00:34 <discord-> n​eph. — Nope, Standecco's got the right idea!

07:14 <discord-> S​ir Mortimer. — @egg for these contracts to work with time warp, I need to calculate substeps when one single fixed update frame exceeds a defined threshold of simulator time. I'm currently doing that based on the stock orbit of the affected vessels and bodies (by taking the orbit at that time, and then fetching past positions from those orbits) - which will yield wrong results with principia. Maybe it *might*

07:14 <discord-> S​ir Mortimer. — But if not, I'm going to need a way to query the history of vessels and bodies.

07:18 <discord-> S​ir Mortimer. — It is going to be a problem when the time warp steps is a significant portion of the orbital period, and that orbit doesn't look anything like something that kepler would have thought of.

07:18 <discord-> S​ir Mortimer. — It is going to be a problem when the time warp step is a significant portion of the orbital period, and that orbit doesn't look anything like something that kepler would have thought of. (edited)

20:41 <discord-> e​gg. — @Sir Mortimer yeah, an API to query the history probably makes sense anyway

20:43 <discord-> e​gg. — I do wonder how bad your current approximation might get in practice

20:43 <discord-> e​gg. — how big do the time steps get?

20:44 <discord-> S​ir Mortimer. — 2000 seconds in stock, I think, and according to @DRVeyl about 6 hours in RO

21:13 <discord-> B​utcher. — 6 million times is RO max.

21:14 <discord-> B​utcher. — Which would be 33 1/3 hours I think.

21:39 <discord-> D​RVeyl. — Oh, thought it was 1M multiplier.

21:40 <discord-> D​RVeyl. — 6e6? Yikes.

21:56 <discord-> B​utcher. — I believe so.

23:06 <discord-> G​ot. — Yeah, I confirm that RO max is 33.3 hours per step. Which makes any position based discrete simulation subsystem in KSP output total random values.

23:15 <discord-> D​RVeyl. — I'd say it's fair to say that we're just doing that to ourselves. I'm not certain if or why we need a month/real-world second as a step size. Probably because historically, it didn't cost anything to set that -- there was no background processing and Keplerian orbits don't care what the size is... barring funny tricks like teleporting through encounters (or entire planets!) based on physics steps.

23:16 <discord-> D​RVeyl. — IOW, I think we'd be interested in what step sizes are reasonable for the simulation. I imagine it's actually worse for stock to have larger steps than RSS scale.

23:21 <UmbralRaptop> Hm. Io's period is 42 hours

23:21 <discord-> A​cer_Saccharum. — And Mimas' is 23 hours

23:22 <UmbralRaptop> whelp

23:22 <UmbralRaptop> What's Laythe's anyway?

23:26 <discord-> A​cer_Saccharum. — ~15 hours

23:36 <discord-> A​cer_Saccharum. — The shortest orbital periods in the solar system are 7hr 4min for Naiad (Neptune III), 7hr 5min for Metis (Jupiter XVI), and 7hr 39min for Phobos (Mars I)

23:37 <discord-> A​cer_Saccharum. — At least those are the numbers I could find in five minutes

23:39 <UmbralRaptop> Neptune: home of the shortest and longest period moons

23:44 <discord-> A​cer_Saccharum. — Saturn's innermost ring motes would orbit in around 4hr 40m

23:45 <UmbralRaptop> Simulating Saturn's rings would probably require rewriting Principia in FORTRAN

23:48 <discord-> A​cer_Saccharum. — It would be cool to have Procedural motes, but that would fall under a different mod

23:48 <discord-> A​cer_Saccharum. — Also colliding with the rings

23:50 <discord-> e​gg. — @DRVeyl I don’t think it is absurd to timewarp fast, not everyone plays career, and if you want to sandbox a single interplanetary mission, you quickly want fast warp

23:51 <discord-> e​gg. — 33.3 hours is not that much, all things considered

23:51 <discord-> e​gg. — Obviously as @Got says, looking at the vessel every 33 hours will yield garbage

23:52 <discord-> e​gg. — but it is short enough that I would expect the Keplerian approximation to remain semidecent in a bunch of cases

23:52 <discord-> e​gg. — (and anyway we should have an API to ask Principia what’s going on)

23:59 <UmbralRaptop> something something runge-kutta-nystrom