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.
<B787_300>
God damn i HATE private data, especially in a NIST document. "Densities were measured with a constant-volume piezometer that operates at conditions up to 745 K and 60 MPa. This study was carried out under a contract to Prof. Ilmudin M. Abdulagatov (Russian Academy of Sciences, Makhachkala) who set up a collaborative project with Prof. Nazim D. Azizov (Azerbaijan State Oil Academy (ASOA), Baku). The uncertainty estimated by ASOA for
<B787_300>
the density measurements is 0.5 kg·m-3 (for T<623 K) and 0.1 % (for T>623 K). Those experimental data were privately communicated to NIST and were used, in addition to NIST measurements, to develop the models discussed in Section 2. "
<B787_300>
ugh and even DTIC doesnt have what i want
<bofh>
yuck
<B787_300>
bofh: and DTIC is the place i turn to for obscure rocketry stuff
<UmbralRaptop>
stabbity.
<B787_300>
like i just want a nice equation for the density of RP1 as a function of pressure and temperature
* UmbralRaptop
feels like more fields need equivalents to NASA's PDS, and astronomy's VizieR.
<bofh>
YES. Yes they really do.
<bofh>
(there is still no website for central exchange of NMR spectra, a problem the crystallographers solved almost 2 decades ago)
<egg|zzz|egg>
e_14159: bofh: genetic algorithm sausagemaking intensifies: currently building up an initial population with smaller runs of a genetic algorithm to see if that helps with things getting stuck on a bad transit
<e_14159>
egg|zzz|egg: Is that problem Lipschitz-continuous (is that the correct word for stetig?)?
<egg|zzz|egg>
yeah that's the correct word
<bofh>
I somehow get the feeling it's not, since if it is, why aren't you doing conjugate gradient or something like that? (or, fuck, an LP).
<egg|zzz|egg>
and uh i guess? I had initially screwed up the metric so that it was not continuous but that should be fixed
<egg|zzz|egg>
bofh: e_14159: I'm confused
<egg|zzz|egg>
a lot of this is probably smooth, right?
<egg|zzz|egg>
but it's also probably full of local minima
<e_14159>
egg|zzz|egg: Well, I'm more concerned about smoothness, basically. If there are significant, sudden peaks in the metric there might be issues.
<egg|zzz|egg>
e_14159: yeah I think I had this issue earlier, but now it should be better. It affected neither the genetic algorithm (which converges fairly well then stalls at maybe 1 h max δt, 5 min average, χ^2 tens of thousands, nor the мcmc (which doesn't converge)
<egg|zzz|egg>
e_14159: maybe I should use the genetic algorithm to generate an initial state for the мcmc :-p
<egg|zzz|egg>
maybe if i mix everything together it will magically work
<kmath>
<whitequark> there's C++ code and then there's C++ code in Principia https://t.co/7yzqs9v8uL
<iximeow>
"coordinates are a separate thing on top of the objects you are manipulating"
* iximeow
considers, unsure if following
<bofh>
"Great old one" INDEED, MADNESS
<egg|zzz|egg>
iximeow: a vector in 3d space is not 3 reals
<egg|zzz|egg>
it's a vector
<egg|zzz|egg>
if you pick a basis, you can use that basis to express it as 3 reals
<egg|zzz|egg>
if you pick another basis, the same vector is 3 other reals
<egg|zzz|egg>
if you identify the vector with the 3 reals in one place and with 3 other reals in another, Fun happens
<iximeow>
ohhh. yes.
<egg|zzz|egg>
Principia has a lot of safeguards in place to make that not compile (the reference frame is part of the type, so is the grade of the multivector, the dimensionality, etc.)
<egg|zzz|egg>
side effects include: there is no such thing as a cross product
<UmbralRaptop>
So no way to wedge in electromagnetism?
<iximeow>
er.. why not?
<egg|zzz|egg>
iximeow: because the result of the cross product of two vectors isn't actually a vector
<egg|zzz|egg>
it doesn't behave like a vector if you go through the looking glass
<egg|zzz|egg>
it's a different object (a bivector) which is also an element of a 3d vector space, but not the original space
<egg|zzz|egg>
you can cross bivectors and get a bivector (that's also the commutator of skew-symmetric matrices)
<egg|zzz|egg>
you can make a bivector act on a vector and get a vector
UmbralRaptor has joined #kspacademia
<egg|zzz|egg>
those are all the cross product in coordinates, but they change the nature of the object
UmbralRaptor has quit [Client Quit]
UmbralRaptor has joined #kspacademia
<egg|zzz|egg>
UmbralRaptop: you can do all the eleggtromagnetism you want
UmbralRaptop has quit [Read error: Connection reset by peer]
<egg|zzz|egg>
but you have to say whether you want a wedge (vector cross vector), a commutator (bivector cross bivector) or the left or right action of a bivector on a vector
* UmbralRaptor
apparently needs a bunch more math to properly do cross products.
UmbralRaptor has quit [Read error: Connection reset by peer]
UmbralRaptop has joined #kspacademia
<egg|zzz|egg>
!wpn whitequark, коте, the котяchromem kitten, and the grey kitten
* Qboid
gives whitequark, коте, the котяchromem kitten, and the grey kitten a nickel inexplicable popup that didn't used to be there and doesn't do what it used to
<e_14159>
So, update on my master's student: With the deadline for handing the thesis in being on the fifteenth (two weeks from now Friday), he apparently planned to send me the last chapter for me to give feedback end of next week.
<egg|zzz|egg>
um
<egg|zzz|egg>
how big are the chapters
<e_14159>
And the latest chapter I received consisted of 1/3rd stuff that belongs in the method section, 1/3rd of irrelevant implementation detail, and 1/3rd was a (literal) itemization of hyperparameters.
<e_14159>
Probably five to ten pages?
<e_14159>
Oh, and the figures had drawing grids in the background.
<e_14159>
(I offered to read through each chapter once as it arrived, then the whole thesis once. For each, I said I could guarantee feedback after a week)
tawny- has quit [Read error: -0x7880: SSL - The peer notified us that the connection is going to be closed]
<egg|zzz|egg>
!wpn e_14159
* Qboid
gives e_14159 a compound pilcrow
<e_14159>
Thanks; I need that :-)
tawny- has joined #kspacademia
<UmbralRaptop>
Pilcrow: a polearm notable for being able to knock down walls of text.
UmbralRaptop has quit [Read error: Connection reset by peer]
* UmbralRaptop
is now armed with the sword of the earth!
<kmath>
<LWoodney> The face of a teenager who pushed out a window screen to sneak out to party with her friends. No remorse. But today… https://t.co/TuDEscMfqn
<kmath>
<sondy> EVERYONE PLEASE STOP ASKING ME TO TAKE IN CATS *sigh* everyone is forwarding me requests to take ownership of cats… https://t.co/cWIuw4QGDY
<kmath>
<bofh453> LRT: doubly amusing since in complex analysis land, exp(z) is simple to define, it's just \sum_{n=1}^{∞} \frac{z^n}… https://t.co/SS9I5ZAS0L
<kmath>
<bofh453> So our solution is "simple": pick a ray in the complex plane over which ln(iθ) is *not defined*, i.e. discontinuous… https://t.co/prG4Xa3Hma
<egg|zzz|egg>
bofh: I mean you could extend it on the boundary of the region where it's holomorphic by picking a side :-p
<egg|zzz|egg>
bofh: and if you don't, it's continuous :-p
<egg|zzz|egg>
bofh: are lie algebras bold fraktur or fraktur
<egg|zzz|egg>
!pet bofh
* Qboid
pets bofh
egg|zzz|egg_ has joined #kspacademia
egg|zzz|egg_ is now known as egg|laptop|egg
egg|laptop|egg is now known as egg|z|egg
<bofh>
egg|zzz|egg: bold fraktur IME
<egg|zzz|egg>
e_14159: do you think it might make sense to produce close candidates for ensemble мcmc via our weirdly-quickly-converging genetic algorithm
<bofh>
egg|zzz|egg: and I'll freely admit I played a bit fast and loose w/some concepts, it's a twitter thread aimed somewhat generally, not a lecture :P
<bofh>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a harmfully-considered point
regex has joined #kspacademia
regex has quit [Remote host closed the connection]
<egg|zzz|egg>
regex!?
<egg|zzz|egg>
very transient regex :-/
<egg|zzz|egg>
argh both great old ones so far have converged to the same bad transit
<kmath>
<geoffwozniak> Kernighan says this (insane!) example comes from "what I believe is the worst C programming textbook ever written."… https://t.co/olVGYGMn4H