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
<mofh>
!wpn egg|cell|egg
* galois
gives egg|cell|egg a parabolic velociraptor
e_14159 has quit [Ping timeout: 190 seconds]
<mofh>
21:23:41 < mlbaker> which is like, why would you say 'diffeomorphism groupoid' then 21:23:48 < mlbaker> just say moduli the action of the diffeomorphism group lol
<kmath>
<bofh453> Yes, this is a spiral branch cut. Yes, this produces a valid Complex Logarithm. No, you don't ever want to actually… https://t.co/OrmdPsRvyD
<SnoopJeDi>
mofh, who hurt your prof
<mofh>
SnoopJeDi: oh so that prof is infamous for liking somewhat tedious counterexamples in analysis
<SnoopJeDi>
It was a rhetorical question, I know who hurt that prof. Math hurt that prof :D
<SnoopJeDi>
mofh, but that scans with that problem
Technicalfool_ has quit [Quit: Leaving]
<mofh>
SnoopJeDi: his complex class was great b/c the structure imparted by complex-differentiability meant you could often via cleverness (and occasionally bootstrapping a few lectures forward, which was explicitly always okay as long as you were fully rigorous) you could sidestep the grindiness
<mofh>
SnoopJeDi: like there was one question on an earlier assignment that was computing 1/z around these silly starlike contours
<SnoopJeDi>
That's a good mind-trick
<SnoopJeDi>
well, fsvo "good"
<mofh>
SnoopJeDi: and instead of bashing arctan integrals for literally 14 pages like the solutions did, me and a friend just derived the principal branch of Log(z) and then threw Cauchy's Theorem at the problem
<mofh>
which came out to two pages and was fun as hell
<SnoopJeDi>
neat, that's where I expected that to go and I feel vindicated that I can remember any maths
<SnoopJeDi>
some days it feels like a fever dream to remember that I technically studied this
<mofh>
I mean I sometimes forget I technically have a *math* degree
<SnoopJeDi>
YUP
<mofh>
Then I get dropped into a complex analysis refresher and the lecturer asks for an eggsample of a non-analytic function (we had just covered polynomials as an example of analytic functions) and my response was "complex logarithm" (which earned a "not now, MUCH LATER" as a response :p).
<SnoopJeDi>
Presumably they were looking for things like 1/z?
<mofh>
|z| or \overline{z} are the canonical standard eggsamples
<mofh>
1/z is holomorphic on the punctured complex plane
<SnoopJeDi>
But you do have to puncture it :P
<SnoopJeDi>
Guess my head is stuck in "removable singularities" mode because the residue theorem came up :)
<mofh>
Yeah, makes sense (well, pole of order 1, but you want poles if you're dealing with the residue theorem)
<mofh>
(whether you necessarily also want Poles is a thing I have yet to empirically determine)
<SnoopJeDi>
is capital-P Poles referring to essential singularities or Polish people
<SnoopJeDi>
(definitely on my list of "sentences I would not have expected to say in my life")
<egg>
essential singularities are not Poles
<SnoopJeDi>
err, oops.
<egg>
whether the entire complex plane except for possibly one value may be found in any neighbourhood of a Pole is a different question
<SnoopJeDi>
yea I'd forgotten the distinction between removable singularities and poles, and the notion that essential singularities are The Other Stuff™
<egg>
then the answer probably depends on whether the Pole is on a Concorde
<SnoopJeDi>
Taking more analysis would have been nice
<SnoopJeDi>
Particularly to shore up my understanding of holomorphisms and co.
<egg>
amusingly, in German, both have a capital P, with Pole being both plural poles and singular Pole
<SnoopJeDi>
o.O
<_whitenotifier-5dfc>
[Principia] pleroy opened pull request #2213: Extract the nome function - https://git.io/fjVgz
<mofh>
rofl did Fukushima just haphazardly shove the nome somewhere in the elliptic f'n code?
<mofh>
(admittedly, that *is* the usual way I see it implemented)
<egg>
my complex analysis professor, who wasn't very good at German, kept mispluralizing Pol (pole) as Polen, so there were Poles in the meromorphic functions
<egg>
mofh: Nice
<mofh>
egg: nice
<egg>
mofh: omnomenome
<egg>
omenomenome
<mofh>
donot eat the nome
<_whitenotifier-5dfc>
[Principia] pleroy synchronize pull request #2213: Extract the nome function - https://git.io/fjVgz
<mofh>
(also gah, calling Mediator "amphetamine-based" is totally missing the point, that's not why it's dangerous, it's the fact that it produces norfenfluramine as a metabolite)
<UmbralRaptop>
mofh: for surprisingly small values of eggspected. Think of Python as Excel++?
<egg>
yeah that "eggspected" only holds for a very specific background
<egg>
(arguably floating-point arithmetic as a default may come with a different category of surprises, I'd eggspect it more in something that screams "numerics" like matlab)
<UmbralRaptop>
Compare with 2^32 in C
<egg>
(then again numpy is a thing)
<mofh>
(fair enough)
<mofh>
“It’s the French paradox: We have doubts about many things; we grumble. But thankfully, vaccine coverage remains high,” says Olivier Schwartz, scientific director of the Pasteur Institute in Paris, which depends on public donations to carry out its work, including vaccine research. “I don’t perceive a hostile climate,” Schwartz adds. “On the contrary, I feel that [people in France have] a
<mofh>
thirst for knowledge.”
<mofh>
honestly this *does* overall match what I see
<egg>
mofh: the *nome* with the arithmeticogeometric mean?
<egg>
or do you mean the integgrals
<egg>
the arithmeticogeometric mean does appear in BulirchCel, but that's only used in one quadrant and only after some argument reduction that makes sure it doesn't cycle too long
<mofh>
egg: right, the integgrals, I was confusing it with the nome asymptotic series that is what I use
<mofh>
(I mean I also use the Taylor series but only for values iirc < 0.25, since above that the asymptotic converges much much faster)
<egg|zzz|egg>
mofh: for the nome it's just pretty high order
<egg|zzz|egg>
mofh: apparently the fukushima method is about 4-5 times faster than the Carlson one, and noticeably faster than Bulirsch
<mofh>
Not bad, I'm definetly taking this elliptic integral impl then (and yeah, Carlson symmetric form *is* a bit slow honestly).
<egg|zzz|egg>
mofh: note, phl has yet to clean it up entirely, and there were some silly bugs due to essentially missing some statements in the code
<mofh>
I mean I'm both familiar with numerics and also completely backlogged to hell and back for the next 2 weeks at least, so.
<mofh>
:p
<egg|zzz|egg>
(bugs that could be caught only if you went through the right branch near a poorly-conditioned part, which the test vectors didn't)
<egg|zzz|egg>
mofh: s/weeks/months/
<galois>
egg|zzz|egg thinks mofh meant to say: I mean I'm both familiar with numerics and also completely backlogged to hell and back for the next 2 months at least, so.
<egg|zzz|egg>
mofh: or maybe years, see also the mofhstack
<mofh>
...OKAY FAIR POINT
* UmbralRaptop
pictures a 2.5 m high stack of papers
<egg|zzz|egg>
put a cat on top
<mofh>
UmbralRaptop: so I don't know if I've mentioned it, but my paper organization method in my room is literal stacks of papers each over a metre high
* UmbralRaptop
meeps
<mofh>
UmbralRaptop: sorting and organization happens when one of the stacks becomes, erm, structurally unsound. in a very sudden fashion.