UmbralRaptor 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> … one of the other grad students just compared me to nomal O_o | <ferram4> I shall beat my problems to death with an engineer.
<Fiora>
my hot-end is iirc....
<Fiora>
nozzle: brass, nickel-composite coated (has some other stuff like cubic boron nitride in it)
<Fiora>
thermal break: stainless steel
<Fiora>
heat block: brass or brass-plated steel (not sure) but i'm getting a replacement that seems to be aluminum
<Fiora>
cold end: aluminum
<egg>
Fiora: THF, also questionable with Ti fwiw
<Fiora>
oh dear
<egg>
and yes, questionable with brass
<bofh>
egg: that I want a [[Citation Needed]] on. I've used it with brass fine before.
<egg>
and aluminium bronze, but since it's questionable with aluminium and bronze we could have guessed
<Fiora>
oh right. that would explain why sodium hydroxide reacted with the cold end when I used it to clean it
<bofh>
Huh. TIL.
<Fiora>
like, not in a horrendous way, but it did
<egg>
!wpn bofh
* Qboid
gives bofh a combustion supercritical loop space
<egg>
!wpn Fiora
* Qboid
gives Fiora a rotary optocoupler
<Fiora>
basically i'm coming to conclude there may not actually be a reliable chemical means to clean PLA off nozzles
<Fiora>
i.e. a chemical which is a solvent (of any strength) for PLA but doesn't destroy common metals either
<Fiora>
.... all the things that say aluminum. aren't those actually not an issue? because all real aluminum is coated with aluminum oxide?
<bofh>
Oh, that's a good point.
<Fiora>
like, aluminum seems to be terribly reactive on that list, more than i would expect
<Fiora>
like, bleach??? reacting with aluminum ????
<bofh>
Yes, non-passivated Al is crazy reactive, but once the Al2O3 layer builds it's kinda okay.
<Fiora>
welp
<Fiora>
chloroform also reacts fairly strongly with iron
<Fiora>
and carbon steel
<Fiora>
lets see if dioxane does
<Fiora>
huh it doesn't have dioxane on the list
<Fiora>
naptha... at least goo gone is probably okay on these metals
<egg>
bofh: and if it can't reform, something something running shoes,
<bofh>
egg: that's metal fluoride passivation layers, but yes :P
<bofh>
Fiora: nonpolar solvents are always okay on metals.
<bofh>
for the most part, at least.
<egg>
bofh: yes, but the same applies right :-p
<egg>
bofh: also if I ever write Remez exchange for Principia, I should call it РемезExchange
<bofh>
ofc.
<bofh>
also you totally should join the I Implemented Remez Exchange club :P
<egg>
hah
<egg>
bofh: we're not there yet though
<egg>
bofh: also, if I'm correct on the 2nd question I asked Atlas and his human, that means a generalist numerics lib should strive to provide any common ill-conditioned function it can think of, because you want to have those first with bounded precision to compose them with well-conditioned stuff afterwards
<egg>
bofh: aka the math lib writer's duty is to tackle the shittiest problems they can :-p
<bofh>
like I'm pretty sure that's always been true, and I even have a Kahan quote I think to that effect.
<bofh>
but don't quote me on that
<egg>
bofh: yeah I guess even if you needed to put the ill-condition last you'd have the same problem
<bofh>
Yep.
<bofh>
today in "fucking seriously, ffmpeg": libavfilter has 3 different audio equalizer filters.
<bofh>
amusingly they all *do* have different usecases, but goddamn if filters.texi is terrible at pointing out how they differ.
* Ellied
spent most of the day trying to navigate the structure of the .mda files EPICS spits out
<Ellied>
PROF: do you want to stay here until tomorrow, or shall I take you h- ME: home please
<Ellied>
mda stands for "multi-dimensional array" which is already a bad sign, but I swear to god the structure fucking changes between collection runs
<bofh>
oh good god that sounds like the usual academic terror
e_14159- has joined #kspacademia
<Ellied>
plus, Fucking MATLAB
<bofh>
Do Not Want
<Ellied>
"No, you can't index that. No, you can't cast it to a datatype that you can index. No. You can't do that! Stop trying!"
e_14159 has quit [Ping timeout: 383 seconds]
e_14159- is now known as e_14159
<Ellied>
Calling the struct mdaload spits out produces a number of answers equal to the number of subscans, and the only way I can find for getting all of them is to do the [var1 var2 var3 var4 var5 var6] = mdadata.scan.sub_scans.subscans.detectors_data shit
<Ellied>
also yes, one sub-struct is called "sub_scans" and the next "subscans" but ONLY SOMETIMES, sometimes both are the same
* egg
stares at bazelbuild/bazel#374
<Qboid>
[#374] title: $ are not allowed in filename of a filegroup | Ultimately any character can be part of a filename. We should probably allow that.... | https://github.com/bazelbuild/bazel/issues/374
<egg>
still no news from that ddwolf giraffe on making it Unicode-friendly :-\
<Ellied>
there's a nice little utility someone made that converts an mda to an ASCII file, but it doesn't convert it to just ONE file, it spits out an unholy flood of files with weird as hell naming schemes
<egg>
!choose z|zz|zzz
<Qboid>
egg: Your options are: z, zz, zzz. My choice: zzz
<Ellied>
on the plus side, I now have my very own broken BNC and SHV cables.
<bofh>
00:50 <@Ellied> there's a nice little utility someone made that converts an mda to an ASCII file, but it doesn't convert it to just ONE file, it spits out an unholy
<bofh>
flood of files with weird as hell naming schemes
<bofh>
Oh these are SO MUCH FUN
<bofh>
Had to deal with a program like this when doing compchem stuff. It was goddamn outright bullshit.
<Ellied>
also, the MDA is probably like 6 GB, which then inflates to over 10 GB of ASCII, which takes longer than three laps around the storage ring by trike.
<bofh>
I love how there was something that could dump individual records but it required you to know (or guess) whether the record was a float or an int (or array of either) DESPITE THAT INFO BEING STORED.
<bofh>
so if you guessed wrong you ofc got utter bullshit
<bofh>
either binary representation of some double in base-10, or a bunch of NaNs depending on which direction the mismatch went.
<Ellied>
ha
<bofh>
oh and very rarely instead of NaNs you'd get utterly nonsensical valid FP values.
<Ellied>
aaaa
<bofh>
like fuck custom bespoke academic fileformats aaaaaaaaaaaaaaaargh
* UmbralRaptor
seems to recall breaking excel with LAMMPS output at one point because a file had too many lines for it.
<Ellied>
also custom bespoke academic programming languages. >:0
<UmbralRaptor>
Bonus points if there are proprietary libraries.
<Ellied>
like, I get why MATLAB finds so much use in physics because of that whole "very fast multiprocessed matrix operations" thing, but fuuuuuuuuuuuuuck
<Ellied>
it's so screwed up
<bofh>
by which you mean it calls out to LAPACK/MKL/etc :P
<bofh>
als, LAMMPS?
<bofh>
also*
<UmbralRaptor>
Well, I didn't run into to much format weirdness since it all was editable in vim.
<UmbralRaptor>
!acr -add:CWP Global Warming Potential
<Qboid>
UmbralRaptor: I added the explanation for this acronym.
<UmbralRaptor>
!acr -remove:CWP Global Warming Potential
<Qboid>
UmbralRaptor: I removed the explanation for CWP
<UmbralRaptor>
!acr -add:GWP Global Warming Potential
<Qboid>
UmbralRaptor: I added the explanation for this acronym.
<UmbralRaptor>
At least it's not as bad as SF6, apparently. >_>
icefire has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
<Fiora>
bofh: nonpolar okay on metals, huh
<Fiora>
actually that reminds me, can you mix nonpolar and polar solvents?
<Fiora>
ex: why does goo gone only have nonpolar solvents, not both?
<Fiora>
also lol that fluorine
<bofh>
depends on the solvents. for purely nonpolar, well, consider mixing hexane & water.
<bofh>
but then there's stuff like acetone which has complicated miscibility properties, & ofc detergents
Majiir is now known as Snoozee
Moistmelon has quit [Ping timeout: 204 seconds]
<egg|zzz|egg>
!wpn bofh, Fiora, UmbralRaptor, et al.
* Qboid
gives bofh, Fiora, UmbralRaptor, et al. a jellied thermistor-like culverin
egg|zzz|egg is now known as egg
<egg>
Fiora: I wonder how you pronounce the name of Wilbell voll Erslied
<e_14159>
egg: With difficulties, obviously.
<egg>
:D
<egg>
e_14159: I mean I'd be tempted to prounce it like it's german, [vilbɛl fɔl eːɐ̯sliːt] or somesuch
<e_14159>
Hm. It looks German-like (I mean, Lied is in there, Wilbell sounds okay, and voll is a German word), but it doesn't make much sense as a name
<e_14159>
Which, arguably, might be the definition of names :D
<Iskierka>
history is sampled like every frame which is a lot of trajectory to render in a long flight
<egg>
bofh: so, our (fixed-step because symplectic) history integration puts a point every 10 s
<egg>
bofh: even worse, under thrust we need to do an integration with the (user-provided, unpredictable) thrust at 50 Hz (the game physics framerate)
<egg>
bofh: that's a lot of segments to draw
<egg>
bofh: amusingly this method also solves the problem of trajectory interpolation for the predictions (which get predicted at a very large stepsize)
<egg>
bofh: but that's not directly the point, the point was that histories droped the framerate in the single digits (esp. after a RSS ascent where you have had 10 min of thrust in space or something like that)
<egg>
there's still #228, which is partially related but probably requires more care
<Qboid>
[#228] title: Keep the whole history | For the moment we compute the histories using `sampling_period == 0`, which results in large gaps in the trajectory at high timewarp (rendered as long straight line segments, see below). We should use `sampling_period == 1` instead, and downsample ourselves as needed when rendering.... | https://github.com/mockingbirdnest/principia/issues/228
icefire has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
<bofh>
egg: ahh I see
<egg>
@#$^ they changed the font in KSP 1.3 and their new font has no glyph for subscript 6
<egg>
which we used for Kerbin 6-hour days
<egg>
bofh: any idea for a symbol for a 6-hour day? >_<
<bofh>
Wait seriously? Why the fuck is their new font missing, like, essential glyphs?
<bofh>
I'd say ⑥ but somehow I suspect it lacks that too.
<egg>
bofh: so I tried the telegraph compatibility hour 6 and day 6 for maximum trolling but they don't have them
<egg>
bofh: also I find it weird that the i18n release drops Unicode support for some things
<egg>
but ⟨KSP⟩
<egg>
bofh: so um they do support 六, so I'm tempted to use that :D
<bofh>
LOL YES
<egg>
⑥ doesn't work of course
<bofh>
also seriously? wow that's kind of a major gap if their i18n font DROPS characters
<egg>
bofh: I'm not doing i18n for principia, because I do that crap at work, so it just has to work for english
<egg>
egg-style english, en-x-egg :-p
<egg>
international egg-style english, en-001-x-egg :-p
<egg>
bofh: 〦 works too :-p
<bofh>
LOL
<egg>
bofh: I guess I'll just be boring and call it d6, CJK will confuse users :-\
<egg>
bofh: FUCK their new font drops support for arabic; so when we have another release name with a persian or arab mathematician users won't see their name >_<
BPlayer has joined #kspacademia
APlayer has quit [Ping timeout: 383 seconds]
<UmbralRaptor>
Silly i18n question: is there a reason it's impractical to use eg: gnu unifont if system fonts aren't practical to get the entire BMP?
<UmbralRaptor>
(probably something of "if system fonts aren't practical")
* UmbralRaptor
wonders to what extent Principia uses the SMP and/or SIP.
<egg>
UmbralRaptor: well principia's comment and the title on GitHub use the SMP (mathematical alphanumerics :-p)
<egg>
UmbralRaptor: eventually we should get some cuneiform in
<UmbralRaptor>
ah
<Iskierka>
does KSP font support cuneiform?
* UmbralRaptor
would be shocked.
<UmbralRaptor>
(IIRC, it took digging around for symbola to see cuniform properly on my boxen)
<egg>
UmbralRaptor: there are nice cuneiform fonts out there though
<egg>
UmbralRaptor: and then on Windows there's Segoe UI Historic which has really good support for many things
<bofh>
egg: Segoe UI Historic cuneiform & Hieroglyphics are pulled directly from Noto's, fwiw
<UmbralRaptor>
Oh, right, noto is a thing also.
<bofh>
it's evdn credited in the font's NAME table
<bofh>
even*
* UmbralRaptor
wants to say that it's several gig, though?
BPlayer has quit [Ping timeout: 204 seconds]
<egg>
bofh: didn't know that, makes sense
<bofh>
several tems of megs
<bofh>
less if you ignore cjk whoch is technically a separate project snyhow
<egg>
[context: Perspective was templatized on the scalar, so you could have a perspective on points in a 3d space of catalytic activities, with a focal catalytic activity >_<]
<bofh>
fucking connection, seriously had a several-second RTT to many places just a second ago.
APlayer has joined #kspacademia
<bofh>
anyway egg I only know this b/c I went spelunking in the font in ttx and happened to read the credits line in the name table, which explained why those two fonts looked identical to me (spoilers: b/c they are, or rather that subset of one is identical to the other)
<kmath>
<dzsenits> 342. On Phobos, an object in a high-potential region can roll, accelerating, into a low-potential region, where it'… https://t.co/7GMMUIi0L9
<UmbralRaptor>
bofh: aaaaa!
<bofh>
Now my next question is what sort of orbit one would land into if they did that.
<bofh>
Like I'm pretty sure tripping and falling either gets you into a stable Mars orbit, or Roche'd, and I'm not sure which.
<bofh>
Hm. Do I want to do orbital mechanics by hand?...
<egg>
Fiora: hm, but then how would you pronounce Malier? [maliːɐ̯]?
icefire has joined #kspacademia
<Fiora>
egg: may-lee-er
<Fiora>
(I mean, the characters pronounce the names in game, so,)
<Fiora>
the only important thing though is that the two protagonists' names from that game were chosen solely to make an extremely bad multilingual pun, and for no other reason.
<egg>
Fiora: yeah, but then they definitely don't pronounce them in the German pronounciation :-p
<Fiora>
oh, they're not *all* german-inspired names. just LOTS of them
<Fiora>
somewhere between "riesengang" and "neuenmeuller",
<egg>
may-lee-er << also, having great trouble making sense of what that sounds like, something something english phonology >_>
<Fiora>
or i guess to use less ambiguous syllables (I can never remember IPA)
<Fiora>
may - lee - ur (schwa for the final vowel)
<egg>
Fiora: huh, so with an r pronounced in some way
Moistmelon has joined #kspacademia
<egg>
Fiora: hm, the way it's pronounced in ja sounds to me almost like [malje] (I'm bad at non-IPA transcriptions, but maybe mah lyay (not quite ay but I'm not even sure [e] is a thing in English), which would be rather more like the french reading of her name https://youtu.be/uWsKYsqYFZU?t=324
<kmath>
YouTube - Atelier Escha & Logy | Let's Play/Playthrough | Escha's Story | Part 1
<egg>
I apparently nested parentheses and forgot to close the outer one)
<Fiora>
in the english she says it with a pronounced R
<Fiora>
japanese transliterations of non-japanese names are always tricky since often the difference just comes from the way they pronounce syllables
<Fiora>
and not actually an intent to pronounce it differently
<Fiora>
since there is no "ur" in japanese
<egg>
yeah
<Fiora>
of course then it gets even messier since half the time the original authors mandate reasonable english spellings/pronounciations that don't match the japanese (because they're non-japanese to begin with)...
<egg>
Fiora: but then Wilbell is probably pronounced like "will" [wil] rather than german-style "veal" [vil] in english too, which seems odd
<Fiora>
... while the other half the time they mandate WEIRD ones that make no sense
<Fiora>
bonus points if they outright contradict themselves
<Fiora>
my current favorite is how Nasu has decided that "Arturia Pendragon" should be "Altria Pendragon", which makes no sense.
<egg>
Ꙩ_ꙩ
<Fiora>
King Arthur -> Arthur Pendragon -> Arturia Pendragon -> Altria (?!?!?!) pendragon
<Fiora>
apparently the western localization team basically had to say "uh... yeah, sorry, they insisted we use that spelling, we have no fucking clue why"
<egg>
at least the pendragon didn't turn into a pendulagon or something
<egg>
also where are my manners,
<egg>
!wpn Fiora
* Qboid
gives Fiora a stretchy grass snake which vaguely resembles a pommel
<Fiora>
but yeah i think she gets a W-sound not a V-sound? heh. i wonder what pronounciation they'd use in a german dub ;-)
kmath has quit [Remote host closed the connection]
cveron has quit [Quit: Leaving]
<egg>
Fiora: also Solle, the German pronunciation would dictate [zɔlə] (zoluh?), but apparently they pronounce it [sɔl] like Sol the star :-p https://youtu.be/oJKR6EVvNaM?t=49
<egg>
icefire: what happened to kmath?
<icefire>
updates
<egg>
Fiora: so there's a german trailer https://www.youtube.com/watch?v=bELX2rkKrhI, they say [malje] for Malier (so like the ja or fr pronunciation, not like a de reading), sadly Wilbell's name does not get said (there's a playthrough of ayesha commented in german, the player certainly does say it with a [v])
<Fiora>
heh, interesting. the players pronounce it using the obvious german approach XD
<egg>
Fiora: so fwiw I'd tend to read [malje] (because it doesn't look very de, and because I'm french), and obviously [vilbɛl fɔl eːɐ̯sliːt] (since I speak german and that looks very german) :-p
<icefire>
oh no, dependency conflicts. not how I wanted to spend my sunday
<icefire>
idk why they chose altria, arturia is closer to アルトリア than altria is.
<bofh>
!wpn egg, Fiora, Iskierka, icefire
* Qboid
gives egg, Fiora, Iskierka, icefire a salted spinlock
<soundnfury>
!wpn bofh
* Qboid
gives bofh a restricted observable
* egg
spins
<soundnfury>
egg: now you'll have to stay awake
<SnoopJeDi>
Thinking with Type ebook is $2.99 today. I wasn't planning on learning to be more miserable with text in general today, but sure let's go for it.
egg is now known as egg|nomz|egg
egg|nomz|egg is now known as egg
Qboid was kicked from #kspacademia by *status [You have been disconnected from the IRC server]
Qboid has joined #kspacademia
Qboid was kicked from #kspacademia by *status [You have been disconnected from the IRC server]
Qboid has joined #kspacademia
<egg>
!wpn bofh
* Qboid
gives bofh a Terran trie
<UmbralRaptor>
!wpn egg
* Qboid
gives egg a hafnium mutex
<egg>
!wpn UmbralRaptor
* Qboid
gives UmbralRaptor a repeater spectrophotometer
<kmath>
<rygorous> @pervognsen @pkhuong (See also the Jacobson Verlet + just-project physics paper for Hitman way back: particle dynam… https://t.co/kdkdGmdN5v
<kmath>
<noundefiner> godmother, n. one who godmoths
<egg>
whitequark: did your cat catch the giant moth
<egg>
or did you catch it by virtue of having been trained by your cat
<whitequark>
i caught it with a beaker
<whitequark>
the cat desensitized me to moths though
<egg>
bofh: oh, also what do you think of my algorithm for downsampling the trajectories
<bofh>
egg: haven't had a chance to read through it fully yet, will do that now. it's open in a tab.
<egg>
bofh: the idea is stepsize control by estimating the 2nd order error in the the tangent of the angle between the segment and the trajectory as a quarter of the tangent of the angle between the linear eggstrapolation of the trajectory and the far endpoint of the segment
<bofh>
okay that seems perfectly reasonable (also similar, but better than what I would've implemented).
<egg>
bofh: out of curiosity, what would you have implemented?
<kmath>
<whitequark> @hackerfriendly in fact I did
<bofh>
egg: like my initial naive approach was just doing 2nd-order error on some distance metric between the linear eggstrapolation of the trajectory and the far segment endpoint
<bofh>
where "some distance metric" in my mind was "ehh... sortaeuclidean?"
<bofh>
doing it with tangents of angles is much much more well-behaved
<egg>
bofh: so a possibility would be the euclidean metric on the projected points, that was my initial approach
<egg>
bofh: but 1. the projected velocity is ugly to compute; 2. shit happens at the edge of your field of view
<bofh>
yep, I was mostly thinking of 2. being the issue, 1. is possible to avoid with ugly approximation hacks (but then you run into ill-conditioning issues potentially)
<egg>
so that led me to angles (and their tangents because we don't want to compute no trig functions)
<egg>
bofh: 1. is doable just by writing the derivative, it's just that we don't have abstractions for it so you end up writing things in coordinates
<kmath>
<tompeck> This morning I've learned that the Otters See a Butterfly video is significantly improved by the addition of some S… https://t.co/0npVn5ru6l
<egg>
(via @spun_off)
TonyC has quit [Ping timeout: 204 seconds]
<egg>
!wpn bofh
* Qboid
gives bofh a diabolic surjection
<egg>
!wpn whitequark
* Qboid
gives whitequark a quacking glider
egg is now known as egg|zzz|egg
<whitequark>
!wpn egg
* Qboid
gives egg a transfinite prescriptive ?
<egg|zzz|egg>
oh a hieroglyphic knife
<egg|zzz|egg>
whitequark: have you eaten the moth
<whitequark>
nnnnno
<whitequark>
i still need to do the dishes from the previous time
<Fiora>
cisfinite
<egg|zzz|egg>
Fiora: yeah I just saw you tw... oh was that prompted by !wpn :-p
<egg|zzz|egg>
Fiora: also you have somehow nerd-sniped me into thinking about transfinite ordinals *proceeds to melt brain*
<egg|zzz|egg>
bofh: I feel like I've asked this question before, but can't we start the Veblen hierarchy earlier than φ_0? if φ_0 is the powers of ω, but I feel like one should have hierarchies of that style for earlier more pedestrian sequences
<bofh>
egg|zzz|egg: so oddly you haven't asked me that question before, but I haven't the foggiest to be honest. like, I agree it makes little sense.
<egg|zzz|egg>
hm, so then the question becomes which fixed-point equation ω^k is the kth solution of
<egg|zzz|egg>
but if you can already produce ω, let alone powers of ω, then you could plug them in that and get things that are higher in the Veblen hierarchy!
<egg|zzz|egg>
so that doesn't make sense...
<egg|zzz|egg>
what if we start from 1+x=x
<egg|zzz|egg>
the solutions of that are ω, ω+1, ω+2, ..., ω+ω, ..., ω+ω+ω, ... the fixed point is ω+k=k, which is ω^2
<egg|zzz|egg>
and then the solutions of ω+k=k are ω^2, ω^2+1, ω^2+2, ..., ω^2+ω, ..., ω^2+ω^2, ..., the fixed point is ω^2+k=k, which is ω^3
<egg|zzz|egg>
bofh: so the start of the Veblen hierarchy is an entire hierarchy starting with 1+x=x?
<bofh>
egg|zzz|egg: hmm. that is weirdly recursively defined.
<bofh>
I guess it's more a set of hierarchies or so.
<bofh>
so I guess hierarchy of hierarchies
<egg|zzz|egg>
bofh: the word set seems dangerous if we touch the transfinite
<egg|zzz|egg>
bofh: well it would make sense that you'd get a hierarchy of hierarchies of hierarchies of hierarchies of... oh wait