« first day (2562 days earlier)      last day (2755 days later) » 

00:00
i don't know any examples off the top of my head, but you can write numbers such as 12^12 - 1
(which is divisible by 11)
and 5
the idea i meant was big^bigger - small
fair enough
and this can be of the form $p_1\cdotsp_n$ for n large primes
Why would that be a problem in ultrafinitism?
00:02
because that prime is "too large"
i'm not doing a great job of explaining
important question: is ultrafinitism willing/able to make the statement
I mean, if you were to write $12^{12} - 1$ in tallies
that the set of natural numbers is not finite
wait, it doesn't allow that set to exist
but does it make any sort of equivalent-ish statement
Then you still can say there's a prime smaller than it. Unless you're not gonna allow some operations to even exist
@eyeballfrog, idk i don't know any in person, just what i've read from blogs, i thought about giving their side of the argument a fair chance a few years ago
00:05
it seems like it makes any sort of useful mathematics impossible
@Daminark, idk, maybe the blog i was reading from was a very biassed source
but then I'm one of those heathen physicists who wants math to do something
biased*
@eyeballfrog GET THE HERETIC!
Lol jk
shh, i won't tell anyone, you could get away with it
00:06
But yeah I mean idk, I have straight up philosophical problems with ultrafinitism
same
the problem I have is that I don't actually understand what problem it's trying to fix is
Mostly with $\mathbb{N}$, like unless you're gonna say "This is the last number", which is absurd, I don't see any reasonable way to dodge saying that it's infinite
And there's no real problem with countable sets
the main problem i have with it, i think, is that there is no good textbook giving the construction of everything, (whereas i know the construction (i think) with normal ZFC)
so the problem for some is with axiom of choice and the problem for some is infinity
Stuff like $\mathbb{R}$, there's def some philosophical sketchiness
00:08
yes, I do accept that R seems to be a little bit sketchy
Though honestly I'm fine with it, and hell I roll with the generalized continuum hypothesis
i thought it was fairly complet
philosophically
and then i saw $\mathbb{C}$
but physics seems to work well enough with it
(well, these days we use C, but still(
00:09
it's because it's more complete, it's algebraically and sequentially closed
If push comes to shove, you can still say that you're working in an axiomatically defined model and make no claims regarding its ontology
it's just the better place
well C has some of its own weirdness
this is wondering more into philosophical talk, not concrete or well defined
well, going to somewhat more concreteness
00:11
Well, I mean the whole debate about ultrafinitism seems to be a philosophical one, yeah?
I'll admit the continuum hypothesis sort of bothers me
seems like one of those things we should be able to know
shrug
After Gödel, continuum hypothesis becomes alright for me
hmm, fair point
A set of axioms which contain a model for arithmetic cannot define a single consistent model
They are either inconsistent or define a class of mutually inconsistent models
be nice to have the structure of R well-established
it's kind of important
00:17
So then ZFC allows both models to make sense
I actually believe in the generalized version, that $2^{\aleph_n} = \aleph_{n+1}$
(well unless you're an ultrafinitist, I guess)
Hey @Semi!
Lol, yeah I'll pass on that, as I said I just take the axioms of math to be formal anyway, no need to make any claims on their having to do with reality/metaphysics in the worst case, and in the best case I can't fathom the nonexistence of infinite sets, at least in thought
But yeah enough of my rambling
Lel
it's nice to take a healthy interest in set theory and axioms, you could ask for reference-request s about good books on them. well worth a read if you are truly interested
00:21
It's all a matter of mathematical taste, of course.
I'm the kind of person who finds special functions way more interesting than foundations, though, so I'm probably a heretic myself.
(Or, even worse, a physicist :P )
the name special functions always sounded odd to me
since the vast majority are just "We defined these to be the solution to X differential equation. Good news, everyone! We solved X differential equation."
I guess the idea is special-purpose functions?
Okay so I never understood them too well. Like, I've heard of them typically labeled under "mathematical physics"
a lot of it comes down to solving boundary value problems via separation of variables
But I've also seen people count the $\Gamma$ and $\zeta$ functions
Which are of number theoretic importance. Maybe they're useful in physics too but still
In physics Gamma and Zeta mostly come up in various integrals
00:26
"which are of number theoretic importance" it's hard to find a function which isn't
for instance, integrals like $\int_0^\infty \frac{x^k}{e^{x}\pm 1}\,dx$ come up whenever you do computations in statistical physics
and those, when you can express them in closed form, are just the Riemann zeta function
I see
how come sometimes it shows your points and sometimes it doesn't
in the chat
usually it's when someone does a few messages in a row
test
00:37
as an example...
no
help
i just want to spam
to test this out
Huh? Shows your points?
did you know... there's a spam filter
i'm 668, eye is 2828
and semi is 10.?k
00:38
rep, yeah
10.7k
Oh weird, I just noticed the points
Wao
on one hand, i want more rep, but on the other, i don't want to do homework questions and the ones i do want to, are too hard
grrr
Where is Mr Eyeglasses? I tried to email him but no response for months already. I hope he is still alive.
Yeah, I've also found it hard to get more conceptual (as opposed to, this is long and annoying someone do it for me?) questions which are at my level to solve
00:42
3
Q: Reflective subcategories of topological vector spaces

Stefan PerkoLet $\mathsf{TopVect}$ be the category of TVS over $\mathbb K\in \{\mathbb R, \mathbb C\}$ with continuous linear maps as morphisms. Do the normed spaces or locally convex spaces form a reflective subcategory of $\mathsf{TopVect}$? (in the latter case: do normed spaces form a reflective s...

Does anybody have a nice slogan what it means for a subcategory to be reflective?
00:59
does unique factorization hold in $\mathbb{Z}[\sqrt[3]2]$?
01:50
@Sophie is that the square root of 3 times 2? If so then let me think for a bit. I know for a fact that its sibling from the square root of 3 is a Euclidean domain. I might be able to piggyback the proof.
no. It's the cube root of 2. There are instructions on how to get the LaTeX working on the sidebar
@Sophie why are you telling me how to turn on latex? I think I should I know that by now. I've been here for years.
sorry if I don't like that stuff running. Plus, I don't run JavaScript on my phone.
oh okay
:p
i know sqrt 2 is integrally closed in one sense
(that term is still kind of weird for me. I've heard it different ways.)
it is a Euclidean domain
I've never looked into the cubic rings. Still examining the quadratic rings. Sorry.
thanks
01:55
No problem
Im heading out
Cya
 
1 hour later…
03:18
Hi.
04:07
Let $f : A \subset \mathbb{R}^2 \to \mathbb{R}$ with continuous partial derivatives at the point $P = (a,b)$. Then there are $\theta_1, \theta_2 ~ \leq 1$ such that, given $\epsilon > 0$ and the point $P' = (h,k)$ with$ P' \in B(P,\epsilon)$, then $f$ holds that $f(a+h,b+k) - f(a,b) = h f_1(a+\theta_1 h, b+k) + k f_2(a,b + \theta_2 k)$
I don't see why we need the continuity of the partial derivatives. Shouldn't be enough the existence of them?
Time to get this thing coded:
No gram schmidt needed. Though I need to check carefully how the numerical error will pile up. If the error is much larger than the one using modified gram schmidt, use modified gram schmidt...
04:22
I can define $\phi : \mathbb{R} \to \mathbb{R}$ such that $\phi(t) = f(t,b+k)$. $\phi(t)$ is differentiable in $a$ since $\phi'(a) = \displaystyle \lim_{h\to 0} \dfrac{\phi(a+h) - \phi(a)}{h} = \displaystyle \lim_{h\to 0} \dfrac{f(a+h,b+k) - f(a,b+k)}{h} = f_1 $ i.e: the partial derivative of $f$ w.r.t the first variable which I said it exists, so $\phi$ is continuous and then I can apply the mean value theorem.
Doing the same with $\phi(t) = f(a,t)$ I can finish the proof without the condition of continuity... Am I wrong in somewhere or is not needed continuity of the first partial derivatives?
Has anyone proved yet that an infinite number of integers satisfy the collatz conjecture?
Oh, to apply the mean value theorem to $\phi$ in $[a,a+h]$, $\phi$ must be differentiable in $(a,a+h)$, that is why I need continuity of the first partial derivative in a neighboorhood of $(a,b)$.
it seems like only a finite number do so
@TedShifrin Think of it this way. What is 1/<0,1>? It's <0,1>. Hence, you can take the reciprocal of a vector. Not that weird really...
04:38
wat
what does 1/(0,1) mean in your sentence?
@anon are you familar with the bayesian interpretation of probability?
I'm not a philosopher. what about it?
I am not sure if bayesian inference can be understood as soem kind of algorithm that updates the probability distribution recursively, like so:
Jul 31 at 12:41, by Secret
in The h Bar, 11 mins ago, by Secret
Meanwhile Bayesianist don't know or don't have the probability distribution of all events they are interested in. They first assign a prior probaility to some given question, and then each trial of an experiment or other incoming evidences will serve to update the probability distribution, so if all the evidences and experiments are not crappy and of high quality, then eventually the probability distribution will converge towards that given by the frequentists
I was trying to compare between frequentist vs bayesian to figure out why bayesian is more practically used
in The h Bar, Jul 31 at 12:26, by Secret
IIRC, frequentist knows the probability distributions of all events they are interested in, and that repeated experiments should corresponds to said probability distributions
(NB everything needed for my question is already quoted in these two paragraphs)
04:47
can't say I'm interested in the discussion
@anon I wasn't talking to you. The person I was speaking to understands. Also, I did not say "1/(0,1)". Learn to read.
okay. what does 1/<0,1> mean? is <0,1> not a vector? Ted Shifrin, the person you're talking to, has a message on the starboard claiming not to understand what the reciprocal of a vector is, and I was assuming you were responding to that.
@Secret I know why. I can give an example of why that is more useful.
if there's something you think I haven't read, feel free to point it out.
...
04:51
@Typhon ok
@Secret routing protocols in delay tolerant networks.
basically imagine routers moving around and trying to send data to each other
what is the probability that the next one a router encounters is some particular router?
if it's something like satellites it is trivial.
people with cellphones as routers?
you're not going to know that probability for real
period
ever
you'd be trying to find the probability of two people meeting
what is 2/<2, 3> tho?
@Dair ...
Think of it this way. What is 2/<2,3>? It's <5,8>. Hence, you can divide scalars by vectors.
@anon I. Wasn't. Talking. To. You.
:-)
04:55
there is perhaps backstory in the trascript I am missing. geometric algebras or something.
@Typhon you are now :-)
or perhaps it's a joke. shrug
@anon Seems legit
rolls eyes
@Typhon uh so how will frequentist and bayesianist differ in terms of determining the router probability?
@Secret hahahaha
fequentist is literally impossible to use.
tell me
put two randomly moving objects in a field
what is the probability they will meet in x seconds?
now put 20 in that field
what is the probability that the next guy I meet is that router?
ok I see, so how does bayesianist get around this problem?
04:59
who said it does?
the probability is probably not accurate
all that matters is that it gets the data through efficiently while preventing flooding to all routers.
(if it were then that would imply that the government can predict who you are going to meet in a day with absolute precision)
note: that includes people you walk past in the street
@Secret simply put. Frequentist assumes a formula is known and exists. We cannot write or engineer one with current statistical knowledge. Bayesianist attempts to use iterative approximating algorithms based on past knowledge. It might be that neither works, but at least we can bank on the latter working in such a way as to get data from point a to point b.
in fact, I would argue that under random motion two routers with the same physical properties have equal probabilities of meeting some other router.
in other words: probability is useless for DTN routing unless that probability is based on some kind of pre-planned non-random motion.
at that point, it boils down to the best way of predicting human motion.
if we could do that, then that would probably lead to crime being ended
(and a general tyranny)
so... we accept that the probability doesn't work. It just somehow gets messages to where they need to go.
I see, makes sense
oh, you're TheGreatDuck...
i remember you
@Secret I ran a lot of mass tests during that thing I had over the summer. The protocol that did the best by a land side was "MaxProp". I'd suggest reading about it. It's pretty complex and I can see why it did so well.
@Dair Nope. I'm Typhon.
Nah, I'm pretty sure you're a duck
(I won't acknowledge a username that has no proper spacing)
@Dair The other guy isn't around anymore. I got rid of him.
That was intended to sound creepy. It's an inside joke.
05:13
well you are a deadly snake thingy, so idk man. Maybe you did kill him
In numerical analysis, the Kahan summation algorithm (also known as compensated summation) significantly reduces the numerical error in the total obtained by adding a sequence of finite precision floating point numbers, compared to the obvious approach. This is done by keeping a separate running compensation (a variable to accumulate small errors). In particular, simply summing n numbers in sequence has a worst-case error that grows proportional to n, and a root mean square error that grows as n {\displaystyle ...
@Dair That's not a snake... that's just a corkscrew shaped stone thingy.
The nonassociativity of floating point arithmetic makes it hard to understand how to optimise those errors
and I doubt a 3d model can kill a deity.
well you either mistyped typhoon, or you are Typhon the deadly snake from greek mythology...
05:16
@Dair Correction: Typhon the giant living hurricane from greek mythology...
@Dair there is a lot more to it then you might think. The joke is lost on most people simply because it is obscure as ****.
heck
I might not even be able to find a link anymore
> function KahanSum(input)
var sum = 0.0
var c = 0.0 // A running compensation for lost low-order bits.
for i = 1 to input.length do
var y = input[i] - c // So far, so good: c is zero.
var t = sum + y // Alas, sum is big, y small, so low-order digits of y are lost.
c = (t - sum) - y // (t - sum) cancels the high-order part of y; subtracting y recovers negative (low part of y)
sum = t // Algebraically, c should always be zero. Beware overly-aggressive optimizing compilers!
@Secret read up on maxProp. idk those terms you used. It might be that it uses the frequentist methodology. idk.
uh, that block of code is kahan summation, not probability stuff
This one, I think might be what you are referring to:
Routing in delay-tolerant networking concerns itself with the ability to transport, or route, data from a source to a destination, which is a fundamental ability all communication networks must have. Delay- and disruption-tolerant networks (DTNs) are characterized by their lack of connectivity, resulting in a lack of instantaneous end-to-end paths. In these challenging environments, popular ad hoc routing protocols such as AODV and DSR fail to establish routes. This is due to these protocols trying to first establish a complete route and then, after the route has been established, forward the actual...
no
i mean it is
but read the actual academia
XD
(that page is sh**)
"Beware overly-aggressive optimizing compilers!" rip -O3
05:24
@Dair don't laugh at me. This goes for everyone. It is from my high school days. It always will have a special place to me. After all... I wrote most of it.
so... you're saying you started a cult?
no dear god no
hahahaha
it's basically an inside joke from an old forum that went quite far tbh
It looks like you were trying the best your highschool self could to start a cult lol.
basically there was a game called "you must not say duck"
like word associated
but where nobody is allowed to say duck
the goal was to go the longest without duck
the opposite occurred.
I meant to say "this thing is spam"
instead I wrote
"this thing is a cult"
next thing you know the resident dorkface decided to write a short story about it
and then the rest of us lemmings followed suit
XD
(if anything the "cult" is the name of the story, not the name of any real life thing)
@Dair newbie. (sorry but like literally everyone has said that at one point or another)
i was going to say I've been around here quite some time, idk why you are calling me a newbie.
05:30
true
but if you were on the other site that it originally started on you'd be a newbie
and plus they'd all say you were horribly misunderstanding it
"Grammar when Referring to Duck"
shrugs
I would of found it funnier if you required that the middle letters were censored.
remember this is a forum filled with teenagers with nothing better to do.
that would be more of a teenager thing to do lol.
than just capitalizing the word lmao
05:31
I primarily focused on writing actual chapters
I kind of told everyone we should take the writing more seriously
not in theme mind you
but like... quit self-contradicting itself.
"The Pickle is always male. However." Classic teenager.
tbh the story ends up with them both being male
(like, confirmed as such)
I doubt you're going to read it so I'll just spoil it for ya.
Basically the entire universe is a computer simulation in the duck's pc.
(note: he's a human)
the pickle is also a human
"Duck's" plz.
and the only reason they have that weird situation
is cause the AI is retarded
and took the name Drake as "male duck"
so let me get this straight... it started as a joke and you told everyone to take the writing seriously?
05:35
and to make it worse the other guy was sarcastic and called himself a pickle.
@Dair yeah. Why not?
XD
I mean. I probably dumped 30 pages of writing into that thing back then.
$((sum_0 + (a_1 - c_0)) - sum_0) - (a_1 - c_0) := c_1 \text{\\\c is negative .bbb of y}$

$sum_1 := sum_0 + (a_1 - c_0) \text{\\\value of $sum_0 + a_1$. Check value of c afterwards to for signs of over aggresive optimizers} $

Therefore, $c_i$ accumulates the low value part of each input and add them back in
@Dair I take everything seriously. Stuff can be fun but that doesn't mean you cannot make a decent effort towards doing your best. (Or at least don't do stuff to wreck other things.)
@Dair tbh, I probably never transferred it all. I should get around to doing that some day.
like... the one page with all the chapters doesn't include the 50 page google doc I was writing a couple summer back. Decided to do writing regarding the older time periods. Latter stuff was getting too weird. At least back then there were no ducks and no pickles. Well... or at least there wasn't a misunderstanding about stuff.
XD
Just regular Greek mythology.
with... programming
and disasters
and time travel
ugh
@Dair anyway, Typhon ended up being a more practical villain in the end.
well... good luck with that lol. idk what to say...
oh it's dead
like D.E.A.D.
I mean technically that game. The one in my avatar. It's technically set in the past but it's really just obscure references here and there. I doubt anyone would ever pick up on them.
@Dair if you can get past the bad writing style it actually isn't as bad of a read as it sounds. I would argue that the point of it is that it isn't meant to be taken seriously yet written in a serious manner. That's the punchline. It's a joke told so seriously that you cannot help but laugh at it.
(well and I would say that some of the lacklusterness of it came from me after a while just getting burnt out on writing the darn thing. There's parts I never bothered to write. I just wanted to finish.)
06:04
@Balarka how does one geometry?
Like it seems that my brain can't process shit
I had hard time with Ted's diffgeo exercises.
@Daminark Which ones are you doing in particular
Right now I'm on this twist of a unit normal field
Though I haven't yet put much time in, I was stuck on this other one
Maybe this'll be aight
Oh wait a sec
Ok. Let me know if you want a hand or anything.
Yeah you can commute things wrt the cross product
Maybe that'll help
Or I mean, combo of cross and inner
Okay wait so
We can write $tw(C,X) = \frac{1}{2\pi} \int_0^L \langle T(s),X(s)\times X'(s)\rangle ds$
So then $|tw(C,X) - tw(C,X^*)| = \frac{1}{2\pi}|\int_0^L \langle T(s), X(s)\times X'(s) - X^*(s) \times (X^*)'(s)\rangle ds|$
06:30
What are you trying to do
Trying to prove the difference is an integer
My guess is that we might be able to say the sign of that shouldn't change
So maybe we can pull in the absolute value
I don't get notation
$C$ is a curve?
$X$ is a vector field along it I guess
Oh, so starting from scratch here
$C$ is a curve
$T$ is tangent field of $C$
$X$ is a normal field on $C$
06:33
ah
Unit normal field, in fact
Oh oh ...
hides
lol
hi @Ted
Hey @Ted!
hi Balarka and Demonark
06:34
But yeah if we can pull in the absolute value here I think we might be able to pull some fanciness wrt the whole, magnitudes of cross and inner products
I'm wondering if that sounds at all reasonable?
So there's $X$ and there's $X^*$.
Hmm ...
what does $tw(C, X)$ really mean
"Twist" of the field
Ah, makes sense
06:36
not with the $X$ in there
BTW, this stuff shows up in the explanation of DNA/RNA replication
I kept thinking I should have written a page appendix on that, but I still haven't done it.
so, Demonark, who are $X$ and $X^*$?
They're two unit normal fields
So what does that tell us?
So for like $X = N$, $tw(C, N)$ is integral of the curvature or something isn't it
Not quite, Balarka.
Curvature measures how $N$ is twisting toward $T$.
ah, how about $X = B$
06:42
At any point $\alpha(s)$ they're both orthogonal to $T$, so you could put them on a plane
How 'bout it?
OK, Demonark. And they're both unit vectors. So what do you conclude?
[Chemistry] In the past 3 hours, I have been massive googling and SE searching things like "robust taking square roots", "robust adding numbers", "robust vector normalisation", "robust matrix and quaternion multiplication" etc.
Well, you can place them both on a copy of the unit circle, which you can sorta vary to get a kinda tube on which they lie
Oh maybe it's something about the angle between them being constant?
Well, OK ... Why should the angle between them be constant?
But, now that you mention it, there's an angle between them. Hmmm ...
My reasoning was that the angle between them being constant felt like something that would be nifty, it'll prob be something about their derivatives though
06:48
They're two arbitrary normal fields. Why should the angle be constant?
Actually I'm no longer of the belief that this angle needs to be constant
At least the picture I had in my mind no longer seems tenable
I'm imagining them curving away from each other somehow
You're assiduously avoiding the obvious thing to do.
Oh differentiating their inner product?
I don't see how that comes in naturally in the formulas you're comparing.
Hmm, what else can detect their angle?
Oh wait cross product
06:57
No. This is a standard game which you'd be more used to if you'd done section 2.
$X$ and $T\times X$ give an orthonormal basis for the normal plane. Now what?

« first day (2562 days earlier)      last day (2755 days later) »