« first day (4647 days earlier)      last day (669 days later) » 
00:00 - 20:0020:00 - 00:00

00:22
In category theory, I've noticed that it seems to be "safe" to take a category theory that doesn't have all products, and add product objects wherever they don't exist. What I mean by "safe" is that the original category is preserved as a full subcategory of the new category.
Does anyone know of any articles that discuss the idea of "completing" a category in this way, by adding product objects (or other kinds of constructions) wherever they don't exist already?
Let $\Omega_\epsilon$ denote the uniform solid region described in spherical coordinates by $0 \leq \rho \leq a$ and $0 \leq \phi \leq \epsilon$. Find the center of mass of $\Omega_\epsilon$

I set it up and all, but I'm not sure if this is all you wanted from it?

$vol(\Omega_\epsilon) = \int_0^2\pi \int_0^\epsilon \int_0^a \rho^2 \sin(\phi)d\rho d\phi d \theta = \frac{2\pi a^3}{3}\int_0^\epsilon \sin(\phi) d\phi$

$\int_{\Omega_\epsilon} \mathbf{x} dV = \frac{2\pi a^3}{3}\int_0^\epsilon \mathbf{x} \sin(\phi) d\phi$
When you get a chance @TedShifrin
Well, what is the center of mass?
Well all the constants cancel out and I'm left with the expression with the integrals
First, by symmetry, what is the only interesting coordinate?
I would say $z$ coordinate because $\phi$ will affect that the most
01:21
You can do better. What are $\bar x$ and $\bar y$?
hmmm....
Look at a picture?
in the process of doing such
well if I drew the correct sort of picture which is a sphere but it will be affected by whatever happens with $\phi$, my $\bar{x}$ and $\bar{y}$ will both be $0$.
Yes, but can’t you describe the region we’re looking at in everyday terms?
I would've just called it a ball, globe, sphere.....
01:34
No!
Where is $\epsilon$?
$\epsilon$ is the rotation over the $z$ - axis. So the latitude
or I mean longitude
I mean longitude
No, latitude.
So what real-world object are we talking about?
The Globe
Come on.
01:41
Clueless.
Longitude goes from north to south and latitude goes east to west. ...
oh were you talking about the limting position?
🎶Here comes Santa CLause, Here comes Santa Clause right down Santa Clause lane...🎶
01:57
@PlaceReporter99 snazzy
@MissMae Why is the website so openly available? I wasn't aware it was a crime
02:31
@CottonHeadedNinnymuggins there's so much here to unpack in why it's openly available even if it's giving unauthorized content (domain country the site servers are in, site could already be seized and is allowing traffic to monitor who access what with what IP, IP changes per shutdown, and etc), but nonetheless the original discussion was professors helping students get access to books, and by giving sites like those to students it would be equivalent to a professor speedrunning how fast they could get fired
02:51
@MissMae You also said students could have astronomical punishment. I've never heard of any punishment for something like this.
Lol, astronomical as in a really large punishment, not like they're gonna rain down meteors or something
whoever does the IT best practices training at your school is doing a very, very good job
don't copy that floppy
Lol...I was about to comment on that.
you wouldn't steal a hearse, drive it to the funeral, flip it over and dance on the opened coffin
why would you pirate a textbook
heard some teachers tell their students to avoid those websites because the books are free there
giving them the addresses without explicitly telling them it's okay
cool teachers
@leslietownes For starters because prices are ridiculous and most textbooks are a waste.
Let’s all get fired for ….
@shintuku Yea I have had profs do that.......they are fully aware of the "sites" or rather I should say "The Site", most understand that textbooks are priced very high and are sympathetic to that.
But it is still illegal even if not enforced
03:04
you get emails like: you are explicitly forbidden from downloading the classes' required textbook at (website)
Yes, I've had a lot of professors provide pdfs or other free textbooks
Sometimes not though
roger sir will not visit that website, sir
@CottonHeadedNinnymuggins providing them is different to downloading them from the website you linked above, which you should delete by the way
that
that's the spirit
03:07
lol
@onepotatotwopotato That probably is related, yeah! Those posts are a bit above my level, so I'll have to spend some time studying them.
@leslietownes I work the IT department at my university, and used to assistant teach some CS ethics courses, the people who had discussions with students who were caught were us (which we investigated only if it was obvious, or a hard DMCA letter was received), and it was never great and it always sucked. They usually got kicked out of the university after the university passed their info on to the DMCA lawyers
for the record i have never downloaded a pirated pdf
let the record pls record that statement
@MissMae Ooooh this is interesting... do these students get caught because they are using the school's internet connections to do their downloading and so they get tracked? Never was aware of publishers sending DMCAs out just because of the cost-to-reward from it
i don't know anything about the circumstances at your school. i agree that it is a bad practice to do anything resembling actively distributing unlicensed material on a campus network. that is a different ballpark from students using a campus network to download unlicensed material that they may have found whatever way.
you wouldn't issue a DMCA letter against the latter activity, anyway, because there is no ongoing offending activity to stop. the offense has already happened by the time it is detected.
03:17
AH...that makes sense
remember kids, if you're gonna break the law limit it to one law at the time
@D.C.theIII It’s via the network, but almost all DMCA requests are automated, as a lot of domains that give out materials are seized and the systems track what IP access what - lawyers make a lot of money out of whoever responds to the request. Campuses are required to investigate it tho, and it’s much easier to track because students have to use their school ID to access the network
ah sorry, wrong ping
Correct ping actually
automatic suing
i fixed it afterwards lol, RIP the other person who got pinged and didn’t know why
@TedShifrin tangential to the whole discussion, but i feel like it may interest you, anon merged all his accounts (artic tern, whacka, blue, runway44) so he’s still active
he doesn’t like the idea of people upvoting or rating his answers based on merit or score, or be distracted by anything other than math, ie wants to be judge to the same degree as a new user every time, but it was cool watching it all merge
03:29
dirty communist
there's something kind of quaint in the idea that people would upvote/rate based on rep. once upon a time, i did get the impression that people did that. i don't get that feeling anymore.
@MissMae Well, tell him I miss seeing him!
I personally tend to upvote people with lower rep.
03:51
and at least to hear copper tell it, people with lower rep tend to snipe on you folks.
at least, that's what i do. it makes me feel big.
some people just have too much rep. it's sinister and worrying
why are they stockpiling it? what are they planing to do with it? mischief, i say
@leslietownes we’ll it’s also the fact that some low rep users sometimes feel intimidated by high rep users, and if there is an error in his answer, there may exist a perceived artificial wall between the low rep user and him that may intimidate them from pointing it out (under the assumption that he knows more and it may not be an error). He wants to invite that discussion and remove any perceived barrier and wants to show that he’s always on the same level
@TedShifrin will do!!!
that makes sense. i could see people deferring to rep instead of asking follow-up questions.
04:09
I get a lot of people asking questions about my answers. Perhaps some do defer to rep, but some don't.
I wouldn't know who defers to rep, the same way I don't know why I get most downvotes since hardly anyone explains downvotes.
How can the Banach Fixed Point theorem be used to achieve Fractal Image compression?
04:27
I think graduate students try to pick their specialty too early
Year 2 or 3 is too early?
In my college, graduate students should choose their specialty in the first semester
That’s unheard of for mathematics. Not so unusual in experimental science where people commit to grant funding.
yeah, wow. if a math department is doing that, i hope they make it easy to switch specialties.
Anyway, a common side-effect of this is that students deny some courses that don't fit their specialty.
04:34
There’s also no good excuse for that. Breadth of knowledge is important for teaching and research.
i can only think of two people my entering year who had anything resembling an idea of what they wanted to do research in, and both were way off. in terms of what they ended up doing.
there was maybe a little bit of one-sidedness in advising on coursework, in that people were always told to take the basic algebra sequence because it might be good for them, and sometimes other algebra-adjacent courses for that reason, while nobody was ever told to take the basic analysis sequence, let alone any other analysis class, for cultural reasons. but that wasn't about specialization, really, and i don't think faculty opinion had that much effect on what courses people actually took.
Weird.
I made my one PhD student take functional analysis :)
no ducks at the pool today. phbhbhbt.
04:48
Too cold?
nah, they're usually out when it's a little overcast. that was the weird thing, actually.
yesterday there was a coyote about and it had all of the ducks upset, that might have disrupted the usual routine.
Ah, Munchkin’s friend, the vile coyote
she identifies with other agents of chaos.
Ready for politics.
i was actually in the office today for some reason, so when i picked her up from school i looked different. she said: "you found your work pants?" i'm not sure she has memories of seeing me in work clothing.
05:05
Glad you didn’t go to work in pajamas.
well, i went to the office in my good pajamas.
Is there a universe I understand what perfectoid space is?
05:20
No clue.
I don't know what that is, but I certainly like its name.
05:36
I still don't understand what's good about hyperelliptic riemann surface
06:13
What’s good or bad? Why judge? They’re the branched double covers of $\Bbb P^1$, which means they have a certain type of linear systems.
I mean I lack motivation on that surface. Something good is happening on that surface or relatively easy to study? etc. Isn't that what people ask all the time?
I know it's the branched double cover of $\Bbb P^1$ and each branch points are exactly Weierstrass points.
So what is your motivation for non-hyperelliptic ones?
What is your motivation for canonical curves?
Just seems like a stupid question when we’re trying to understand different curves/Riemann surfaces.
How to find the winding number of $\gamma(t) =e^{4it }+e^{it}+1 $ , $t\in[0, 2\pi]$ about the origin?
Umm... nothing. btw I'm not studying RS in AG language (I'm planning to do that later on)
@TedShifrin Hmm I see.
@robjohn sir
06:42
quit: the shortest route may depend on your definition of 'winding number,' but the winding number is given by an integral that you can evaluate fairly explicitly in this case
are y0 and y0' related?
if the above theorem holds and g(t)=0 then y=0 is a solution and should be the only solution?
since the solution is unique
because then how can they claim here that z(t)=0
07:03
@QuitMSE yes?
Sir, how to find the winding number of $\gamma(t) =e^{4it }+e^{it}+1 $ , $t\in[0, 2\pi]$ about the origin?
I am interested in finding number of zeros of $f(z) =z^4+z+1$ inside the unit disc.
Use Rouche's theorem then
I can't apply Rouche's theorem as I am not able to find dominating function.
So,my strategy is to use Argument principle directly.
@robjohn
$f$ has no poles( even no singularity, as entire), so number of zeros = \frac{ 1 }{2\pi i } \int_{|z|=1} \frac{f'(z) {f(z) } dz$
Change of variable yields, number of zeros inside |z|<1 is equal to the winding number of f($\gamma(t)) =e^{4it }+e^{it}+1 $ , $t\in[0, 2\pi]$ about the origin.
07:27
so evaluate that and see that it's 2
or, the roots occur in complex conjugate pairs, call them a, a', b, b', and the polynomial factors as (z-a)(z-a')(z-b)(z-b'), equating constant terms you see |a|^2 |b|^2 = 1 which implies |a| = 1/|b|. so if you somehow know that the polynomial has no roots on the unit circle, one pair of roots is inside the unit disc and the other pair of roots is outside the unit disc, 2 again
or, plot the thing (ok, you have to do slightly more than plot the thing to see how it's oriented) and compute from the picture tinyurl.com/22u6vcsy
(still 2)
I need to learn mathematica at some point. I'm just putting it off
08:07
Oh so the branch points of a branch covering space are points that are not identified after the quotient
ParametricPlot[{Re[#], Im[#]} &[z^4 + z + 1 /. z -> Exp[I t]], {t, 0, 2Pi}] gives
08:27
From the image it's clear that winding number about 0 is 2.
@leslietownes How to evaluate?
numerically integrate it? it's an integer so it almost doesn't matter what algorithm you use
Since it's a logarithmic derivative, I have to choose a branch.
well, i'm telling you not to try doing it symbolically
i gave a symbolic, theory, whatever suggestion above too, so i don't know why you're focusing on this if you don't like alternate approaches
in general, if you don't know what some number is, but you know it's an integer and you have an integral formula for it, you can get your answer by numerical integration
and then look for less numeric ways later
the beauty of numerical integration is you don't need to branch anything, you just need to be able to numerically evaluate the thing you're integrating
which you can (because you have a symbolic formula for it)
this isn't a situation where a numerical method in an off the shelf computer or calculator is going to spit out 5.4563067026 and you'll have to guess at what the 'real' answer might be or what happens if you do more digits or choose a better method
it's going to be very close to exactly one integer
for the reason you mention, i'd actually be worried about asking a computer to evaluate it symbolically
 
2 hours later…
10:54
Hm, was this generally bad answer or bad path to e? math.stackexchange.com/a/4685552/1091751 Sometimes you indeed wish there was an option for anonymous comments so downvotes were less puzzling.
11:26
0
Q: $A = \sum_{i=1}^{\infty} \frac{i}{a_i}$ and $1 = \sum_{i=1}^{\infty} \frac{1}{a_i}$?

mickI have the following problem : For a given positive real $A>1$ $$A = \sum_{i=1}^{\infty} \frac{i}{a_i}$$ $$1 = \sum_{i=1}^{\infty} \frac{1}{a_i}$$ $$a_{n+1}>a_{n}$$ where $a_i$ is a strictly increasing sequence of strict positive integers. How to find $a_i$ for a given closed form $A$, in particu...

 
1 hour later…
12:27
Given a Riemannian manifold $(M,g)$ with the Riemannian volume form (in coordinates) $\sqrt{|g|}dx^1\wedge...\wedge dx^n$ hypersurface. Given a hypersurface $S$ with induced metric (pullback) $\gamma$, can I conclude that the induced volume form is (in coordinates) $\sqrt{|\gamma|}du^1\wedge...\wedge du^{n-1}$?
12:56
how can we assume this highlighted equation
I think that you will find much better success if you don't make people deal with an image in order to understand text. For example, I am not going to try to parse your notes. You can MathJax in chat: tinyurl.com/cfqcvpc .
Let $f$ be continuous on $[a,b]$, i.e. integrable in the sense that for any $\epsilon>0$, there exists a partition $P$ of $[a,b]$ such that $U(f,P)-L(f,P)<\epsilon$. Define $K=\mathrm{max} f$ on $[a,b]$. Now, let $\epsilon$ be given and suppose there is a $P$ with $N$ points satisfying the previous inequality. Let $P_n$ be another partition of $[a,b]$ with mesh $\delta$. Define the refinement $Q=P\cup P_n$.
Then according to this answer, since $Q$ is obtained by adding $N$ points to $P_n$, we have $$U(f,P_n)-U(f,Q)<2NK\delta .$$ I would be very grateful if someone could explain this inequality. I currently do not see how this is derived.
13:13
0
Q: Show that the differential equation $\frac{dy}{dx}=\sqrt y,y=0$ has non-unique solution.

FranklinShow that the differential equation $\frac{dy}{dx}=\sqrt y,y(0)=0$ has non-unique solution. To solve this problem I used Lipschitz theorem. My solution goes like this: Let $f$ be continuous in an unbounded domain $$D:a\leq x\leq b,-\infty\leq y\leq \infty.$$ Let $f$ satisfy a Lipschitz condition...

Can anyone please help me with this?
 
1 hour later…
14:33
Today's work suddenly brought up a differential equation problem and I was embarrassed that I couldn't recall a single trick from that class years back. At least there's something residual that makes looking at the textbook a breeze.
For those that want to join along if they want a puzzle:
$y'+\frac{1-2x}{4x^2+x}y=\frac{1}{4x^2+x}; y(0)=0$
have you tried getting the integrating factor $e^{\int\frac{1-2x}{4x^2+x}\ dx}$
that's the usual way to solve first order linear equations
15:09
Yeah that's what I'm doing right now, I just now have to deal with ugly absolute value integrals.
Can anyone please explain how to determine the order of a differential equation in which we have more than one dependent variable say, $\frac{d^2y}{dx^2}+\frac{d^{2}z}{dx}+\frac{d^3p}{dx^3}=f(x)$ ?
But the proof behind indefinite integrals of absolute value feels mind-blowing which makes me happy. I like to be surprised by calculus tricks of substitution and perhaps there was a way I could have used the usual 'udv-vdu' business to find it, but this felt much more powerful.
https://math.stackexchange.com/a/1980123/524816
prime notation is only used for functions of variables, not variables themselves right
$y' = \frac{d}{dx}y$ for example
then what is $x = g(u) \to dx = g'(u)du$
like what is the operation
@Franklin I think order is still the highest differential, is it not?
Differentiation and then multiplying by $du$:
$x=g(u), \frac{d(g(u))}{du}=g'(u)$
$\frac{dx}{du}=\frac{d}{dx}\left[g(u)\right]$
$\frac{dx}{du}=\frac{d(g(u))}{du}$
$\frac{dx}{du}=g'(u)$
$dx=g'(u)du$
A bit shoddy on choosing a consistent notation on my part, but I hope that clears it up for you @Obliv
It's similar to
$y=f(x)$
$\frac{dy}{dx}=f'(x)$
$dy=f'(x)dx$
15:24
Trying to understand what the jacobian is
of an integral
I'm not grasping where it's coming from
Looks like a determinant and a total differential
If I want to change the variables of an integral, how would I "derive" the jacobian
I guess the steps
It's the lovechild of linear algebra and differentiation, I'm not in a strong position to give the why off-the-top of my head, I can delve into it if you want, but the steps are quite consistent:
If you have some $x=g(u,v)$ and $y=h(u,v)$, meaning you can express $x$ and $y$ in some planar parametrization, then little bells in your head should go off saying linear algebra has become all-powerful for this problem.

Think back to gradients, where we would be able to just do, for example, $\nabla g=<\frac{\partial g}{\partial x},\frac{\partial g}{\partial y}>$ this is a vector, but we have thi
column space of jacobian of integral gives you approximation of integral @Obliv
15:40
Pausing to make sure the $\LaTeX$ behaved,

So then our matrix would be
$\begin{bmatrix} \frac{\partial u}{\partial x} & \frac{\partial v}{\partial x} \\ \frac{\partial u}{\partial y} & \frac{\partial v}{\partial y}\end{bmatrix}$

And its determinant would be $\frac{\partial u}{\partial x}\frac{\partial v}{\partial y}-\frac{\partial v}{\partial x}\frac{\partial u}{\partial y}$ which should seem eerily familiar to something else you've covered.
Whoops wrote my fractions upside down, one moment
Ah the edit timer wore off, guess it's on my permanent record.
Anyways just flip all of those partials, I clearly need more coffee.
16:18
ohh
@A016090 Thank you, that was very helpful.
16:41
Looks like asking simple questions here is just alright? Is it correct that such function f(x+y)=f(x)∗f(y) is unique (continuous, real numbers only, etc)? Is there some incorrectness if you try to introduce e number via exp function, skipping all the limits, derivatives and series?
you need more conditions to get uniqueness there (i'm assuming for functions R to R)? e.g. f(x) = 0 and f(x) = a^x are solutions for any a > 0 and sadly there are also other solutions
it is tragic, even
if f satisfies f(x+y)=f(x)f(y) for all real x, y then with the exception of the identically zero function, f is positive-valued (e.g. because f(x) = f(x/2 + x/2) = f(x/2)^2 for any x), and g(x) = ln f(x) satisfies g(x+y) = g(x) + g(y) [the "cauchy functional equation"], and solutions to the cauchy functional equation give you solutions to your equation by exponentiating.
144
Q: Overview of basic facts about Cauchy functional equation

Martin SleziakThe Cauchy functional equation asks about functions $f \colon \mathbb R \to \mathbb R$ such that $$f(x+y)=f(x)+f(y).$$ It is a very well-known functional equation, which appears in various areas of mathematics ranging from exercises in freshman classes to constructing useful counterexamples for s...

@Obliv Review your linear algebra and understand why you get the area of a parallelogram in $\Bbb R^2$ by taking the determinant of the matrix whose columns are the vectors.
@leslietownes f(0+1) = f(0)f(1) seems means automatically pointless soultions or f(1) = 0
16:47
the guess that the only functions f (other than the zero function) satisfying f(x+y)=f(x)f(y) are the exponential functions f(x) = a^x for some a > 0 is equivalent to the guess that the only solutions g to the cauchy functional equation are linear functions. this guess becomes real if you impose more conditions, e.g. continuity
halt: well, either f(0) = 0 (in which case f = 0) or f(0) = 1 (which is indeed forced the rest of the time, but does not determine what f is)
some of the answers in the survey page above are phrased in terms of your equation above, which it calls "cauchy's exponential functional equation"
17:00
@leslietownes $f(x)=n^x$ satisfies that property for all $n$.
@leslietownes of course, misplaced 0 and 1. Thanks for details, will read/recall everything a bit more carefully. Oh finally, at least I have best guess for downvotes. But these details seem so unimportant when you just try to have simplest path to e number! Like simplest path to Pi is divide length to radius. You skip all the details and explanations during this introduction, you don't prove that L/R is constant. Seems so natural to have same path to E instead of Limit.
17:45
When typing up a question on the main site it's not rendering immediately
Takes a good 5 full seconds to render
and then when I type more math I have to constantly manually render
It used to just render automatically
@geocalc33 Does immediately for me on short (and long too) answer, not after 5 sec
18:00
@halt9k my chrome is up to date, internet connection is fast, and this was not a problem in months prior. the lag in rendering only happens when I'm typing up a question
@geocalc33 could be something specific in answer markdown, something on your PC, or specific communication between server (firewall, etc). Post test answer example?
@halt9k What do you mean post test answer example?
@geocalc33 Err, question. What exact question text you typing to try it too?
ohh it renders immediately when trying to answer a question
weird
It's same for me both on asking and answering, at least on random examples. Less than 1 sec to update. Browser is Firefox
I'd guess it's some markdown complication, try to divide question on two parts and to try them separately.
18:36
Ok, I think a plausible route forward for incommensurability is a certain realization I just had for extending the idea of numerical bases to the reals via the floored logarithm.
18:46
Proposed rigorous definition: Two reals $x$ and $y$ are incommensurable iff $x$ or $y$ have finite representations in the base of the other real. We can define real-valued bases rigorously as a sum $\sum_n^{-k} R^n = a$ where the LHS is qualifiedly equal to the RHS depending on whether it is a limiting sum or finite sum.
Each term $R^n$ can be computed given reals $a$ and $R$ by successively computing $a(n+1) = a(n) - R^{\lfloor\log_R a(n)\rfloor}$.
with $a(0) = a$
Can't really say it has a certain measure of each term per say as $[0, R)$ then. In this case, the coefficient for any given term is 0 or 1.
19:04
We can think of the Fourier transform as a canonical isomorphism $L^2(K) \to L^2(\hat K)$ where $K$ is a locally compact abelian group and $\hat K$ is its Pontryagin dual. What happens when you trade $K$ for an abelian semigroup?
hello, if $f $ is continuous on a compact set from $\mathbb{R}^N$, is $\nabla f$ bounnded ?
@Semiclassical I don't understand this obsession with CGPT
19:20
There's nothing too understand. People are bored and stuck online "after" the pandemic lockdown.
It's a sign of the times.
yeah, idgi
the impact of it as a technology is interesting enough: how it affects student submissions of work, what sectors it can further automate, etc
but treating chatGPT as some kind of all-knowing entity is duuuuumb
The question looks like some sort of marketing scheme that should be moved to a chatroom.
it's bizarre, yes
hooray, it's been deleted!
19:29
🙏
i wish there was a good way to say: "this is a well-defined question for which i have no hope that there's a nice answer"
@Semiclassical You don't think that numerical answers are "nice"?
if you can get them, yes
@Vrouvrou you'll need more hypotheses than 'continuous' just for nabla f to be defined, and maybe some care in defining nabla f at the boundary of the compact set. as you see e.g. with N = 1 and f(x) = sqrt(x) on [0,1], to say nothing of examples where you have continuous functions on compact sets that are nowhere differentiable. is there a more precise version of this question?
the only nice point in the question i linked is that it's a spherically-symmetric normal distribution
19:34
+1 👍
@Semiclassical I mean, anything continuous is approximable---just approximate the integral using a Riemann sum on a grid. Convergence is going to be slow as f---, but that is a method which we know is going to work.
And there are better quadrature methods out there.
true enough
Not a terribly satisfying answer, but likely good enough for, say, engineering. :D
in the context of interest I'd just say "do Monte Carlo"
which is to say, view it like a probability problem and accept that's all you can do
problem is that's not a very helpful answer when it has that many free parameters
@Semiclassical Yeah, that's probably the right answer in this case.
19:42
@leslie Pretty soon Munchkin will be playing with ChatGPT ducks.
Let $x\in (0,1)$ be any number and let $x=0,x_1x_2x_3...$ be the decimal expansion of $x$. Then define $D(x)=\lim_n \frax{x_1+...+x_n}{n}$. I want to find $x$ s.t. $D(x)=\frac{1}{2023}$. But I don't see how. Can someone help me?
nobody answer that question with 2023 in it. it's a contest problem. change it entirely by changing it to 2054 so you'll have to wait until then to use it in a contest.
How do we flag a contest problem in chat?
user: try finding examples that would solve the problem for smaller values than 2023. try 2, 3, 4, 5 for example.
@leslietownes it is not a contest problem, I have seen this in dynamical system but we also can take 2022 instead
19:48
@user123234 the only intelligent thing i can note is that, if your number never has 0 in its decimal expansion, then each average will be at least 1. so your number had better have a looooot of zeros
indeed, i suspect you can approach this using only 0 and 1
Dynamical system(s)?
@TedShifrin yes
i was kidding. the joke is that, often contest problems will use the year of the contest for "an arbitrary number that is large enough so that brute force doesn't immediately come to mind, or solve the problem."
@leslie It probably can't be a contest problem since I immediately saw how to do it.
wouldn't entirely shock me, i've seen symbolic dynamical systems stuff
tho i can never remember how to reference it
19:49
@TedShifrin could you give me a hint then?
so if you see a problem on the internet with e.g. 1979 as a random parameter, it likely appeared in at least one contest in 1979, and we're just dealing with that ever since.
@user123234 I think leslie's suggestion is the standard and correct suggestion to use whenever you have problems like this.
ted: i agree, it isn't much of a contest problem if changing the contest-year parameter to 2, 3, 4, 5, more or less immediately suggests a solution.
but if you think it's bad now, wait until it's 2054.
as in, D(x) = 1/2 etc?
19:50
oh, i get it
@TedShifrin the problem is that I'm confused with this limit.
@user123234 The only hint I'll give besides leslie's excellent suggestion is that you should think about rational numbers with repeating decimals.
Why are you confused?
i'll also encourage keeping it simple and only use the digits 0,1
Now we're giving way too many hints.
eh, i'm just being consistent with what i said at first
19:52
I just stood in line over an hour to get a few things as Bed Bath and Beyond closes down forever. Very sad. I've been shopping there for at least 30 years (I think).
I should have thought to have ChatGPT stand in line for me.
so assume I want to find a number $x$ s.t. $D(x)=1/2$ this would mean that $\lim_n \{x_1+x_2+...x_n}{n}=1/2$ I thought about what if $x_i=1$ if $i\in 2\Bbb{N}$ and $0$ else. But then how do I deal with the limit so how do I prove it rigorously
Wow! Closes 360 stores.
Look at an obvious subsequence, @user123234 and then think about how to prove the entire sequence has that limit.
Another pandemic business casualty.
The shift to online shopping, I guess.
@TedShifrin so I mean $(\sum_{k=1}^2n x_k)/(2n)=1/2$
19:59
what i wonder about is how fast the convergence will be
00:00 - 20:0020:00 - 00:00

« first day (4647 days earlier)      last day (669 days later) »