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>
<✔PA> #Breaking Professor Stephen Hawking has died at the age of 76, a spokesman for his family said
<iximeow>
speaking of (de)obfuscators, whitequark you still use binary ninja?
<rqou>
egg: random question: is there a simple example of a normed vector space that isn't a banach space?
<rqou>
(btw, midterm tomorrow in case you can't tell :P )
<bofh>
rqou: well a Banach space implies normed & *complete*, so uh ℚ^n under the Euclidean norm? :P
<rqou>
right, dug
<rqou>
*duh
<whitequark>
iximeow: define use
<iximeow>
whitequark: look at programs with it, even casually. i suspect i must be missing a lot of what they're doing because it's not exposed through the UI
<iximeow>
whitequark: it's roughly 50 times slower than i expect and at this point i'm profiling to figure out why, but i haven't looked closely at the API docs yet
<whitequark>
occasionally, yes
<iximeow>
are you aware of anything more intensive than building SSA that they might be doing?
<iximeow>
nothing i've seen suggests dead code elimination or *much* constant propagation.. nothing that seems worse than doable-in-linear-time anyway
<whitequark>
not yet
<whitequark>
i mean, no, they aren't doing even dce
<whitequark>
and whatever constant propagation there is is incidental
<kmath>
<henryfhchan> @Informoji @BabelStone @Random_Guy_32 @FakeUnicode Finally, half of the characters in Ext. B could be axed had the… https://t.co/kfvhY9cmxa
<egg>
cc rqou et al.
<rqou>
not relevant for midterm :P
<rqou>
oh what
<rqou>
CJK extension _G_?!
<rqou>
egg: what's IDS/IVD?
<egg>
IDS?
<Qboid>
egg: [IDS] => Index catalogue of visual Double Stars
<rqou>
i'm still amazed that a white guy actually went and actually properly understood the history and/or potential geopolitical issues rather than making something up and hoping it works
<egg>
Ken Lunde or Andrew West (NBR for GB before ISO/IEC JTC1/SC2 are worth following on those questions
<rqou>
NBR?
<egg>
(Andrew West did the tangut stuff)
<egg>
!acr -add:NBR National Body Representative
<Qboid>
egg: I added the explanation for this acronym.
<awang>
Of course *now* Google returns relevant results
<awang>
Looks like a VS bug
SilverFoxy has joined #kspacademia
<SilverFoxy>
whitequark, just saw your message about obfuscation
<SilverFoxy>
so
<SilverFoxy>
Ifuckin love ternaries
<SilverFoxy>
they can compact code and make it smaller and fits in one line and that turns me on
<SilverFoxy>
I like that kind of syntactic sugar shit
<awang>
SilverFoxy: Learn Haskell then
<awang>
Or APL
<awang>
Or J
<awang>
Or K
<SilverFoxy>
naw naw naw
<SilverFoxy>
C# is my shit
<SilverFoxy>
I think I need more object-oriented stuff for the human body than functional stuff
<SilverFoxy>
although I could reduce certain things down to functional levels, but that would take a mindshift
<SilverFoxy>
would be neat, but not until I get the design down in the way I am most familiar
<SilverFoxy>
really the stuff I like to see is things where it's like "remember how you have to do all this shit to do X?" "Yeah?" "Well fuck that you can do it much more efficiently like Y"
<SilverFoxy>
also, in Haskell everything is a function right?
<SilverFoxy>
so one thing I want to do, for shits and giggles, is to use some weird coding techniques, and maybe some bad ones, without fundamentally coding terribly
<SilverFoxy>
I'm pretty not good at coding, so I'm just embracing it at this point
<SilverFoxy>
basically my goal is to code in a way that makes good programmers feel disgusting and dirty but keeps them intrigued enough to keep reading on. Where it's not bad enough, but it's functional and efficient and does its job correctly
<awang>
SilverFoxy: Look up the IOCCC
<SilverFoxy>
IOCCC?
<SilverFoxy>
qboid?
<SilverFoxy>
add it
<awang>
!acr -add:IOCCC International Obfuscated C Code Contest
<Qboid>
awang: I added the explanation for this acronym.
<SilverFoxy>
ehhh
<SilverFoxy>
I'll save that for later, can't dedicate huge amounts of time to sit down and do things right now
<awang>
If you're trying something new, a why not go for something *really* new?
<awang>
Also, the "much more efficiently" is really complicated these day
<awang>
s/$/s
<Qboid>
awang meant to say: Also, the "much more efficiently" is really complicated these days
<SilverFoxy>
yeah I know but yall know what I mean
<awang>
Cache speeds start becoming more important than big-O times, in some cases
<SilverFoxy>
that's getting past the point i'm talking about
<awang>
I don't think everything is a function in Haskell? You can probably model programming that way, but idk if that's how Haskell does it
<SilverFoxy>
idk enough about haskell to start thinking about how I would write the functions of the human body in it
<SilverFoxy>
also, I need a UI to show the human body physically, dunno if haskell does that stuff?
<SilverFoxy>
the best way to start getting into multiple languages for this project would be to write a backend and frontend and have a middleend communicate, I forget the term for a thing you have that communicates between front-end and back-end and it basically translates or whatever?
<SilverFoxy>
an API?
<awang>
MVC?
<awang>
Haskell definitely does UI
<awang>
idk how though
<SilverFoxy>
i just know how to do unity stuff with the editor and make things so that's why I'm with unity
<SilverFoxy>
UE4 or 5(if that exists) would be cool because of the graphical capabilities I've seen from it
<SilverFoxy>
although the human body, even its insides, suffer from that thing in animation where the more realistic you get, the better, until you hit a plateau of almost realism, and it makes a large trough because it gets super creepy
<SilverFoxy>
until you hit full realism
<SilverFoxy>
but yeah, if I were to delve into haskell, I'd do the backend body functions in it and just make a communication interface so it can relay the info to the body model
<SilverFoxy>
However, right now I am stuck trying to wrap my head around how to program the heart in a good way such that I don't hardcode it too much, but in a way that I can manipulate the function of each chamber individually as per necessary, and yeah
<SilverFoxy>
because it's not just a serial process, there are things happening simulataneously in the heart
egg|phone|egg has joined #kspacademia
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
SilverFoxy has quit [Ping timeout: 383 seconds]
egg|cell|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
<bofh>
whitequark: I need to either get ~6.4V up to 7.5V or ~9.6V down to 7.5V (retrofitting something to use LiFePO4s). which of those two would you recommend (I'm leaning boost 6.4V) and what switching topology would you recommend?
<bofh>
(peak current load would be ~450mA, closer to 350mA in reality, so this probably is relatively simple).
SilverFoxy has joined #kspacademia
awang has quit [Read error: Connection reset by peer]
<SilverFoxy>
but the real question is whether that is simultaneous-dependent or simultaneous-independent
awang has joined #kspacademia
egg|cell|egg has joined #kspacademia
egg|phone|egg has quit [Read error: Connection reset by peer]
egg|phone|egg has joined #kspacademia
SilverFoxy has quit [Ping timeout: 383 seconds]
egg|cell|egg has quit [Ping timeout: 198 seconds]
<egg>
!wpn bofh
* Qboid
gives bofh a Shannon cuddle
<SnoopJeDi>
That sounds like a very underwhelming cuddle.
<SnoopJeDi>
Also consumes more digits than it produces for reasonable truncations
<SnoopJeDi>
Although that said I like the continued fractions as an example of "whoa math has these strange patterns"
tawny has joined #kspacademia
<bofh>
Like it's a very inefficient continued fraction representation in terms of how rapidly its truncated convergents actually converge to the limit.
<bofh>
But the efficient ones don't look ~*~ superficially magic ~*~ sooooooooooooooooo
<bofh>
:P
<SnoopJeDi>
yea there's some fudge room in the convergence rate I suppose, but I still wonder about the "digit consumption" factor :P
<SnoopJeDi>
just having a piss at the expense of maths
<bofh>
Yeah I know :P
<SnoopJeDi>
Clearly Buffon's Needle is the best way to suss out π
<SnoopJeDi>
(I'm actually quite fond of Monte Carlo estimations for how good they work for outreach)
<kmath>
<bofh453> My stance on the whole τ vs. 2π debate is that the true analyst just scales their Lebesgue measure by the inverse o… https://t.co/eajmc35Qc6
<kmath>
<stephentyrone> @Gankro People who are serious about floats are very mad.
<egg>
bofh: tbh I would love to have a constexpr sqrt
<egg>
also rootn probably
<egg>
at least cbrt
<egg>
bofh: don't see a sane way to have constexpr trig portably without requiring correctly-rounded though
<egg>
also we should lure atlas in there
<bofh>
I too want constexpr sqrt/cbrt, but past that the utility drops off very rapidly imho.
<bofh>
Also yes.
<bofh>
We *really* should.
<SnoopJeDi>
Which one of those things is the cause and which is the effect
<egg>
bofh: honestly I'd like the standard to be honest about the definedness of floating-point expression, because UB division by 0 is silly
<egg>
(it was silly and harmless before constexpr, it's silly and obnoxious now)
<bofh>
UB div by zero is fucking idiotic
<whitequark>
bofh: in general I prefer buck
<whitequark>
that's because you have lower current through the cells
<whitequark>
but at 350mA it doesn't really matter
<whitequark>
dropping 2.1V... hm
<awang>
egg: Implement compile-time IEEE 754 so you get control over rounding modes :P
<awang>
Template floating point sounds... fun
<egg>
well principia is full of constexpr arithmetic
<egg>
(but some of it is borked by clang's overzealous interpretation of UB in floating-point arithmetic, hence the CONSTEXPR_INFINITY macro)
<egg>
(I should get back to richard smith about that)
<whitequark>
awang: rust is considering compile-time ieee754
<whitequark>
also, llvm already has it
<awang>
What platforms don't use IEEE 754 these days?
<awang>
Didn't know LLVM already had it
<egg>
awang: I mean, sure, constexpr floating point is very much a thing
<egg>
it's just that the language and clang's interpretation thereof introduce weird limitations
<awang>
whitequark: Does LLVM actually use it, or is it just exposed for front ends?
<awang>
egg: That's true
<awang>
Have you asked on the mailing list why constexpr division by zero is a fatal error?
<egg>
I've asked Richard Smith by mail
<egg>
we've discussed that a bit
<egg>
(my question was actually 1 * infinity, and that's hardly clearly UB by the standard)
<egg>
1/0 is explicitly UB and so the compiler has to reject it (but iirc they want to discuss changing that in the standard)
<egg>
!wpn котя
* Qboid
gives котя a submersible lobster
<egg>
!wpn whitequark
* Qboid
gives whitequark a flammable Eva
<egg>
!wpn awang
* Qboid
gives awang a double
<whitequark>
awang: yes, it does constant propagation
<awang>
I can't wait for the day where I can email people like that without feeling bad
<awang>
egg: I mean, UB is UB, so maybe the compiler can not reject it?
<awang>
Unless UB during compile-time results in ill-formed programs
<awang>
whitequark: Interesting. I'll need to look at that sometime
<egg>
awang: you can't have UB in a constant expression
<egg>
at runtime it was harmless so pre C++11 FP division by 0 was "your behaviour is defined in another international standard"
<awang>
Ah, OK
<bofh>
sigh.
<egg>
OH(phl, translated): we can use the release to compute the date of easter
<egg>
!wpn bofh
* Qboid
gives bofh a tritium zone
<bofh>
egg: LOL PLEASE DO SO
<bofh>
!wpn egg
* Qboid
gives egg a mangled Sachertorte
<rqou>
egg, bofh: what is "the point" of the poincare-bendixson and bendixson-dulac theorems? when are these useful?
<bofh>
rqou: uh I guess you have to care about if your system of ODEs in the plane has periodic orbits or not? like it allows you to determine conclusively when it does *not*, but I don't know when that's actually useful for anything.
<rqou>
lol ok
<rqou>
usually i deal with planar systems by looking at the phase portrait and using human brainpower
<bofh>
I mean same, plus whether it has periodic orbits or not is usually not at all a useful result is more my issue.
APlayer has joined #kspacademia
<SnoopJeDi>
oh cool, @inconvergent has their stuff on GitHub
<kmath>
<bofh453> OH: "I never thought I’d have any cause to speak fondly of Hu Jintao, but at least he had the decency to remain the… https://t.co/O54kzEbaaz