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.
<kmath>
<TiroTypeworks> 18th C. East Syrian liturgical manuscript showing tendency to use stroke extension to justify the last word in a li… https://t.co/h4sFNwi017
awang has quit [Ping timeout: 194 seconds]
<egg>
!u –
<Qboid>
U+2013 EN DASH (–)
<egg>
stabbity
<iximeow>
!u –
<Qboid>
U+2013 EN DASH (–)
<iximeow>
!u —–
<Qboid>
U+2014 EM DASH (—)
<Qboid>
U+2013 EN DASH (–)
<egg>
!wpn whitequark
* Qboid
gives whitequark a transitive Kusanagi
<egg>
!wpn -add:wpn fennec
<Qboid>
egg: Weapon added!
<iximeow>
!wpn the fennec protecting egg
* Qboid
gives the fennec protecting egg a falcon
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Ping timeout: 182 seconds]
APlayer has joined #kspacademia
UmbralRaptop has quit [Ping timeout: 202 seconds]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
<bofh>
egg: huh
egg|phone|egg has quit [Read error: Connection reset by peer]
<bofh>
also yes to the french-latin dictionary, duh
egg|cell|egg has quit [Ping timeout: 202 seconds]
egg|phone|egg has joined #kspacademia
Profound has quit [Read error: Connection reset by peer]
<egg|work|egg>
bofh: well it's rather heavy
<egg|work|egg>
bofh: the question is essentially whether I'm more likely to be writing/reading Latin in the office or at home
egg|work|egg has quit [Quit: webchat.esper.net]
egg|work|egg has joined #kspacademia
UmbralRaptop has joined #kspacademia
<egg|work|egg>
!wpn Iskierka
* Qboid
gives Iskierka a secure potato
awang has joined #kspacademia
UmbralRaptop has quit [Ping timeout: 202 seconds]
awang has quit [Ping timeout: 198 seconds]
awang has joined #kspacademia
<egg|work|egg>
bofh: oh, ねこ is shortened from ねこま "originally a compound of にゃ (nya, “onomatopoeia for the sound a cat makes (compare English mew, meow)”) + こま (koma, “four-legged animal”)."
<egg|work|egg>
so ja is like zh and many others in naming cats onomatopeically
<bofh>
I mean arguably that makes MORE SENSE
<bofh>
what is the etymology of "cat" (other than "latin")
<bofh>
"Late Latin cattus, catta (first attested in the 4th century, presumably with the introduction of domestic cats); ultimately origin obscure"
<egg|work|egg>
bofh: so the germanic/latin root is weird
<egg|work|egg>
bofh: of course various bits of the wiktionary have various opinions on the relationship between the protogermanic and latin roots https://en.wiktionary.org/wiki/k%C7%ABttr
<kmath>
<stephentyrone> @eggleroy @bofh453 @Tiffnixen @rygorous For the very-applied perspective, there’s a bunch of decent graphics resour… https://t.co/iQoUx6hY8D
<egg|work|egg>
bofh: the big timewaster afaict is generally not realizing that coordinates are a thing on top of the rotations, and so taking your rotations the wrong way or rotations in the wrong coordinates by not taking the right ones, rather than really anything about the quaternions
<bofh>
well, yes.
<bofh>
Also tbqh writing something might not be a bad idea!
<egg|work|egg>
bofh: well, yes, but time
<egg|work|egg>
bofh: but yeah, I don't think there's much written in the niche of "not having a prereq on linalg/diffgeo & not using axiomatic generalizations/n-dimensional things &c" and still being coordinate free
<bofh>
YEAH.
<bofh>
Hence my poking you three, since I feel like there *should be* but I can't think of any.
<egg|work|egg>
bofh: as an introduction to SO(3) and so(3) at the abstraction level of "3d vectors are arrows in 3d space" rather than "3d vectors are elements of a real vector space whose minimal generating sets have cardinality 3" or "3d vectors are triples of numbers"
<egg|work|egg>
s/as/as in/
<Qboid>
egg|work|egg meant to say: bofh: as in an introduction to SO(3) and so(3) at the abstraction level of "3d vectors are arrows in 3d space" rather than "3d vectors are elements of a real vector space whose minimal generating sets have cardinality 3" or "3d vectors are triples of numbers"
<bofh>
ahh yes, the linear algebra is about numbers in a box that you do things to approach
<bofh>
fuck that approach
<egg|work|egg>
bofh: and the axiomatic approach is nice, but also not conceptually helpful for reasoning about things in 3d space
<bofh>
yep
<egg|work|egg>
bofh: so yeah, maybe I should write a thing
<egg|work|egg>
bofh: but then again I should also write an akkadian input method
<Iskierka>
I would guess this kwoma might be related to 熊 / くま
<Iskierka>
and now I'm thinking of cats as "meowing bears"
<egg>
bofh: wait they're explicitly saying "matrices" and 4x4 matrices at that, not just 3x3, so in addition to the differential geometry you want representation theory to turn the Euclidean group into a subgroup of GL(4) https://twitter.com/Tiffnixen/status/1000930180301443072
<kmath>
<Tiffnixen> @bofh453 I mostly use matrices specifically in the realm of 3D geometry (so 3x3, 3x4, and 4x4 are pretty much the o… https://t.co/kmlwrzBSC9
<bofh>
egg: ehh, usually when you mean 4x4 you mean GL(4), yeah.
<bofh>
SnoopJeDi: LOL WAT
<SnoopJeDi>
LOL LIGO!
<egg>
bofh: no in this context we're not using the whole of GL(4), this is 3d geometry
<egg>
bofh: this is 3d affine maps as 4x4 matrices
<egg>
bofh: aka, a representation!
<egg>
bofh: so now basically my answer to this question is "take linear algebra, differential geometry, and representation theory"
<UmbralRaptop>
… debunking BICEP2? wat?
<SnoopJeDi>
BICEP2 wasn't a misunderstanding, you see. It was a deliberate lie, or..uh, something
<UmbralRaptop>
(I mean, their data got beaten into the dust with Planck)
<SnoopJeDi>
right, which is a more nuanced and uncomfortable reality than "they lied bc science is fraudulent"
<SnoopJeDi>
an organized conspiracy is way more soothing to the average mind than "reality is complicated?"
<SnoopJeDi>
file under: fake news
<UmbralRaptop>
Also, on of the BICEP2 people published a book on the situation and how they went wrong.
<UmbralRaptop>
… huh, this crank seems very concerned with scientists disproving god.
<UmbralRaptop>
Also Sean Carroll.
<SnoopJeDi>
yea
<bofh>
egg: I mean honestly I feel like *most people* who work with math should know AT LEAST A BIT of rep theory
<bofh>
and this includes both physicists and inorganic/pchem folks.
<egg>
yeah, but at least at ETHZ physicists do get exposed to it whether they want it or not
<SnoopJeDi>
bofh, at *least* the stuff that touches the Eightfold way!
<egg>
bofh: but people dealing with computer graphics/game engines, well
<egg>
you're lucky if they know vectors at a more abstract level than "three numbers"
<SnoopJeDi>
wow
<egg>
!wpn SnoopJeDi
* Qboid
gives SnoopJeDi a dead package
<egg>
!wpn bofh
* Qboid
gives bofh an expanding reactor
* SnoopJeDi
uploads it to npm
<UmbralRaptop>
egg: aaaaaaa
<egg>
UmbralRaptop: a?
<UmbralRaptop>
<@egg> you're lucky if they know vectors at a more abstract level than "three numbers"
<SnoopJeDi>
I've found all the engine-side people I've ever talked to are usually very clever, mathematically.
<SnoopJeDi>
Definitely less exposure to the headier math outside that core, though. AI people might know optimization and automata, procgen people know a pretty good amount of signals...
<SnoopJeDi>
and then there's John Carmack O_O
<bofh>
!wpn egg, SnoopJeDi
* Qboid
gives egg, SnoopJeDi a bugcheck
<SnoopJeDi>
!wpn egg, bofh
* Qboid
gives egg, bofh a Yarkovsky hyperboloid
<egg>
SnoopJeDi: I don't think I have seen any computer graphics APIs that allow you to deal with the underlying abstractions rather than the "buckets of numbers" view though (e.g. typed reference frames, distinguishing points and vectors, and the like)
<SnoopJeDi>
huh, you've never seen Point3D vs Vec3?
<SnoopJeDi>
(or whatever)
<SnoopJeDi>
Although granted, there's not much reason to abstract points as such in an engine. Just calling it `point` when you're treating it like one seems to be the most common pattern.
<SnoopJeDi>
most of my exposure to proper innards is idTech4, and John Carmack is just actually brilliant, so I'm possibly biased in this matter.
<egg>
well, yes, there is, they're different things, much like many programming languages will distinguish a date from a duration and disallow adding dates
<SnoopJeDi>
but like, slerp() is perfectly dull in terms of how common it is, and that's a pretty sophisticated mathematical idea!
<egg>
(it's a horrible name, too)
<SnoopJeDi>
granted
<SnoopJeDi>
egg, sure I see the reason an Undefined Someone might want to do such a thing. I dunno, I guess it wasn't a cycle expense worth paying once upon a time, and it kind of stuck?
<SnoopJeDi>
It's a strange history
<egg>
and it's not a particularly elaborate concept, it's complicated in implementation
<SnoopJeDi>
Well that's just moving the bar
<egg>
no, it's what I'm getting at
<SnoopJeDi>
On what is sufficiently complex to justify as being "real" math
<egg>
it's not the calculations where you screw up
<kmath>
<eggleroy> @stephentyrone @bofh453 @Tiffnixen @rygorous Obviously you need them, but most of the 3D time-wasting I have seen a… https://t.co/4V1U6FqPKT
<SnoopJeDi>
I just don't really agree with the assertion that engine programmers are mathematically braindead
<SnoopJeDi>
they might not have formal training, but they're very mathematical IME
<SnoopJeDi>
IM(limited)E
<egg>
we're not using the word mathematical in the same sense though
<SnoopJeDi>
that's kind of my point
<egg>
they don't provide things that reflect the abstractions, which is the point I'm getting at from the start of that twitter thread
<SnoopJeDi>
To what end?
<SnoopJeDi>
(in a games context)
<egg>
see that tweet, to the end of not wasting the time of the user of the API?
<egg>
it's not exactly fun to debug a mixup of coordinates systems
<SnoopJeDi>
which is, indeed, why it's absurdly common to use frames...
<SnoopJeDi>
well
<SnoopJeDi>
I guess it depends on what that means. I'll agree that the abstractions aren't usually there. The culture emphasizes "just doing it right" more often
<SnoopJeDi>
My view on it is that the tools that are the most useful are the ones that have come to exist
<SnoopJeDi>
i.e. fromEuler() exists because it's heckin' useful
<SnoopJeDi>
but there are two threads of conversation here and I was addressing a separate one altogether
<egg>
yes, none of this approaches the point, I am not saying that Euler angles are not a useful chart, nor Cardano angles
<SnoopJeDi>
then we are speaking past each other I suppose
<egg>
I am saying that most often the users of the API don't know what a chart is, and since the API doesn't help them with it, they shoot themselves in the foot
<egg>
I am talking about a problem of mathematical education, not of API capability
<bofh>
^
BPlayer has quit [Ping timeout: 182 seconds]
<bofh>
also wait what's wrong with slerp() as a name? it's linear interpolation along great circles, so it makes sense.
<egg>
bofh: mostly that lerp is a derpy name to start with :-p
<SnoopJeDi>
Perhaps that's why it stuck
<SnoopJeDi>
like `lusers`
<SnoopJeDi>
err, `luser`
<bofh>
egg: I mean that's the POINT
<egg>
bofh: yes but I find it far less legible than just spelling things out >_>
<bofh>
linear_interpolate()?
<bofh>
I mean I also agree, much like wtf does `dd' actually stand for acronym-wise? (yes I know it's "disk destroyer" tyvm)
<egg>
bofh: clearly interpolating polynomials should have a "derp" property, the Degree of the intERPolation,
<bofh>
ROFL
<egg>
bofh: so it seems that phl has made progress on the genetic algorithm, we might eventually get a solution
<SnoopJeDi>
which is, in fact, a defense (albeit a bad one)
<SnoopJeDi>
but my point was never that the APIs are mathematically good
<egg>
yes, that's my poin
<SnoopJeDi>
I agree wholeheartedly that there's low-hanging fruit mathematically
<egg>
bofh: is there any structure left with Euler angles in a vector, or does it reach the point of "three numbers"
<egg>
bofh: I guess the zero vector still gets mapped to the identity so there's that
<SnoopJeDi>
vecN tends to get used for any group of N numbers that are semantically related
<SnoopJeDi>
but yea, there's no structure there, I think
<SnoopJeDi>
I guess maybe it's a side-effect of DRY
<SnoopJeDi>
(lest you go further down the rabbit hole and do it "right" and implement things like rings)
<egg>
it's not quite true though, often a unit quaternion is distinguished from "other kinds of 4-tuples of numbers"; it's more that while there is a concept of rotation, there is no concept of vector
<egg>
(beyond tuple that is)
<SnoopJeDi>
oh sure, I didn't mean to imply it *always* happens that way
<SnoopJeDi>
I was thinking of color stored in a vec4, mostly, heh.
<egg>
!wpn bofh
* Qboid
gives bofh a ppc_f128 acid
<egg>
so, two acids but you have to react them at the same time?
<kmath>
<henryfhchan> @wtnelson Summary of IRG #50: ⏎ More than a dozen documents from individual contributors reviewed (including missed o… https://t.co/QMDuhGthoL
<bofh>
admittedly this reminds me of every time I've had to vacuum basis set elements for quantum chemistry out of a PDF
<bofh>
mostly consisting of a combo of pdftotext and paranoidally checking every single coefficient, which considering they're triple-zeta with polarization for row 2 transition metal atoms, means there's a goddamn ton of coefficients.
<UmbralRaptop>
Note to self: offer TeX files if someone asks you for a PDF.
<bofh>
this is more "offer supplemental data in plaintext ffs, particularly if it's something like coefficients to a numerical expansion of Gaussian orbitals"
<egg>
it's not that it's the transit observations
<egg>
bofh: I mean parsing plaintext records wouldn't be very different from parsing this copy-pasted pdf