<UmbralRaptor>
egg|zzz|egg: I assume you will be stabbing Tone in the numerics at some point.
egg|phone|egg has joined #kspunofficial
<UmbralRaptor>
egg|phone|egg: I assume you will be stabbing Tone in the numerics at some point.
<egg|phone|egg>
Hmm?
<egg|phone|egg>
Tu Caen de thé stabbity
<egg|phone|egg>
Autocowreck
<egg|phone|egg>
You can do the stabbing
<UmbralRaptor>
Basically they're suggesting rescalling distances by a factor of 1e6 as a way to get around KSP damping down low speeds. So you can ingame have the equivalent of a bowling ball orbiting a person.
egg|cell|egg has joined #kspunofficial
egg|phone|egg has quit [Read error: Connection reset by peer]
<egg|zzz|egg>
UmbralRaptor: ........ but FP is scale-free
<UmbralRaptor>
egg|zzz|egg: ¯\_(ツ)_/¯
<UmbralRaptor>
I don't know enough numerics to usefully get into this.
<UmbralRaptor>
Hence my request that you do the stabbing.
<Iskierka>
I don't know what their logic is but the physics engine is somewhat scale-sensitive for various reasons. Though that's why RO doesn't work completely stable so if that's an increase scale of 1e6 ...
<egg|zzz|egg>
UmbralRaptor: nah, you know enough, and you have better claws
<egg|zzz|egg>
Iskierka: the scale sensitivity is because of relative sizes, the Earth is much bigger wrt an IRL spacecraft than Kerbin wrt to a KSP spacecraft
egg|zzz|egg is now known as egg
<Iskierka>
I have seen that physics engines can struggle with using large mass numbers in general, and that even comparably large objects often don't like each other
<Iskierka>
though possibly physX is nicer about those
egg|cell|egg has quit [Quit: Bye]
<egg>
Iskierka: hmm; there might be some scales baked into the engine? there is certainly the time scale of your physics step (insert rant about Euler integration here)
<egg>
Iskierka: not sure where a mass scale would come from though
<egg>
(obviously if you have very large masses but keep having small masses your problem is that, rather than inherent scale)
<Iskierka>
probably would come from approximations made that are sane at smallish number ranges around 1-10 but break more and more the further you go outside
<egg>
possibly, but I can't think of anything obvious that would do that
<kmath>
<iamdevloper> Today we celebrate the day when Jeff Goldblum blind deployed untested code into a completely hostile environment and it ran first time ??