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.
<egg|zzz|egg>
!u b̂ʹ
<Qboid>
U+0062 LATIN SMALL LETTER B (b)
<Qboid>
U+0302 COMBINING CIRCUMFLEX ACCENT (◌̂)
<Qboid>
U+02B9 MODIFIER LETTER PRIME (ʹ)
<bofh>
I am slightly curious where you are finding all these circumflexed letters in various unicode sequence form :p
e_14159 has joined #kspacademia
e_14159_ has quit [Ping timeout: 202 seconds]
<egg|zzz|egg>
bofh: not so much finding as producing,
<egg|zzz|egg>
bofh: well, Estrin = dumb monomial for degree 2
* UmbralRaptop
chirps at iximeow
<iximeow>
chirping diapsids
<egg|zzz|egg>
bofh: I had taken the habit of hornering all of the things, so I had stuff like q̂[k].value + h * (c[i] * v̂[k].value + h * Σj_a_ij_g_jk)
<egg|zzz|egg>
bofh: which I just changed to q̂[k].value + h * c[i] * v̂[k].value + h² * Σj_a_ij_g_jk
<egg|zzz|egg>
which is also how it generally appears in papers, so that's nicer to read anyway
<bofh>
I mean I still have that habit, tho I guess for a degree 2 polynomial you don't save much by Hornering it.
<bofh>
I *do* honestly find the Hornered forms more readable tho, from years of numerics.
<egg|zzz|egg>
bofh: well that's the thing, nowadays you pretty much always lose by hornering :-p
<egg|zzz|egg>
(unless you have several polynomials being hornered at the same time)
<egg|zzz|egg>
hornering is latency-bound on every operation, it's awful
<bofh>
yes but just evaluating in naive polynomial form I *believe* can be more ill-conditioned numerically for certain input/poly coefficient sets.
<bofh>
assuming high-order polynomials ofc, but we've already established those are good on current h/w, so,
<egg|zzz|egg>
bofh: yeah, but it usually either is in the realm where it doesn't matter (here an integrator), or where you do the error analysis carefully for the eggsact evaluation you end up doing (e.g. that cbrt)
<bofh>
okay, point.
<bofh>
wait, why does it not matter in an integrator? b/c it'll get smoothed out on subsequent iterations, or?
<egg|zzz|egg>
because you're not doing a careful error analysis anyway :-p
<egg|zzz|egg>
and if the errors from a second-degree polynomial evaluation are a problem you are not out of the woods, regardless of its condition
<egg|zzz|egg>
there are things that matter, e.g. accumulating the state in double precision, or in multistep methods with highly alternating coefficients, doing the whole combination in double precision
<egg|zzz|egg>
but the RKN q̂[k].value + h * c[i] * v̂[k].value + h² * Σj_a_ij_g_jk doesn't
<egg|zzz|egg>
aaaaaaaaaaaaaa Fine uses hat for the low order method, and Dormand for the high-order method
<egg|zzz|egg>
WHYYYYYY
<bofh>
okay, true, but for a second-degree polynomial *in general* it won't matter unless you are *deeply* in the woods.
<kmath>
<mrkgrnao> @rygorous @stephentyrone on fire: doing this for e and pi by subdividing beats into finer and finer pieces and doin… https://t.co/0seqMirkz6
<bofh>
ααααα
<bofh>
-ααααα*
<UmbralRaptor>
/me blinks
<egg|zzz|egg>
<bofh> okay, true, but for a second-degree polynomial *in general* it won't matter unless you are *deeply* in the woods. <<< yeah, it's just that there's no point in Hornering it either, so might as well go back to writing it as you would on paper :-p
<bofh>
fair enough, I agree. :p
<egg|zzz|egg>
bofh: found a spanish thesis on RKN methods
<egg|zzz|egg>
bofh: the examining committee is called "Tribunal" which sounds a bit ominous
<kmath>
<profanegeometry> *sobs* you can't just refer to every physical dimension like cross sections.. ⏎ me: *at art supply store* 1.96 zettabarn pencil lead please
awang has joined #kspacademia
<egg|cell|egg>
!wpn -add:adj tricyclic
<Qboid>
egg|cell|egg: Adjective added!
<egg|cell|egg>
!Wpn whitequark
* Qboid
gives whitequark an invertible handle
awang has quit [Ping timeout: 183 seconds]
<egg|cell|egg>
!Wpn UmbralRaptor
* Qboid
gives UmbralRaptor a negative fire extinguisher
<UmbralRaptor>
!wpn egg|science|egg
* Qboid
gives egg|science|egg a starboard mount
<UmbralRaptor>
!wpn egg|nature|egg
* Qboid
gives egg|nature|egg an expander clowder
<UmbralRaptor>
… one of those
awang has joined #kspacademia
egg|zz|egg has joined #kspacademia
<egg>
UmbralRaptor: egg|a&a|egg
* egg
meows at bofh
<egg>
bofh: do you have an eggsample of a nice 2nd order differential equation where the right hand side depends on the derivative and which has a known solution?
<bofh>
egg: any differential equation which has as a solution an orthogonal polynomial family representable by a 3-term recurrence?
<bofh>
I'd do Legendre's Eq'n personslly but Chebyshev's also works & you might be more familiar w/it.
egg|zz|egg has quit [Ping timeout: 180 seconds]
egg|zz|egg has joined #kspacademia
* UmbralRaptor
pokes transit with a car.
<UmbralRaptor>
Looks like driving to Udvar-Hazy takes ~1/3 to 1/9 as long as buses+subways. Having to go through Dulles probably doesn't help.
<egg>
bofh: so I can do both, but clearly I *have* to add another identifier with Чебышёв in it
<egg>
γιτ χηεχκουτ
<egg>
um.
<UmbralRaptor>
Are you going to submit a pull request on a Byzantine emperor?
<egg>
UmbralRaptor: so, did IAU 2018 resolution B4 pass?
<egg>
"All five Resolutions were accepted by the General Assembly, although Resolution B4 — the proposition to rename the Hubble Law as the Hubble-Lemaître law — will, in addition, be voted on electronically by all IAU members in the fourth quarter of 2018, in order to ensure wider participation in the decision. Prior to the General Assembly, Resolution B4 was also submitted to a period of extensive consultation with
<egg>
"that, as from 1 January 2019, the fundamental realization of the International Celestial Reference System (ICRS) shall be the Third Realization of the International Celestial Reference Frame (ICRF3), as constructed by the IAU Working Group on the Third Realization of the International Celestial Reference Frame"
<egg>
yay
<bofh>
Nice!
<egg>
!u ”
<Qboid>
U+201D RIGHT DOUBLE QUOTATION MARK (”)
<egg>
stabbity
<egg>
!wpn whitequark, коте, and the котяchrome kitten
* Qboid
gives whitequark, коте, and the котяchrome kitten a colorless dysprosium nib
tawny has joined #kspacademia
<UmbralRaptor>
!wpn egg
* Qboid
gives egg an Algol constructor
<UmbralRaptor>
!wpn tawny
* Qboid
gives tawny a basilect
<tawny>
oh thank you, I've been looking for one of those
<kmath>
<ManishEarth> @arjache 猫 is one of my favorite hanzi because if you break down the phonosemantic compound the character literally… https://t.co/3HXIUUvpeb
<rqou>
yes, that is correct
<egg>
rqou: sadly it seems hard to make a pun between 折反射 and 猫
awang has joined #kspacademia
<egg>
bofh: whitequark: today in transliterations, 馬克蘇托夫 (Максутов)
<bofh>
Augh.
<bofh>
Also eagerly waiting the first document that tries to translate that as if it were a sentence.
<egg>
bofh: Чебышёв is 切比雪夫 apparently
<rqou>
oh yeah RU->CN is amazingly "fun"
<rqou>
i should ask my father/grandfather about those
<kmath>
<eggleroy> (T(s?j|[xšr])|(Ts?c|C)h|Č|Ç|C[sz]?)[eê]b[îiıyư](s[jz]|sch?|[sc]h|[šșşsx])[ëéeo](ff?|[vwp]) ⏎ ⏎ Consistency: there is a… https://t.co/aR9v3paUYM
<bofh>
(cf. Antonio Stella Bottom Tile)
<egg>
Especially Чебышёв
<rqou>
egg: i'm surprised there's consistently a b :P
<rqou>
no language managed to put something like v there instead?
<egg>
rqou: none that I found
<egg>
rqou: I looked for romanizations only, obviously 切比雪夫 is out of the competition
<egg>
it has a b in pinyin fwiw, not that this means much
* egg
stabs pinyin
* rqou
hands egg a zhuyin fuhao
<rqou>
now you can have a new set of symbols to learn _and_ touch geopolitical issues
<egg>
!u ㄅㄆㄇㄈ
<Qboid>
U+3105 BOPOMOFO LETTER B (ㄅ)
<Qboid>
U+3106 BOPOMOFO LETTER P (ㄆ)
<Qboid>
U+3107 BOPOMOFO LETTER M (ㄇ)
<Qboid>
U+3108 BOPOMOFO LETTER F (ㄈ)
<egg>
rqou: pretty sure you can't render anything CJKV without touching geopolitical issues :-p
<rqou>
yeah i suppose
<bofh>
what egg said :p
<egg>
honestly you can't do much to Latin without running into interesting things too
<egg>
something something capitalization of i and Turkish
<rqou>
<insert story of the murdered ex because of mangled turkish i>
<bofh>
Like I got one the day of my Khorsabad trip, since I realized it *very* rapidly paid off resp. day admissions.
<bofh>
(actually that's a thing I've consistently noticed; the price of a one-time X is absurd, but like a week or a year pass for X often isn't that much more, often only 3-10x that of the single-use).
<egg|nomz|egg>
bofh: oh, you got one already? nice
<egg|nomz|egg>
!wpn bofh, whitequark, et al.
* Qboid
gives bofh, whitequark, et al. a Cassegrain shattered conflict
<bofh>
!wpn egg|nomz|egg
* Qboid
gives egg|nomz|egg an extended kernel
egg|nomz|egg is now known as egg
<egg>
cannot convert argument 1 from 'const std::vector<...>' to 'std::allocator<...>' Ꙩ_ꙩ
<bofh>
UHM.
<egg>
!u tÙ_
<Qboid>
U+0074 LATIN SMALL LETTER T (t)
<Qboid>
U+00D9 LATIN CAPITAL LETTER U WITH GRAVE (Ù)
<egg>
bofh: they haven't figured out that they can talk about the documents referenced by the note in section 1 of ISO/IEC 10646:whichever like WG9 does
<egg>
(I think section 1 has become clause 2 in recent versions)
<bofh>
Oh, does ISO 10646 not actually refer back to Unicode WGs? Not sure I'm understanding that correctly.
<egg>
bofh: ISO 10646 does refer to Unicode, in the note in section 1 (Unicode doesn't have WG, the WG9 I'm talking about is ISO/IEC JTC 1/SC 22/WG 9). In general however you can't refer to a non-international standard from an international standard (10646 is magic and harmonized with Unicode), hence why Ada goes through that weird circumlocution (JP complained if Unicode was mentioned directly)
<bofh>
(ROFL, WAIT, WHY???)
<bofh>
Also I thought ISO10646 is just Unicode in ISO standard form?
<egg>
no, Unicode does a lot more crud
<egg>
ISO 10646 is just the encoding; everything from case transformations, general categories, to emoji BS is Unicode
<bofh>
Ahh.
<egg>
ISO and Unicode have to agree to *encode* something, but the Unicode properties are Unicode's only
<egg>
bofh: "you can't refer to a non-international standard from an international standard" << that part is fairly sensible, you can't start depending on arbitrary documents that are cooked without the transparency and openness requirements of ISO, otherwise the whole process is pointless
<egg>
bofh: and the JP NBR pointed that out
<egg>
now, 10646 is magic, in that the encoding is jointly standardized by ISO and the consortium, which means it does refer to Unicode (the standardization process for 10646 is extremely weird, so I guess something happened at the JTC 1 level to get that compromise)
<egg>
which means that there are the documents referenced by the note :-p
<bofh>
Wait a second, *Unicode* is less transparent/open than ISO some...how? Huh?
<egg>
bofh: good luck making sense of the ESC decisions :-p
<egg>
bofh: but the rule about non-international standards isn't about Unicode
<egg>
it's a general principle for international standards, you can't say "and the definitions are some obscure book published by the academy of sciences of st. petersburg in 1800"
<bofh>
Okay, point, but in that aspect I also find fault with "and the definitions are in this easily obtainable book that requires you to fork over 199CHF first". >_>
<egg>
bofh: yeah, that's annoying; otoh, the process matters too
<egg>
a random third party standard isn't voted on by NBRs
<bofh>
Point.
<egg>
so the interests of <insert country here> have not been considered in drafting it
<bofh>
Like, I don't really have any fault with ISO standards and understand both why they exist *and* why they'd be preferred in use in a context where international agreement is necessary.
<bofh>
It just irritates me to no end that most of them claim to be "open" yet are obscenely paywalled (for no good reason???).
<egg>
I suppose the ISO process takes resources, but yeah the paywall is a pain
<bofh>
I mean sure, but why are they dumped on the users of the standard rather than the implementors of the standard?
<bofh>
(So far I've had luck finding leaked copies every time I've needed one (which is more often than you'd expect, btw >_>) but those often took hours to days to find).
<egg>
bofh: re. the "open", it is about the standardization process (you can get decent minutes etc.) rather than the actual standard
<UmbralRaptor>
egg: AAAAAAΑΑΑAΆĀÅ
<bofh>
Yeah, hilariously you can get minutes and a draft standard and *usually* that's sufficient IME, which makes the paywall even stranger?
<egg>
bofh: I mean sure, but why are they dumped on the users of the standard rather than the implementors of the standard? << well, for a programming language, the implementors are the compiler writers, so, that's what happens :-p
<bofh>
Okay, I suppose.
<egg>
bofh: if you put the cost on the NBRs, you gate participation to the standardization process, which compromises the standard
<egg>
(i.e. you'll lose experts who can't afford it)
<bofh>
Okay, point, but I'dve expected that for any standard that's being heard at the ISO, the experts would be sponsored by some company or nationstate or whatnot.
<egg>
bofh: well, the experts often *are* national body representatives
<egg>
bofh: but that's, e.g., West
<egg>
(NBR GB before ISO/IEC JTC 1/SC 2)
<egg>
bofh: it's still actually eggstremely individual at the end of the day
<egg>
"This is sure to be very popular amongst fans of Lewis Carroll. It goes well with hedgehogs."
<egg>
"Fine. Encode."
<bofh>
ROFL
<bofh>
I mean, as justifications to encode an Emoji that's better than a *lot* that I've heard.
<egg>
bofh: see, the ESC's output might be freely available, but its process is a fucking mess
<egg>
"Well, “originally proposed as ICEBERG” shows just how disorganized the Emoji Subcommittee’s process is."
<bofh>
ROFL
<egg>
"A white bear is a bear which is white. A white bear is not a whiteness which is bearish."
<bofh>
wait, ICEBERG WAS ORIGINALLY PROPOSED AS ICE *CUBE*???
<bofh>
ROFL IT RANDOMLY GETS INTO HERALDIC PATTERNS AT THE END WHAT THE HELL
<egg>
bofh: it *is* Everson, if he can take advantage of this madness to get heraldry encoded he will :-p
<bofh>
rofl
<egg>
bofh: (there's some horrifying proposal to add colour swatches that somehow modify things, I don't want to know about the details; Everson points out that they have squares with heraldric hatching already, could add some more and use those instead of new codepoints)
<bofh>
Yeah that honestly seems far saner than colour swatches (also I'm increasingly annoyed that it's getting to the point where you *need* a colour font renderer to properly render things... at least for emoji just converting to grayscale works okay, but once we get into colour modifiers... *ugh*).
<egg>
bofh: eh, just wait it out :-p from the same document, "The UCS will outlast the emoji fad. Does anyone really believe that an Emoji Subcommittee will be beavering away 70 years from now?"
<kmath>
<ptrmsk> tired: me ⏎ wired: also me ⏎ ⏎ (i have an anxiety disorder)
<egg>
!wpn bofh
* Qboid
gives bofh a harmful map
<UmbralRaptor>
!seen ferram4
<Qboid>
UmbralRaptor: I last saw ferram4 on [28.08.2018 21:10:37] in #RO saying: "Right now I'm looking at an unfunny situation because the laptop I was previously using won't start. So I'd prefer to move the HDD from that to another machine rather than rebuild my workflow from scratch, still looking into what that will take"
<ferram4>
hmm?
<UmbralRaptor>
sorry, your mod came up in KSPO, and I realized that I hadn't seen you active in here in ages.
* egg
pokes ferram4 in the partial differential equations
<ferram4>
Yeah, I need to update it, but some PC issues mean my workflow is really disrupted right now