a break from solving DEs I don't want to talk about DEs. It kind of defeats the purpose of taking a break from DEs.
Go away!
I've spent all day solving DEs because I write an exam on them next monday, so just go away with it, will ya? I'm trying to avoid maths here right now!
A picture of your thumb and whatever you managed to draw on it.
That doesn't need artistic skills, requires only a mobil-phone-cam and is something creative.
1. Each AI saves *all* it's data in it's .knw files. That includes all knowledge about all cars and all knowledge about passing, etc.
2. The AI saves it's data when crossing the finish line.
(On longer tracks you can notice that when the field crosses the finish line. The framerate gets jumpy because the knw files are written.)
3. The AI always learns. In practice, singleplayer, multiplayer, quali, race, always.
You could use that technique to kill the slowest AIs after a certain amount of time and replace him with a clone of the best AIs.
After a couple of races you do that again, and again, and there you go, AI evolution in LFS.
I'm thinking about something like 5 10 minutes races with reversed order each restart (use a LAN-server to do that), automatically scoring the AIs using InSim and automatically swapping files. That way you breed AIs that can race quickly and pass nicely.
Do you still have the AI1-11 knowledge-files from the old AIs? I think they should still be there, so your driving-heroes aren't lost. I think the folder is called data/knw or something along those lines.
I believe it has to do with the checkpoints for speed-measuring.
Imagin a corner where right after the apex the AI checks how quick it was in the last couple of corners. Each time the AI gets a too high corner-entry speed it scores a better time than it would when driving sensibly.
It writes down "I was quicker when instead of braking I just powered into the 90° bend!" and repeats that in the next lap because it looks like a good idea to go faster. Of course it is very unhappy that in the next corner it's times went down a lot from the crashing, but there is no mechanism to connect the two experiences.
Basically, there are some combination the AI just fails at. You can make the AI a set to avoid such things, but it'll never work well.
I know you said something about improved camera handling and I know about cinematographic trends these days, but is it supposed to be like this?
That happens with an Intel 915GM chip.
Live For SM works great though. Except for the fact that positional audio is influenced by the absolute position of the car relative to (0,0).
Well done so far.
Though for some reason I expected LFSM to have skidmarks. Silly me...
@Passing:
Set the AIs up with your 11 driving heros and one bad-setup-FXR on Fe Club and let them do some passing-practice.
They'll learn that "My position = his position => bad" and that "My position = his position + 0.5*FXRwidth + 0.5*FOXwidth = good". Fiddle with the FXR's gearbox to make it slow so it gets passed a lot. Use the FXR or any other wide car as the obstacle. If they practice passing the UF1 they'll be caught by surprise when they notice that other cars aren't as slim. Use a small track to practice passing so there's lots of lapping going on.
The AI drives a very strange line on some tracks. They need a setup that suits that line. My attempt at LX4/BL1 utterly failed because with my set they always oversteer out of the chicane, hitting the curbs and losing several seconds on the straight.
Also:
- The AIs always use the same line, even for reversed configs. So those won't ever work.
- The AIs aren't aware of downforce-issues and drafting. So don't ever expect the AIs to battle for positions in DF-cars.
- The AIs don't know about tyre heat. Try using cars that are easy on the tyres for best results. -> RB4
- Practicing passing on AIs is great for your driving style. You learn to keep your distance and to stay behind the AI-obstacle for several corners without touching it. It generally teaches you good manners. :go:
@al heeley: Each lap the AIs learn "This was too quick!" or "I can still go quicker there!". So after some laps with well heated tyres they find the limit and stay there. After a restart they try to go just as quick as with heated tyres because they don't realize that the good speeds were just down to tyre heat, which obviously causes havoc.
I don't think it's contradictory at all. It just implies that LFS is no serious racing simulator, which is true.
Right now LFS compares very well to CounterStrike. Oh no, CS is more realistic, it doesn't allow you to respawn until next round.
This is not a 'We want better analogue dials because they are the superior measuring system'-discussion, so the point 'digital speedos are more precise' is like saying everyone should use the BF1, because if you want to race it's the quickest car.
I am fully aware that a digital speedo is more precise. But I don't want a digital speedo. I want a properly readable analogue speedo.
And in other languages it's difficult to distinct some gerunds from verbs, because they are both not capitalized.
Names tend to be a lot more unique than gerunds, which are often identical with some forms of the verb. So I think capitalizing all nouns is the better trade-off.
However, such comparisons of structures don't work, because the use of the language is formed around the structure. So if you think a foreign language has a certain drawback you're just not using it right.
I always chuckle when some one says "Oh no, the Ring wouldn't be fun, it's actually a bad track. No one would drive it."
Why? Because people drive on that track in real life. And they spend money for it. A lot of it. Just to drive there. Not to race, just to drive. 'Hotlapping', if you will.
Oh, and saying that racing is impossible is also a great statement. After all the Ring hosts a 24h race. 'Overtaking is impossible!' ... Right....