raptop 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. | We can haz pdf
<bofh>
LOL ON PAGE 150 THERE IS A DIAGRAM OF THE GROUND STATION AND ONE OF THE ITEMS ON IT IS "FIRE EXTINGUISHER", THAT'S WHY
<B787_300>
lol
<bofh>
also it's not that odd of a grouping of terms, like it literally "happens" in Gravity
<bofh>
Sandra Bullock does a transfer from the ISS to Tiangong-1, an orbital inclination change of 8.88 degrees.
<bofh>
Hint: you need more Δv than you can get from a fucking fire extinguisher to pull that off.
e_14159 has quit [Ping timeout: 190 seconds]
e_14159 has joined #kspacademia
<whitequark>
how much do you get from a fire extinguisher
<bofh>
okay, so let's take a CO2 fire extinguisher. standard ones seem to contain 2kg of CO₂ pressurized to 70 bars. I'm being lazy and lifting figures from a KSP thread here, but basically 3 * P * V gives you an internal energy of ~116kJ. Assuming that all of this gets converted to kinetic energy we get 116kJ = 1/2m*v^2, let's assume a mass of 126.5kg (70kg for the human, 50kg for the spacesuit, 6.5kg for
<bofh>
the extinguisher less propellant) -> v ~ 21.4126m/s
<bofh>
now correcting for gravity, we use a g of ~8.6735m/s^2 (for an orbital altitude of 408km & a mean Earth radius of 6371km) and so at this altitude get a v ~ 18.93193m/s for our exhaust velocity
<bofh>
now plugging it into the rocket equation we get v * ln(m0/mf), m0 = 128.5, mf = 126.5 -> Delta_v ~ 0.297m/s. meters. not kilometers.
<bofh>
basically for a system where your propellant is only ~1.556% of your system mass, you're going to get tiny Delta_v pretty much no matter what
<UmbralRaptor>
egg|zzz|egg: sorry, was dead. Should be somewhat less dead now.
<egg|cell|egg>
Meow
<egg|cell|egg>
Zzz,
<UmbralRaptor>
bofh: uncertain about switching the g, but yeah. I'd recommend replacing the fire extinguisher with a similarly shaped hydrazine thruster
egg|laptop|egg has joined #kspacademia
<egg|laptop|egg>
bofh: what's an interesting function to test finite difference formulae on
<bofh>
egg|laptop|egg: x^3*sin(1/x), particularly near 0 :p
<egg|laptop|egg>
bofh: can you think of something interesting that involves only +-*/ sqrt cbrt
<egg|laptop|egg>
so that it's reproducible in the current state of the principia libm :-p
<bofh>
egg|laptop|egg: so like all the "interesting"/evil examples I can think of use sin or cos, b/c the things that finite difference tends to choke on (IME) are highly periodic/oscillatory functions
<egg|laptop|egg>
bofh: yeah
<egg|laptop|egg>
bofh: but that's inherent to the formula, which is not what I'm trying to debug
<egg|laptop|egg>
bofh: I'm trying to see if the implementation is too cancelly
<bofh>
egg|laptop|egg: wouldn't just a simple product/chain rule with a sign in it work there? like x^2*sqrt(1-x) or something along those lines
<egg|laptop|egg>
dunno
<bofh>
so I've literally never worked with finite difference methods tbh, so I'm not entirely sure what sort of functions would be likely to hit implementation-related pathology. but if cancellation is the main worry (which I guess yeah, that makes sense), just create something where the function and second derivative almost cancel but the first derivative doesn't
<egg|laptop|egg>
bofh: well, it's most cancelly for higher-order methods obviously
<egg|laptop|egg>
bofh: I guess I want functions with smol derivative to get the cancelly goodness
<bofh>
...in retrospeggt, yeah.
<egg|laptop|egg>
bofh: how about Cbrt(x) * (x-1) near 1
<egg|laptop|egg>
or do i want the derivative to be smol everywhere
<bofh>
I feel like this is a case of "try it and see"
<egg|laptop|egg>
s/(x)/(x-1)
<galois>
egg|laptop|egg meant to say: bofh: how about Cbrt((x-1)) * (x-1) near 1
UmbralRaptor has quit [Read error: Connection reset by peer]
UmbralRaptop has joined #kspacademia
egg|laptop|egg has quit [Ping timeout: 180 seconds]
egg|laptop|egg has joined #kspacademia
egg|laptop|egg has quit [Ping timeout: 180 seconds]
egg|cell|egg has quit [Ping timeout: 190 seconds]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 183 seconds]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 190 seconds]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 202 seconds]
egg|cell|egg has joined #kspacademia
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has joined #kspacademia
<bofh>
!u
<galois>
No info for U+f8ff (I only know about Unicode up to 9.0)
<galois>
title: Shared album - Patrick N - Google Photos
<SnoopJeDi>
bofh, I heard this on the radio this weekend and it is so bizarre but I love the creative ways the saxophone sound is used in it: https://www.youtube.com/watch?v=tJqHGNPBCQU
<galois>
title: Songs for Tony: I. â - YouTube
<bofh>
SnoopJeDi: so I have a question, why are *so* many webfonts just fucking terrible?
<SnoopJeDi>
I confess to actually liking some of that font
<SnoopJeDi>
although the TALLBOI capital C is...extremely cursed
<bofh>
(also honestly the lowercase in Amaranth is basically just Comic Sans Lite but with nonsensical advance widths)
<SnoopJeDi>
oh I guess the top of G is also not flush with the rest
<SnoopJeDi>
anyway, I guess the answer is because nobody cares about fonts and the web encourages cheap aesthetics at all layers
<egg|zzz|egg>
bofh: not sure how best to implement finite differences
<egg|zzz|egg>
Σᵢ aᵢ f(xᵢ)? Σⱼ (Σₖ aₖ) (f(xⱼ) - f (xⱼ₊₁))? make the aᵢ integers and divide by a common denominator? should the accumulation be in double precision?
<bofh>
I'd prolly try the common denominator trick at first and see if it helps (since I get the feeling it might), but not sure. If you're doing differences, I'd do accumulation in double-double or use Kahan summation.
<egg|zzz|egg>
I haven't seen an eggsample where the accumulation in double has a noticeable effeggt
<egg|zzz|egg>
bofh: is there a difference between kahan sum and accumulation in double-double?
<egg|zzz|egg>
bofh: otoh on some eggsamples (but not others) differences are a clear improvement over the naive linear combination
<egg|zzz|egg>
bofh: the denominator seems to help most in cases where the formula should be eggsact
<egg|zzz|egg>
(e.g. because it has higher order than the thing being integgrated)
<egg|zzz|egg>
s/integgr/differenti/
<galois>
egg|zzz|egg meant to say: (e.g. because it has higher order than the thing being differentiated)
<bofh>
21:18:51 <@egg|zzz|egg> bofh: is there a difference between kahan sum and accumulation in double-double?
<bofh>
I mean they *should* be identical
<bofh>
expressionwise
<egg|zzz|egg>
there might be a difference in order, kahan sum might be equivalent only if the increment is smol
<egg|cell|egg>
bofh: s/ex/eggs/
<galois>
egg|cell|egg thinks bofh meant to say: eggspressionwise
egg|laptop|egg has quit [Ping timeout: 180 seconds]