egg|nomz|egg changed the topic of #kspacademia to: https://gist.github.com/pdn4kd/164b9b85435d87afbec0c3a7e69d3e6d | Dogs are cats. Spiders are cat interferometers. | Космизм сегодня! | Document well, for tomorrow you may get mauled by a ネコバス. | <UmbralRaptor> egg|nomz|egg: generally if your eyes are dewing over, that's not the weather. | <ferram4> I shall beat my problems to death with an engineer.
<kmath>
<AdamSerwer> Tonight is the first night of Hanukkah, where we celebrate the defeat of a vast modern empire at the hands of a sma… https://t.co/Y9yGnia2pd
egg|zzz|egg has quit [Read error: Connection reset by peer]
icefire has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
egg|mobile|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 186 seconds]
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
<egg|work|egg>
!wpn
* Qboid
gives egg|work|egg a numerical pounce
<bofh>
!wpn egg|work|egg
* Qboid
gives egg|work|egg an associative trihexaflexagon/IDE hybrid
awang has quit [Ping timeout: 198 seconds]
awang has joined #kspacademia
awang has quit [Ping timeout: 207 seconds]
<UmbralRaptor>
bofh: yeah, my QM teacher is silly.
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 186 seconds]
APlayer has joined #kspacademia
<APlayer>
Hi there! Sorry for crashing right in like this, but this might be the first time I need outside help with school maths... I have this function, a part of which is e^(-2 * (x^2)). How do I differentiate this? WolframAlpha spits out -4 * e^(-2 * (x^2)) * x, but I have no idea how this works
<Fiora>
chain rule
<APlayer>
That is, I believe I have trouble splitting this up for the chain rule
<Fiora>
iirc
<APlayer>
Yes, chain rule
<Fiora>
F'(x) = f'(g(x)) * g'(x)
<Fiora>
g(x) = -2*x^2
<APlayer>
But what are the two individual functions that are chained? e^x and -2 * x^2?
<egg>
awang: have you had the time to look at that branch?
<egg>
awang: the new moon is on the 18th, so it would be nice if you could confirm that it fixes the bug (and doesn't break other things) by the week-end
<awang>
egg: Haven't had much time, unfortunately
<awang>
I'll definitely be able to test it more thoroughly by the 18th
<awang>
The one time I got to try it I got an unrelated KSP crash
<awang>
Helps that my last final is today
<awang>
egg: Also, would time passing in the editor violate any assumptions Principia makes?
<egg>
time doesn't pass in the editor iirc?
<egg>
awang: well we'll build the release this week-end, so it's not really "by the 18th"
<awang>
It does if you have EditorTime installed
<awang>
I'll be able to test some more by then at least
<egg>
huh, yeah time passing in the editor might break things
<awang>
I wonder
<awang>
Might it be that if I stay in the editor longer than the minimum history time problems would arise?
<awang>
Wasn't a problem earlier because I used the default history length
<egg>
awang: I mean, we think there's an actual bug in the C# there, I just want to make sure it does fix the problem and doesn't introduce new ones
awang has quit [Ping timeout: 186 seconds]
* egg
pokes awang in the timeout
egg is now known as egg|nomz|egg
<SnoopJeDi>
lewd
<SnoopJeDi>
egg|nomz|egg, I shoehorned a discussion of cosine similarity into a dinner date last night because I had Spotify running :D
<kmath>
<adambroach> @mozilla's privacy engineers ran a study to see whether various privacy protection techniques break the web for use… https://t.co/TL8iOrx7Ow
<egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor a metal tarrasque
<UmbralRaptor>
*a* tarrasque?
<UmbralRaptor>
!wpn egg
* Qboid
gives egg a maximum variety
<egg>
UmbralRaptor: ????? ?☿♄♃♂♀☽☉ (there doesn't appear to be a "turned ?")
* UmbralRaptor
gives egg a codex of alchemical engineering
<kmath>
<ChemHeritage> This #NEH grant will fund the prototype of a digital game that builds on CHF’s expertise in the history of alchemy,… https://t.co/XVZ2d65UWM
<SnoopJeDi>
I bounced off of Spacechem pretty hard and TIS-100 intimidates me
<SnoopJeDi>
LOL the manipulator in the middle
<SnoopJeDi>
actually no, just the entire thing
<egg>
SnoopJeDi: which one?
* UmbralRaptor
likes that back and forth bounce,
<egg>
UmbralRaptor: yeah, also the fact that I'm binding a giant blob of water then slicing it up and rebinding it >_>
<SnoopJeDi>
the one that swings RTL egg, I love the advancing blob and also the radial bounce
<egg>
UmbralRaptor: I think there's a dead cycle because of the bounce though, I wonder whether I can get the water to be fet optimally
<egg>
s/fet/fed
<Qboid>
egg meant to say: UmbralRaptor: I think there's a dead cycle because of the bounce though, I wonder whether I can get the water to be fed optimally
<SnoopJeDi>
it just occurs to me to ask this, but...does the game render these gifs itself?
<egg>
no, I mean x and I expand at x at the appropriate place
<UmbralRaptor>
Ah, so that's how a SnoopJeDi uses -forcepush.
<SnoopJeDi>
!wpn -add:wpn ≈
<Qboid>
SnoopJeDi: Weapon added!
<egg>
SnoopJeDi: so you get pi-π = ~half an ULP at π, *and* that value ends up in the result because the derivative of the sting of death is 1 in absolute value
<egg>
the second part matters, a function with a smol derivative will eat the error in its argument
<egg>
SnoopJeDi: and when you look at relative errors on either side, that's how you get the condition number
<egg>
anything that has a 0 far from 0 with non-0 derivative is terribly-conditioned there, because a small relative error in the input (here the error in π, sub-ULP) results in an arbitrarily large relative error in the output (10^-16 rather than 0)
<egg>
note however that the 0 at 0 is perfectly well-behaved
<egg>
it doesn't amplify errors (there sin x = x so it literally does nothing)
<egg>
SnoopJeDi: if you search your distant backlog you'll find me asking bofh about rewriting Keplerian-element-related things because of such condition considerations
<SnoopJeDi>
vaguely rings a bell yep
<egg>
SnoopJeDi: e.g. if you want to compute x^2-y^2 (x close to y, nonzero reals), and you compute it that way, the errors in the rounding of the squares will be arbitrarily amplified by the difference (even though computing the difference is exact by Sterbenz). If you compute it as (x-y)(x+y) everything's fine and you just get a couple of ULPs
<soundnfury>
egg: ahh, because (x-y) is ill-conditioned and you're doing that step first?
<egg>
you can see this amplification by looking at the condition of u-v, e.g. in u, u/(u-v) which becomes infinite for nonzero v and u -> v
<egg>
soundnfury: yeah, since you do it first there's no previous rounding error to amplify (and by Sterbenz it's even exact if x and y are within a factor of 2 so you don't get any error from that)