UmbralRaptor changed the topic of #principia to: READ THE FAQ: http://goo.gl/gMZF9H; The current version is Desargues. We currently target 1.3.1, and 1.4.x. <scott_manley> anyone that doubts the wisdom of retrograde bop needs to get the hell out | https://xkcd.com/323/
egg|zz|egg has quit [Ping timeout: 202 seconds]
egg has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
_whitelogger has joined #principia
_whitelogger has joined #principia
Qboid_ has joined #principia
Qboid has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
egg|zz|egg has quit [Remote host closed the connection]
egg|zz|egg has joined #principia
egg has joined #principia
_whitelogger has joined #principia
Jesin has quit [Quit: Leaving]
Jesin has joined #principia
awang__ has quit [Ping timeout: 183 seconds]
awang_ has joined #principia
<lamont>
egg: i’ve been on sierra for some time now (and laptop is on high sierra)
<egg>
lamont: yay
<lamont>
i only needed el cap back when we couldn’t compile principia on sierra
<egg>
lamont: well, now it's faster!
<lamont>
noice
<egg>
(at least if you have multiple cores)
<lamont>
primer vector guidance is getting close to an alpha
<Qboid>
5d 0h 0m 0s left to event #28: Horizons 3e & Azerspace 2 (Intelsat 38) / Ariane 5 VA245 [at 2018-09-17 22:00:00]. Say '!kountdown 28' for details
<egg>
lamont: NICE!
<egg>
lamont: so then I guess we get to reimplement it in C++? or what's the plan on getting it to work for Principia
<lamont>
hah
<lamont>
i think i need to drink for a week straight then think about n-body physics…
<egg>
haha
<egg>
lamont: I'm fine with implementing your guidance algorithm if that simplifies things, I think the integration pretty much has to be on our side for it to work
<egg>
lamont: you're solving the boundary value problem by shooting, right?
<lamont>
eyah
<lamont>
yeah
<lamont>
multiple shooting
<lamont>
i’m probably on a fbi list somewhere for googling “multiple shooting methods”
<egg>
lamont: yeah, so we can do the integration if you specify the algorithm, and you can call us to shoot
<lamont>
i still need to do some work on the accuracy of the node executor
<lamont>
i might find i still have some conceptual math bugs there
<egg>
eggsecutor
<lamont>
but i think i fundamentally just need two things. one of them is ANY position and velocity vector at a given time on the target orbit
<lamont>
then i need to be able to integrate n-body trajectories
<egg>
lamont: well, n-body trajectories with thrust I suppose
<lamont>
oooh yeah
<egg>
lamont: but anyway we have long-term plans of doing basically Sarbian's fuel flow simulation but in Principia and more accurate :D
<lamont>
i need to propagate the costate even during coasts
<egg>
costate?
<lamont>
can you back port your fixes to mechjeb because i need to overhaul the fuel flow sim as well
<lamont>
primer vector and primer rate variables == costate — the pointing direction and the rate of change in the pointing direction
<egg>
well, if you give us an ODE we can integgrate it
<egg>
be careful to specify everything including reference frame
<lamont>
yeah that may be the start
bofh has joined #principia
<bofh>
why am I not in here,
<lamont>
yeah i finally figured out the Planetarium.rotation so i’m doing math in mostly proper ECI space
<egg>
not sure I can sanely backport to mechjeb (it's a rather different environment than principia, what I'll be doing is expressing KSP as an ODE)
<lamont>
although for some reason i’m doing it in swizzle-space still
<egg>
yeah you'll have to rigourously define your frames because there are a lot of swizzly spaces,
<lamont>
yeah are you ultimately gunning towards having ion engine thrust curves on “ODE rails”?
<egg>
maybe? it's just that for now I'd like the flight plan to be smarter than "the ship is entirely fuel and we burn to singularity"
<egg>
in particular I want to support SRBs with thrust curves
<egg>
but those have m' expressed as a function of m/m0 because wtf
<egg>
which gives you a free ODE to get acceleration as a function of time
<egg>
which means I get that complexity already, which means that in a later iteration I might as well add fuel
<egg>
(my first iteration will still be "liquid engines obey the rocket equation and everything is fuel", but combined with actually simulating solids)
<lamont>
yeah i was more thinking though that for ion engine burns you could solve the primer vector equation TPBVP and then run that as the source of actual truth of how the rocket moves during high warp
<lamont>
but that means baking some kind of transfer planner directly into principia
<egg>
yeah using that for ions on rails sounds fun in the future
<egg>
the distant future
<lamont>
the year 2000
<egg>
eggsactly
<egg>
well, baking planners into principia is probably where it will go eventually :-p
<kmath>
YouTube - The Humans Are Dead - Full version
<egg>
and then we'll drop the KSP plugin part, and we will all play Principia
<egg>
the way it's meant to be played: Principia on the command line
<lamont>
yeah so to solve the TPBVP you’d need to also have numerical jacobian code in principia along with a quasi-newtonian root solver like powell’s dogleg or levenburg-marquardt
<lamont>
and so far i don’t know enough to recommend there
<lamont>
but i’m getting fairly frustrated with alglib.net’s implementations and like MATLABs fsolve() better
<egg>
lamont: well, I don't think we want to solve the boundary value problem ourselves, you can just call us to shoot and do the solving on your side?
<egg>
integration has to be entirely in principia, but if you can call it, I think you can use that to solve the BVP, right?
<lamont>
yeah
Jesin has quit [Quit: Leaving]
<egg>
lamont: unrelated, not sure whether I showed you that, but there are histories in atmospheres now https://imgur.com/a/78Qiwll (since Δημόκριτος)
egg changed the topic of #principia to: READ THE FAQ: http://goo.gl/gMZF9H; The current version is Descartes. We currently target 1.3.1, and 1.4.x. <scott_manley> anyone that doubts the wisdom of retrograde bop needs to get the hell out | https://xkcd.com/323/
<lamont>
hah cool
<lamont>
need to play with principa, haven’t in ages
<lamont>
really haven’t properly played kerbal in ages