00:40 <discord-> e​gg. — @Kerbinator Again, you are making incorrect technical judgments, because you do not understand how Principia works.

00:40 <discord-> e​gg. — > discard planet history

00:40 <discord-> e​gg. — No planet history is kept. This statement is not just a bad idea, it is nonsense.

00:40 <discord-> e​gg. — > because history discard is what we need >= Uranus

00:40 <discord-> e​gg. — This is not what you need.

00:40 <discord-> e​gg. —

00:40 <discord-> e​gg. — What you need is faster scene changes.

00:40 <discord-> e​gg. — As @Standecco points out, there are multiple steps to getting there, they are complex, and they nothing to do with what you would naïvely think is needed.

00:40 <discord-> e​gg. —

00:40 <discord-> e​gg. — What is causing the slow scene changes tracked by issue #2400 is a combination of two things:

00:40 <discord-> e​gg. — 1. large saves, stemming from old low-orbit satellites whose histories are big (note: *vessel* histories are kept, not *planet* histories);

00:40 <discord-> e​gg. — 2. games where the current time is far from 1950, because the ephemeris needs to be recomputed from 1950.

00:40 <discord-> e​gg. —

00:40 <discord-> e​gg. — To resolve 1., you need smaller histories. However, smaller (in terms of data size) is not the same thing as shorter (in terms of duration). Picking different representations will lead to different behaviour; right now, the representation is very efficient for relatively straight trajectories (for instance, interplanetary trajectories), but is woefully inefficient for orbits. We are looking at an alte

00:40 <discord-> e​gg. —

00:40 <discord-> e​gg. — To resolve 2., we need to not recompute the ephemeris. Since we need to have the ephemeris over the length of the longest trajectory plotted, which may be very long for outer solar system trips, we need instead to store the computed ephemeris over a long period of time.

00:40 <discord-> e​gg. — This reduces the issue to problem 1.: right now storing the ephemeris is impossible because of the massive amount of data involved in representing the orbits of the planets, and, more critically, the moons, which would exacerbate the slow load times instead of improving them. With a representation that is efficient on orbits, this should become possible.

00:40 <discord-> e​gg. —

00:40 <discord-> e​gg. — As a result, the current work on #2400 focuses on implementing trajectory representations that are efficient for orbital trajectories.

00:43 <discord-> b​adgermasher. — I would've just gone with a lolno, but nice thorough response

02:57 <discord-> K​erbinator. — thx

02:59 <discord-> K​erbinator. — well

02:59 <discord-> K​erbinator. — I recontacted the original reporter who is Chinese and not on discord

03:00 <discord-> K​erbinator. — He said after he fly a vessel and reverted to VAB there is still noticeable lag increase