<UmbralRaptor>
If not, the next 2 months will require a thesis and ~20 credit hours.
<UmbralRaptor>
… though a surprise shortening of my MSc program could save the school ~10-15k?
UmbralRaptor has quit [*.net *.split]
UmbralRaptor has joined #kspacademia
ferram4 has joined #kspacademia
GreeningGalaxy has joined #kspacademia
ferram4 has quit [Read error: Connection reset by peer]
ferram4 has joined #kspacademia
icefire has quit [Read error: Connection reset by peer]
GreeningGalaxy has quit [Ping timeout: 180 seconds]
egg|phone|egg has joined #kspacademia
* egg|phone|egg
stabs timezones
<egg|phone|egg>
Stabbity
<UmbralRaptor>
!wpn egg
* Qboid
gives egg a Schwarz compression
<egg|phone|egg>
Moo
<egg|phone|egg>
UmbralRaptor: how goes the flattening of the earth
<whitequark>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg a monopole
* egg|phone|egg
has a monopoly over monopoles
<whitequark>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg a sharp ray tube
<whitequark>
!wpn -add:adj Sharp
<Qboid>
whitequark: Adjective added!
<whitequark>
!wpn -add:adj Renesas
<bofh>
!wpn whitequark
<Qboid>
whitequark: Adjective added!
* Qboid
gives whitequark a state-of-the-art ?
<whitequark>
!wpn -add:adj Hitachi
<Qboid>
whitequark: Adjective added!
<whitequark>
!wpn -add:wpn cryocooler
<Qboid>
whitequark: Weapon added!
<egg|phone|egg>
!wpn -add:adj Darboux
<Qboid>
egg|phone|egg: Adjective added!
<whitequark>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg a submillimeter bird
<whitequark>
that... is impressive
<whitequark>
!wpn -add:wpn tardigrade
<Qboid>
whitequark: Weapon added!
<egg|phone|egg>
Diapsid !
<whitequark>
!wpn -add:adj biologically active
<Qboid>
whitequark: Adjective added!
<whitequark>
!wpn -add:adj harmless
<Qboid>
whitequark: Adjective added!
<whitequark>
!wpn -add:adj completely harmless
<Qboid>
whitequark: Adjective added!
<whitequark>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg an elliptical base/surjection hybrid
<whitequark>
!wpn egg|phone|egg
* Qboid
gives egg|phone|egg an unit nonagon
<egg|phone|egg>
!Wpn whitequark
* Qboid
gives whitequark a Renesas measure
* egg|zzz|egg
pets UmbralRaptor with a bird
* egg|zzz|egg
pets whitequark with a cat
* whitequark
pets egg with an egg
<egg|zzz|egg>
hmm, eggs don't seem quite fluffy enough for that purpose
<egg|zzz|egg>
whitequark: UmbralRaptor: eggsplanation time, because I woke up early. Consider this thing http://vf.rohanp.xyz/ and put in something like y, -x. The plane then represents the phase space of a harmonic oscillator, say a mass on a spring: horizontally you have position, vertically you have velocity. As stuff flows around the phase space it happily rotates forever, and you immediately see why you get sinusoids
<egg|zzz|egg>
from a harmonic oscillator: they're one coordinate of a circle in the phase space.
<egg|zzz|egg>
(so here it would be saner to rename y to v, you get x'=v, v'=-x)
<egg|zzz|egg>
or maybe we could rename x to q and y to p, because it's usual to deal with qositions and pomenta
<UmbralRaptor>
Are we measuring Henergy?
<egg|zzz|egg>
right, so let's talk about Hnergy: that particular force derives from a potential (1/2 x**2), so to every point (x, y) in the phase space we can associate its energy
<egg|zzz|egg>
kinetic (y**2 / 2) plus potential (x**2 / 2)
<egg|zzz|egg>
(behold it's half the squared distance from the origin)
<egg|zzz|egg>
but since it's distance from the origin, it means that the region of equal energy is just a circle
<egg|zzz|egg>
(and now we've shown that we'll follow circles just from conservation of energy)
<egg|zzz|egg>
now, conservation of energy alone predicts our trajectory here, because our phase space is 2d, and that's a single constraint on there (yielding our 1d trajectory)
<egg|zzz|egg>
in higher dimension it would not suffice
<egg|zzz|egg>
e.g. were our physical system 2d rather than 1d, our phase space would be 4d, and energy would constrain us to 3d surfaces in the phase space
<egg|zzz|egg>
whitequark: UmbralRaptor: now, if you let a rectangle flow in that particular vector field, you have noticed that it rotates, and doesn't change shape nor surface. The harmonic oscillator is mildly boring at times, let's look at slightly more complex (but still 1d) physical system
<egg|zzz|egg>
the pendulum!
<egg|zzz|egg>
pick something like y/3, -Math.sin(x/3) as the vector field
* UmbralRaptor
doesn't have a fast enough bird to run that page well.
<egg|zzz|egg>
the /3 are there just to magnify things a bit
<egg|zzz|egg>
UmbralRaptor: ow
<egg|zzz|egg>
so, here, x is the angle of a pendulum from the vertical, y is its angular velocity
<egg|zzz|egg>
near the origin, it's a spinny thing that looks like the phase space of the harmonic oscillator from above
<egg|zzz|egg>
(that's why grandfather clocks work: for small oscillations the pendulum is close to harmonic)
<egg|zzz|egg>
(because sin x = x)
<SnoopJeDi>
~=
<egg|zzz|egg>
if you start far enough above or below, you just get shoved endlessly right respectively left, with some deformation, but never going back: that's a pendulum started with enough energy that it keeps turning in the same direction, going over the top
<egg|zzz|egg>
if you start reasonably close to the origin though, you see your rectangle get deformed over time, because it's still not quite a harmonic oscillator
<egg|zzz|egg>
but an important property of this kind of physical system is that the red area will remain constant
<egg|zzz|egg>
it turns out that you can make integrators that do preserve that area, even if they inaccurately propagate actual trajectory: those are called symplectic integrators
<egg|zzz|egg>
it also turns out that symplectic integrators not only preserve that seemingly abstract property, but also they have no long-term energy drift, which is slightly surprising and really useful
<egg|zzz|egg>
(nonsymplectic integrators either boost you or drag you down eventually)
<egg|zzz|egg>
oh yeah and there's a strong dependence on initial conditions here if you start your pendulum upside-down. notice that this stretches your rectangle madly (because its area has to remain constant while it gets looong). Also notice that eventually the simulation breaks down and it doesn't have enough points on the rectangle and your screen turns red.
<egg|zzz|egg>
whitequark: SnoopJeDi: UmbralRaptor: this has been this morning's eggsplanation, you may now return to a normal activity :-p
<SnoopJeDi>
normal or nomal?
<egg|zzz|egg>
is there a difference?
<SnoopJeDi>
ouch
<egg|zzz|egg>
whitequark: also, has this been understandable?
<UmbralRaptor>
Did it make you anxious?
* UmbralRaptor
stabs things, and wonders why multiling o keyboard is suddenly so laggy/crashy.
<whitequark>
egg|zzz|egg: ummmm i'm cooking right now
<whitequark>
haven't read that yet
<egg|zzz|egg>
whitequark: oh that would eggsplain the twitter :-)
<egg|zzz|egg>
whitequark: also, pancakes? now I'm hungry :-p
<egg|zzz|egg>
(it's half past six, haven't had breakfast yet)
<UmbralRaptor>
mmm... pancakes
<UmbralRaptor>
And srsly. There's only 1 IKEA in Kansas, and it opened in 2015?
<whitequark>
egg|zzz|egg: I don't even try to tie my schedule time to solar time anymore
<whitequark>
UmbralRaptor: are you @newpa_hasai?
<UmbralRaptor>
I think still 0 in Missouri, Iowa, and Nebraska.
<egg|zzz|egg>
whitequark: he is
<UmbralRaptor>
whitequark: yeah
<egg|zzz|egg>
he is a diapsid, too
<egg|zzz|egg>
whitequark: yes, I noticed that the time in Hong Kong was utterly useless to predict whether you were asleep
<egg|zzz|egg>
it's also quite close to ICT, I always forget that it's mostly north
<UmbralRaptor>
Though this leads to the question of why a package with instructions in Chinese chose US measurements for the English portion.
<egg|zzz|egg>
did the english of HK not undergo decimalization?
<whitequark>
that's the first time I see °F anywhere
<egg|zzz|egg>
UmbralRaptor: maybe because pancakes?
<whitequark>
or non-metric
<whitequark>
except for uhhh
<whitequark>
for some reason apartment surface is measured in square feet
<egg|zzz|egg>
Wikipedia: "Hong Kong has three main systems of units of measurement in current use:" (Chinese Qing, British imperial, metric)
<UmbralRaptor>
Aaaaaaa
<egg|zzz|egg>
"The traditional measure of flat area is the square foot (Chinese: 方呎, 平方呎) of the imperial system, which is still in common use for real estate purposes. The measurement of agricultural plots and fields, however, is traditionally conducted in 畝 (mau) of the Chinese system."
<egg|zzz|egg>
whitequark: so if you buy a field it will not be in square metres or perches or roods or acres! otoh it still won't be in hectares :-p
<egg|zzz|egg>
s/metres/feet
<Qboid>
egg|zzz|egg meant to say: whitequark: so if you buy a field it will not be in square feet or perches or roods or acres! otoh it still won't be in hectares :-p
<whitequark>
hk has agriculture?
<whitequark>
I thought it just imported everything :p
<egg|zzz|egg>
whitequark: yeah tbh I'm not sure where O_o
<egg|zzz|egg>
now they don't have an airport where you land between the buildings though :-p
<whitequark>
yeah, land reclamation is huge
<whitequark>
wonder what's HK going to do once sea level rises 3-5m or so
<whitequark>
most of the new residential complexes are built under 5m
<UmbralRaptor>
4D dairy farms.
<egg|zzz|egg>
!wpn whitequark
* Qboid
gives whitequark a parallel mission-like guillotine
* Qboid
gives whitequark and UmbralRaptor a boron complex alternator
<egg|zzz|egg>
wait, no, that doesn't follow math citation conventions!
<egg|zzz|egg>
!wpn UmbralRaptor and whitequark
* Qboid
gives UmbralRaptor and whitequark a Hitachi quadrupole photomultiplier tube
<egg|zzz|egg>
here we go
<egg|zzz|egg>
whitequark: context: in math we don't have first and last authors, you just list all authors alphabetically
<egg|zzz|egg>
(this works well in part because the number of authors in maths papers is small, in high-energy physics experimental monstrosity that might get weird)
<egg|zzz|egg>
whitequark: does it support Galileo?
<egg|zzz|egg>
GLONASS?
<Qboid>
egg|zzz|egg: [GLONASS] => ГЛОбальная НАвигационная Спутниковая Система
<egg|zzz|egg>
\o/
<whitequark>
ГЛОНАСС?
<Qboid>
whitequark: [ГЛОНАСС] => ГЛОбальная НАвигационная Спутниковая Система
<egg|zzz|egg>
whitequark: yeah, for acronyms that are commonly used with both alphabets I do that
<egg|zzz|egg>
I think there's another acronym for another arm
<egg|zzz|egg>
VKM?
<Qboid>
egg|zzz|egg: [VKM] => Верхней Кабель-Мачта
<egg|zzz|egg>
whitequark: those two get used in latin too, because of Союз lauches from Kourou
<egg|zzz|egg>
whitequark: btw, were the eggsplanations above understandable? if so, were they interesting?
<whitequark>
!acr -add:VKM Верхняя Кабель-Мачта
<Qboid>
whitequark: I already know an explanation for VKM! (Update it with !acr -update:VKM Верхняя Кабель-Мачта)
<whitequark>
!acr -update:VKM Верхняя Кабель-Мачта
<Qboid>
whitequark: I updated the explanation for VKM
<whitequark>
!acr -update:ВКМ Верхняя Кабель-Мачта
<Qboid>
whitequark: I updated the explanation for ВКМ
<egg|zzz|egg>
wrong case I gather?
<whitequark>
declinations.
<whitequark>
yes.
* egg|zzz|egg
should learn russian someday >_>
<whitequark>
about the explanations
<whitequark>
they're not especially insightful to me because I've never had the problem they're trying to solve
<whitequark>
so it's more of a "sure, i guess"
<whitequark>
the high-level concept of a symplectic integrator makes sense to me and its existence is interesting
<whitequark>
the exact details of operating one seem irrelevant
<egg|zzz|egg>
well those aren't so much the details of how they operate, as it is "what it means to be symplectic"; tbh the implication that energy doesn't drift is what you actually use, rather than symplecticity
<egg|zzz|egg>
but the difference can matter, e.g. with conjugate-symplectic integrators not being the same as symplectic integrators when it comes to doing things like nesting them or whatnot
<whitequark>
knowing these differences makes absolutely no, uh, difference to me
<whitequark>
I'm sure I'm just not an appropriate audience for this.
<egg|zzz|egg>
so it can matter if you're trying to nest things that work at different timescales, e.g. integrate many steps of the earth around the sun between steps of Neptune around the sun
<egg|zzz|egg>
that's one way to do solar system integration (there are a few papers in the sea about that), but it requires actually symplectic integrators, conjugate-symplectic won't do
<egg|zzz|egg>
(conjugate symplectic means a step of it is, or is equivalent to, Something [a step of some symplectic integrator I] Something^-1, which means that if you repeat it 100 times, you get the equivalent of Something [100 steps of I] Something^-1; so for practical purposes it's as if you were using I, that perturbation by Something only occurs at the ends. otoh if you do more than composing it with itself, e.g.
<egg|zzz|egg>
nesting it at different timescales, this breaks down)
<whitequark>
now *this* is much more insightful
<egg|zzz|egg>
\o/
<egg|zzz|egg>
whitequark: it's probably more practical, too, so I'd classify it closer to the details of operating one (one really doesn't need to know what symplectic means to use such an integrator). We might not be using the same definition of operate, or eggsact details :-)
<egg|zzz|egg>
whitequark: also this is the same "conjugate" you might know from linear algebra, it's just that we're doing it with arbitrary operations in some group (a Lie group I think) rather than matrices
<egg|zzz|egg>
!wpn
* Qboid
gives egg|zzz|egg a partially-ordered missile/hammer hybrid
<Thomas>
!wpn egg
* Qboid
gives egg a surjective pommel with a thyristor attachment
<egg|zzz|egg>
!wpn Thomas
* Qboid
gives Thomas a Trojan bundle
<whitequark>
!wpn egg|zzz|egg
* Qboid
gives egg|zzz|egg a depleted pentode
<whitequark>
!wpn -add:adj [REDACTED]
<Qboid>
whitequark: Adjective already added!
<whitequark>
!wpn -add:wpn [REDACTED]
<Qboid>
whitequark: Weapon already added!
<whitequark>
...
<whitequark>
okay
<egg|zzz|egg>
whitequark: that would probably be Greening's work
<egg|zzz|egg>
hmm, so apparently YouTube has some snakes in it, I might end up writing snakes
<soundnfury>
hmm, what's the canonical way to !acr something with nested acronyms?
<soundnfury>
(do the subacronyms have to be expanded in the definition?)
<egg|zzz|egg>
MaDCoWS?
<Qboid>
egg|zzz|egg: [MaDCoWS] => Massive and Distant Clusters of WISE Survey
<egg|zzz|egg>
soundnfury: also, precedence for astronomy over jargon file nonsense:
<egg|zzz|egg>
!acr -add:WOMBAT Wavelength-Oriented Microwave Background Analysis Team
<Qboid>
egg|zzz|egg: I already know an explanation for WOMBAT! (Update it with !acr -update:WOMBAT Wavelength-Oriented Microwave Background Analysis Team)
<soundnfury>
!acr -add:LDIR LoaD, Increment and Repeat
<Qboid>
soundnfury: I added the explanation for this acronym.
<Thomas>
!acr -remove:WOMBAT
<Qboid>
Thomas: I removed the explanation for WOMBAT
<egg|zzz|egg>
soundnfury: nah, stop with the opcodes, you'll instantly clog the space of 3-letter thingies
<soundnfury>
!acr -add:DJNZ Decrement and Jump if Not Zero
<Qboid>
soundnfury: I added the explanation for this acronym.
<egg|zzz|egg>
or ask Thomas for a !opcode
<egg|zzz|egg>
!acr -add:WOMBAT Wavelength-Oriented Microwave Background Analysis Team
<Qboid>
egg|zzz|egg: I added the explanation for this acronym.
<soundnfury>
egg|zzz|egg: but I'm not doing the 3-letter ones
<egg|zzz|egg>
soundnfury: stop with the 4-letter words
<soundnfury>
I'm only doing long and interesting ones
<soundnfury>
!acr -add:FSCK File System Consistency checK
<Qboid>
soundnfury: I added the explanation for this acronym.
<soundnfury>
;)
* egg|zzz|egg
slaps soundnfury with a porpoise
<whitequark>
that's not even an acronym
<whitequark>
that's just stupid unix naming
<Thomas>
hah
<soundnfury>
if "sinh" is pronounced "shine",
<soundnfury>
how is "sith" pronounced?
<soundnfury>
whitequark: hey, at least I didn't -add GREP
<egg|zzz|egg>
sinh is properly written sh anyway
<soundnfury>
egg|zzz|egg: eresyh!
<whitequark>
egg|zzz|egg: does "anciens" really just mean "old"
<egg|zzz|egg>
whitequark: well it's plural masculine, and yes, mostly? I mean you wouldn't say that to say someone is old, but in most uses it works I think?
<whitequark>
egg|zzz|egg: so whats the french word for "ancient"
<egg|zzz|egg>
so if you really want really old, maybe you can use antique?
<egg|zzz|egg>
but ancien can work too?
<egg|zzz|egg>
I mean translation is rarely bijective :-p
<Qboid>
Thomas: Aliased "!csharp opcodes_.ContainsKey("{0}") ? ("[{0}] => " + opcodes_["{0}"]) : "Opcode not found."" to "!opcodes"
<egg|zzz|egg>
whitequark: for your ancien/ancient question
<Thomas>
!csharp -persistent opcodes_.Add("DJNZ", "Decrement and Jump if Not Zero")
<Thomas>
!opcodes DJNZ
<Qboid>
[DJNZ] => Decrement and Jump if Not Zero
<egg|zzz|egg>
!acr -remove:DJNZ
<Qboid>
egg|zzz|egg: I removed the explanation for DJNZ
* soundnfury
applauds Thomas
<egg|zzz|egg>
soundnfury: here, have fun (in #bottorture though, because this is likely to get spammy)
<Thomas>
And please dont fill the evaluator with opcodes
<Thomas>
:P
<egg|zzz|egg>
whitequark: so wiktionary points out something that sounds reasonable, which is that it depends on whether it's pre- or postfix https://en.wiktionary.org/wiki/ancien#French
<egg|zzz|egg>
prefix might be the previous one, postfix is closer to the meaning of ancient
<egg|zzz|egg>
oh, and 5.: it serves for "the elder"
<whitequark>
egg|zzz|egg: regarding Черенков, you're using a noun as an adjective
<whitequark>
that doesn't work
<egg|zzz|egg>
whitequark: well, the idea is for things named after people, e.g. a Galois extension, the Alexandroff compactification, Cherenkov radiation...
<whitequark>
yes. that works in english. but not in russian
<egg|zzz|egg>
right, because declension
<whitequark>
"Galois extension" becomes something like "extension of Galois" with "of" merged into "Galois"
<egg|zzz|egg>
yeah, I'm familiar with languages that have declension
<egg|zzz|egg>
whitequark: my idea here was that Qboid generates weapons in English, but I don't transliterate names because that avoids the confusion over Tchebychoff/Chebyshev/etc. (much like in the identifiers in my code, that class isn't called МногочленыЧебышёва)
<egg|zzz|egg>
if it generated weapons in other languages it would pretty much instantly break down because it doesn't support grammatical gender for one thing
<whitequark>
egg|zzz|egg: I guess...
<egg|zzz|egg>
(grammatical gender is a fun mess, a recurrent topic here is the gender of planets in french)
<whitequark>
it's not particularly fun
<egg|zzz|egg>
Jupiter the dude is masculine, but in modern usage I'm pretty sure I'd refer to Jupiter the planet as feminine (because there Jupiter stands for "la planète Jupiter")
<egg|zzz|egg>
and then for moons, one can wonder whether one uses the word for moon (feminine) or satellite (masculine)
<egg|zzz|egg>
(or the grammatical gender of whatever the moon is named after)
<Iskierka>
what about planetoids?
<egg|zzz|egg>
ah planetoid is masculine, as is asteroid and anything in -oid I think?
<egg|zzz|egg>
comets are feminine though
<egg|zzz|egg>
(what grammatical gender do you use for things that have dual comet/asteroid designation?)
<Iskierka>
you abolish the silly idea of grammatical gender
<egg|zzz|egg>
huh, in German comets are masculine
<egg|zzz|egg>
Iskierka: so apparently in German comets, asteroids, satellites, moons, and planets are all masculine, so I guess this simplifies things?
<egg|zzz|egg>
stars are masculine too, but the sun is feminine
<egg|zzz|egg>
(in french that's the reverse)
<egg|zzz|egg>
and then you have Gestirn which is neuter
<egg|zzz|egg>
<egg|zzz|egg> ah planetoid is masculine, as is asteroid and anything in -oid I think? <<<< wait no stupid egg, plenty of feminine stuff in -oid, all kinds of curves for instance
<egg|zzz|egg>
and surfaces
<egg|zzz|egg>
huh, I never noticed the discrepancy: catenoids are feminine, and hyperboloids are masculine
<egg|zzz|egg>
(hyperbolae are feminine though)
<egg|zzz|egg>
Iskierka: it *is* quite silly :-p
* whitequark
stares at a C++ lambda
<whitequark>
x = [=]() { ...
<whitequark>
this is fucked up
<egg|zzz|egg>
whitequark: implicit capture by copy?
<whitequark>
no
<whitequark>
the syntax
<egg|zzz|egg>
ah yes it is
<egg|zzz|egg>
whitequark: somehow they haven't quite managed to put all of their brackets into that syntax
<egg|zzz|egg>
whitequark: I await a language construct that's []<>(){} so we can be completely mad
technicalfool has joined #kspacademia
<whitequark>
egg|zzz|egg: they're almost there.
<whitequark>
[=]()->{}
<egg|zzz|egg>
whitequark: the fuck is that
<egg|zzz|egg>
oh, right, postfix return type?
<egg|zzz|egg>
but can it be empty?
<whitequark>
no I just omitted it
<whitequark>
>auto y = [&r = x, x = x + 1]()->int
<whitequark>
the fuck is that
<egg|zzz|egg>
capture expression by reference
<egg|zzz|egg>
(for the first one)
<egg|zzz|egg>
(the second one is by copy)
<egg|zzz|egg>
so in the body of the λ, r refers to x, by reference; x otoh, refers to the value x+1 by copy at the point of declaration of the λ
<egg|zzz|egg>
whitequark: the syntax is nuts, but also doing this in particular seems quite nuts too
<whitequark>
question
<whitequark>
does std::bind(..., x) take x by copy or by reference
<egg|zzz|egg>
I *think* copy, you have to do std::ref to pass by reference
<egg|zzz|egg>
I've gotten my arse bitten by that
<egg|zzz|egg>
std::bind is a booby trap
<egg|zzz|egg>
(or std::cref)
<egg|zzz|egg>
whitequark: yes, C++ is barking mad
<whitequark>
egg|zzz|egg: it's not even that
<whitequark>
C++ is like American politcs.
<whitequark>
every day I wake up with the expectation that it can't possibly be worse, and yet it is worse.
<egg|zzz|egg>
(also this syntax annoys me even more since I'm more of a keywordy syntax kind of person, but this is beyond such matters of taste)
<egg|zzz|egg>
whitequark: I'd say you're better off using lambdas than bind though, less of a trap with arguments expected by reference
ferram4 has joined #kspacademia
<egg|zzz|egg>
whitequark: at least it has value genericity with a turing-complete type system I guess (though it's not contractual and breaks in the most awful ways)
<egg|zzz|egg>
so, we're having issues replaying our journals in principia
<egg|zzz|egg>
are we stomping memory? is the journaling code wrong? is there some other external state around? doesn't look like the FP environment is at play, even though it's FP checks that fail, because redirected rounding would error everywhere, and not suddenly quite late with a big error...
<egg|zzz|egg>
(it's quite possible that we're stomping something)
<egg|zzz|egg>
(sob(
<egg|zzz|egg>
))
<whitequark>
um
<whitequark>
valgrind it?
<egg|zzz|egg>
whitequark: hmm, I don't really think we have adequate tooling to do that :-\
<egg|zzz|egg>
also, somehow the error is just 300 ULPs, that seems very small (I'd expect stomping to yield garbage mostly) and that's too big to be redirected rounding ........
<egg|zzz|egg>
stabbity
GreeningGalaxy has joined #kspacademia
<egg|zzz|egg>
!wpn GreeningGalaxy
* Qboid
gives GreeningGalaxy a ½λ power supply
<GreeningGalaxy>
!u ?
<Qboid>
U+1F52A HOCHO (?)
<GreeningGalaxy>
dammit, why doesn't that show up
<GreeningGalaxy>
I have Noto installed, ffs
GreeningGalaxy has quit [Read error: No route to host]
* Qboid
gives Iskierka a really big late 80s otter
<egg|zzz|egg>
!wpn GreeningGalaxy
* Qboid
gives GreeningGalaxy a potassium trihexaflexagon
<egg|zzz|egg>
!wpn whitequark
* Qboid
gives whitequark a trapezohedron
<egg|zzz|egg>
ok, at least it appears to work with RSS
<egg|zzz|egg>
(or at least not be more broken)
<egg|zzz|egg>
we'll probably ignore the nonreproducibility for Cardano :-\
* egg|zzz|egg
suspects *some* sort of stomping
<egg|zzz|egg>
the pile up stuff has way too many tricky cross-references
whitequark has quit [*.net *.split]
Norgg has quit [*.net *.split]
Orkeren_ has quit [*.net *.split]
X has quit [*.net *.split]
icefire has quit [*.net *.split]
Snoozee has quit [*.net *.split]
soundnfury has quit [*.net *.split]
Soozee has quit [*.net *.split]
Iskierka has quit [*.net *.split]
Hydragyrum has quit [*.net *.split]
Thomas has quit [*.net *.split]
kmath_ has quit [*.net *.split]
technicalfool has quit [*.net *.split]
egg|phone|egg has quit [*.net *.split]
TheKosmonaut has quit [*.net *.split]
GreeningGalaxy has quit [*.net *.split]
UmbralRaptor has quit [*.net *.split]
egg|zzz|egg has quit [*.net *.split]
Parenthesie has quit [*.net *.split]
Orkeren has quit [*.net *.split]
TheKosmonaut has joined #kspacademia
Hydragyrum has joined #kspacademia
whitequark has joined #kspacademia
egg|zzz|egg has joined #kspacademia
Norgg has joined #kspacademia
Orkeren has joined #kspacademia
technicalfool has joined #kspacademia
kmath_ has joined #kspacademia
icefire has joined #kspacademia
X has joined #kspacademia
GreeningGalaxy has joined #kspacademia
UmbralRaptor has joined #kspacademia
egg|phone|egg has joined #kspacademia
Snoozee has joined #kspacademia
Thomas has joined #kspacademia
soundnfury has joined #kspacademia
Orkeren_ has joined #kspacademia
Iskierka has joined #kspacademia
Soozee has joined #kspacademia
Parenthesie has joined #kspacademia
ferram4 has quit [Read error: Connection reset by peer]
ferram4 has joined #kspacademia
<egg|zzz|egg>
whitequark: so yeah, we're going to make a release on the 28th, but there will probably be some weird stomping in it >_> <_< >_> (also other bugs probably? but non-reproducibility worries me most)
<whitequark>
mmm
<whitequark>
nice
<egg|zzz|egg>
whitequark: well, I'd rather it be reproducible :-p
<kmath_>
<whitequark> @uxcn_ yeah I've banned non-const reference arguments in my projects
<egg|zzz|egg>
whitequark: in principia we allow pass-by-reference, but taking the address of an argument passed by reference is a cardinal sin
<egg|zzz|egg>
(and references as class members get you thrown into a bucket of interhalogens)
<egg|zzz|egg>
(with an exception for constreferences to static constants for integrators)
<whitequark>
it's really bad that I can pass something in and not know if it's mutated or not
<egg|zzz|egg>
that doesn't fundamentally bother me? the lifetime questions otoh worry me
<egg|zzz|egg>
whitequark: a lot of our stuff is in pointers anyway, so the argument that "you see the &" if you take mutable arguments by pointer doesn't really hold
<whitequark>
that's kinda the opposite of the solvespace situation
<whitequark>
everything is either a local, a semi-global, is owned by the sketch and accessed by index, or in shared_ptr
<egg|zzz|egg>
whitequark: I guess in lower-level stuff you tend to have less "everything on the heap cross-referenced" madness
<soundnfury>
But Szch would be Щ!
* soundnfury
performs void pointer arithmetic on egg|zzz|egg
<egg|zzz|egg>
whitequark: amusingly (sorry I keep blathering about dead languages) Ada's in out and out parameters don't have a marking at the call site; Green (the earlier draft, before Ada 83) had a syntax where if you named the parameter at the call you could do Formal_Parameter := Value, Formal_Parameter :=: Variable, or Formal_Parameter =: Variable
<kmath_>
<whitequark> @oshepherd @mcclure111 if you don't terminate the file with a newline the compiler splices a "trusting trust" attack into its binary
<egg|zzz|egg>
!wpn
* Qboid
gives egg|zzz|egg a bottom wing-like trihexaflexagon
<kmath_>
<Newpa_Hasai> @whitequark I Thea that it was Nice and large, but you'd need to find someone more Wiley, Elsevier the numbers won't be precise.
<egg|nomz|egg>
aaaaaa
<UmbralRaptor>
<_<
<egg|nomz|egg>
UmbralRaptor: also there's a Nice 2 right? o_O
SnoopJeDi has quit [Read error: Connection reset by peer]
SnoopJeDi has joined #kspacademia
<egg|nomz|egg>
!wpn whitequark
* Qboid
gives whitequark a Bolzano-Weierstraß missile
<egg|nomz|egg>
the missile is bounded, and has a submunition that converges on its target
<egg|nomz|egg>
Bolzano-Weierstraß jokes \o/
egg|nomz|egg is now known as egg|zzz|egg
<SnoopJeDi>
An unbounded missile prototype would be a shoe-in for DOD funding
<egg|zzz|egg>
SnoopJeDi: but then there's a risk that all submunitions be divergent
<SnoopJeDi>
feature, not a bug
<egg|zzz|egg>
SnoopJeDi: so in french divergence means the start of chain reaction for nuclear reactors, that would make a fun pun; sadly I don't think that's the case in English?
<SnoopJeDi>
what's wrong with using `less` to read diffs? ;P
Majiir is now known as Snoozee
<UmbralRaptor>
"Why are you editing a novel with github?"
* UmbralRaptor
hands Snoozee a liter of coffee.
<GreeningGalaxy>
I think I've sneezed fifty times in the last hour
<GreeningGalaxy>
fuckin allergies man
<GreeningGalaxy>
One more hour, then I'm taking benadryl and going to be early
<UmbralRaptor>
Plants are evil.
<GreeningGalaxy>
going to bed* early
<GreeningGalaxy>
bofh / whitequark etc: I got an idea; since I can't find SOIC-pitch perf for less than $30 per ~20cm sheet, I thought it might work well to make a design for a simple grid of squares at 0.05" pitch and send it to some circuit board print shop. We're already planning on doing that for the photon counter design I've been working on, so it seems reasonable.
<GreeningGalaxy>
Depending on cost, I could even do cute little things like add a 20x2 PTH socket in 0.1" pitch along one edge with traces linking each pin to the first row of pads, to accommodate an RPi GPIO header.
<GreeningGalaxy>
custom protoboard sounds like a lot of fun though, and it sounds like I might be able to get it done for only about a buck per 10x10cm sheet if I get it in quantity.
<GreeningGalaxy>
I am also apparently not going to be the first to try something like this: http://imgur.com/N3Clans
<GreeningGalaxy>
right, I have endless uses for those.
<GreeningGalaxy>
there's a huge range of different kinds of protoboards available. There's even ones that are the same size, shape, and wiring layout as a solderless breadboard, but those are extremely stupid IMO because they constrain you too much and force you to use a mess of jumper wires.
<GreeningGalaxy>
I like the stuff that's just got a grid of holes with copper pads, and then you can use solder bridges to form traces. Only problem is that it's too large for SMD chips.
<GreeningGalaxy>
so enter the stuff I have now, 0.05" Vectorbord. Regular old DIP chips fit just fine, but now you can squeeze solder bridge traces between their legs, and also solder on SOICs as you would a DIP to normal-size perf.
<GreeningGalaxy>
that's expensive though and the holes are all connected through to the other side, so you're kind of constrained to one layer of traces.
<GreeningGalaxy>
I want something without holes everywhere (although still some, so I can make vias by threading wire through) in the pitch of SOIC chips. It will be perfect for making fairly complex prototypes in a small space (e.g. RPi HAT-sized board)
<GreeningGalaxy>
Iskierka: basically, the solderable protoboard stage is the stage in between "got it working on a solderless breadboard" and "sending it into some company to get it printed into real boards"
<GreeningGalaxy>
for research I make a lot of one-off devices that I don't need in quantity but I need to stick together for a long time and not take up a lot of room; solderable perf is perfect for this.
<soundnfury>
!wpn GreeningGalaxy
* Qboid
gives GreeningGalaxy a jellied measure
<soundnfury>
!wpn add:wpn salmiak
* Qboid
gives add:wpn salmiak a Максутов-Cassegrain hypothesis
<soundnfury>
!wpn -add:wpn salmiak
<Qboid>
soundnfury: Weapon added!
* egg|zzz|egg
should bind Ꙭ to some key instead of using O_O
<egg|zzz|egg>
or maybe I could use Ꙩ_Ꙩ
<soundnfury>
!u Ꙭ
<Qboid>
U+A66C CYRILLIC CAPITAL LETTER DOUBLE MONOCULAR O (Ꙭ)