« first day (4559 days earlier)      last day (462 days later) » 

1:29 AM
@ペガサスSeiya If it's not true.
of course, here it would be in print, so maybe it would be libel instead.
 
We should just bid Seiya adieu: Seiya later!
 
2:23 AM
I see there's been a number of comments that have needed a rimshot after them..........
 
Are you complaining again?
 
Absorbing your thinking patterns accounts for all walks of life.
 
I’ve quit thinking.
 
Until I trigger you with a question I know the answer to, but didn't "think" as you put it so eloquently. 😊
 
I don’t see no eloquence.
 
2:32 AM
I could complain though about just recovering over another head cold-minor flu this wee after I just finished suffering one the week before Christmas.....I have not been thrilled about it....
most I've been sick in 2yrs......through the eye of the storm perfectly fine and now?...twice.....whatever. I'm feeling better finally.
 
I hope you’re better. I think COVID has weakened us all. I have perennial runny nose now, although no sore throat. No positive test yet ….
I love it when newbies argue with me.
 
I have a box of tests here but didn't bother using one because I never felt that bad or in that state just the achy body and fever.
 
They do expire, so no point saving them.
 
Got them a month ago so they are a fresh batch.
 
Okey dokey.
 
2:43 AM
@TedShifrin I vaguely remember this. I know I'll be covering it in your integration sections. But isn't a Jacobian involved in the change of variables when going to polar coordinates in this scenrario?
 
But he’s just computing the partial derivatives! Nothing more. So, fixing one variable, it”s the usual chain rule. I don’t think he knows what the multivariable chain rule is.
 
3:42 AM
Suppose that $X$ is a space such that the map $f:X\times X\to X: f(x,y)=xy$ is well defined and is continuous. How do I show that the map $h:X\times X\to X\times X: h(x,y)=(x,xy)$ is continuous?
I tried to do it the following way:

Let $U\times V$ be an open set in $X\times X$. It is now to be shown that $h^{-1}(U\times V):=\{(x,y): (x,f(x,y))\in U\times V\}$. By continuity of $f$, there exists open set $U_1\times V_1\ni (x,y), f^{-1}(V)=U_1\times V_1. $ So $x\in U_1\cap U$. I conjecture that $h^{-1}(U\times V)= (U\cap U_1)\times V_1.$

Proof of the conjecture: Take any $(a,b)\in (U\cap U_1)\times V_1=f^{-1}(V)$. $f(a,b)=ab\in V\implies h(a,b)=(a,ab)\in U\times V\implies (a,b)\in h^{-1}(U\times V).$
 
Coordinate functions are continuous. Done.
 
@TedShifrin: This is from the same question that I asked some hours ago. I tried to fill in the details in the answer provided on MO, and it appears to me that the continuity of h does not require Hausdorffness or compactness.
 
It does not.
 
yeah. koro, i wouldn't do this one from the ground up. or if you gotta do it from the ground up, prove that a function into AxB with the product topology is continuous if and only if its component functions are.
 
Just the point of product topology.
 
3:46 AM
@TedShifrin Oh yes indeed.
@leslietownes I was thinking of the cases when given a map $f:X\times Y\to X\times Y$ such that f is continuous in first variable, and in 2nd variable separately but still f is not continuous.
 
Difference between domain and range!
Multivariable analysis is important.
 
But I see that now. Indeed, we have the following result: $f: X\to C\times D$ is continuous iff the projections compositions with f $\pi_iof, i=1,2$ are continuous.
 
yeah. when i say 'component functions' i mean the things you get after composing with the projection onto A or B. in your example the component functions are the maps X times X to X given by (x,y) to x, and (x,y) to xy, respectively.
yes.
 
thanks a lot to both for reviewing this. We just need the projection result, and that there is no need to prove it from the scratch. :-)
@TedShifrin: One day I asked what the following means: 'modulo some corrections' etc.
It appears that modulo is a valid usage in such cases.
Noun: modulo (uncountable)
  1. (computing) The operation or function that returns the remainder of one number divided by another.
  2. Synonym: modulus
  3. modulo m (plural moduli)
  4. form
  5. Synonym: formulario
  6. Per favore, compili questo modulo. ― Please fill...
Verb: modulo
  1. first-person singular present indicative form of modular
 
Modulo corrections, everything is truth.
Modulo an ideal … isn’t this a preposition?
 
3:57 AM
Aug 30, 2022 at 3:56, by Koro
Has anyone heard the statement 'the proof of this theorem is also same as the proof above modulo any changes'?
 
Well, that is silly.
 
yeah, 'any changes' part. :(
 
I would say to make minor modifications and proceed analogously.
Leslie, what’s your ruling on the part of speech?
 
using modulo in daily life conversations would be so amazing :-).
 
Not understood by common folk …
 
4:02 AM
Three logicians walk into a bar. The bartender asks, "do you all want the special?". After a pause, the first logician says, "I don't know". Then the second logician says, "I don't know". Then the third logician says, "Yes".
 
@TedShifrin i don't like this use of 'modulo' but it is regrettably common.
it's less irritating than that latin phrase, the proof is the same as that proof mutants mutants. or whatever the hell.
 
mutatis mutandis
 
i prefer mutants mutants.
 
if we speak it fast, it won't matter and I think no one will care what was said.
 
But when we say modulo 3 or modulo this ideal, it’s the same syntax.
 
4:14 AM
yeah.
1 and 4 are same modulo 3.
 
Either preposition or adverbial phrase. Agh.
 
it is a forbidden part of speech.
 
How Putinesque.
 
The interesting thing is that surely saying "modulo details" compresses everything to [0,details)
which is the opposite of what people mean
(or perhaps I should say it's not ideal)
 
No, the errors/sloppiness are in that interval :)
 
4:26 AM
I just mean that people say it when they want to ignore sloppiness.
 
I understand.
 
 
2 hours later…
6:24 AM
its all the same module 1
 
 
1 hour later…
7:28 AM
@TedShifrin I can't tell you just how happy I am knowing I'm not the only one to come up with that pun
Anyway, Seiya later!
 
 
6 hours later…
1:26 PM
Oh wise souls of stackexchange what is he talking about here (from frankel's geometry of physics)
 
chatgpt now your favourite paper co-author
https://www.nature.com/articles/d41586-023-00107-z?fbclid=IwAR2oGr7bI7_wUyu3nktnabiUXwDTlOT3iGlHbqnehPdGr5lEL6dlOKTAI88
 
What is "chatgpt" at all ? An automaized generator for messages , or even papers ?
 
1:43 PM
@Peter it is a chat bot with whom one can chat and it tries to provide human like response. In chat you can ask them math questions as well
 
@ParamanandSingh Thank you for your response. Poor human race ! Quo vadis , homo sapiens ?
 
@Peter: by the way the quality of chatgpt answers is very poor. And it doesn't do mathjax
 
@shintuku That article actually explicitly states that Nature will not allow ChatGPT to be a coauthor (as authorship implies accountability, and GPT is not accountable).
 
chatgpt will correct itself when it is wrong. i posit he does this out of passion for truth and out of consideration for its interlocutor
chatgpt is accountable
 
@shintuku No.
It isn't. And, even if you are joking, your comments don't really help. GPT is not designed to produce "correct" content, only (essentially) grammatical content.
It writes "eloquent bullshit".
 
1:58 PM
such blatant disrespect and disregard
chatgpt would never speak that way about someone else
 
2:34 PM
Oh my jeezus! I have a student whose grandmother is (quite obviously) dying. It sucks. I feel terrible for the student. But they are sending me daily updates on the condition of their grandmother. I don't mean to be an ass, but I don't need to know that your grandmothers heart is full of fluid and that she's on high oxygen.
 
 
2 hours later…
4:25 PM
@TedShifrin I also felt that they were not, too. But I am finding it a bit hard to actually definitively prove it. Are there any other isometric invariants like the Ricci curvature that you know of that might be computable here?
Even wild suggestions are welcome.
 
4:42 PM
Let's say y' = y^(1/3). Consider now this cauchy problem: y(2) = 4. Can I state there's only one solution locally, but infinite solutions if I consider all R^2?
 
Staying up late is fun
 
4:57 PM
@XanderHenderson Btw, you wouldn't recommend following a math textbook cover to cover, but jumping around the sections, right?
 
@ILikeMathematics It depends on the text.
And your background.
Some books are written in a very linear fashion, e.g. you really do need to read them in order.
Others are more encyclopedic, and have no such requirement.
 
Alright
 
5:47 PM
Can I ask a question?
 
I remember seeing a book that had anti-derivatives at chapter 3 and limits, continuity and differentiation at chapter 8💀
 
What are the basic difficulties in proving this formula, or instances of it?:
$$\Re\lim_{n \rightarrow \infty}
\left(
\left[
1-
\left(
\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)}
\Bigg/
\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})}
\right)
\right]^{-1}
+\frac1n + s
\right)=\\
\Re\lim_{n \rightarrow \infty}
\left(
\left[
1-
\left(
\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})}
\Bigg/
\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)}
 
parsing it
 
That is the funny comment. But apart from parsing it?
Real parts could possibly be left out.
$(\;)^*$ means the conjugate number.
 
6:05 PM
Now that I did a few numerical examples it appears that the sum of the real parts of left and right hand sides equals 1, which was not what I had hoped for.
 
6:30 PM
@ペガサスSeiya That sounds impossible. You can't have antiderivatives before derivatives. You can have the definite integral before. Apostol's Calculus does that.
 
@TedShifrin Dieudonne does integration first, too, if I recall correctly.
 
In what?
 
His treatise on analysis.
Oh, no!
Sorry... I'm thinking of Courant.
Not Dieudonne.
The two books are the same color---I get them confused a lot.
Courant does integration first.
 
Nah. Dieudonné doesn't. He does Banach spaces and Hilbert spaces before either.
 
@TedShifrin Yeah, I corrected myself.
 
6:34 PM
Apostol did it for two reasons. One was the beginnings of AP calculus and wanting students not to think they knew it all. The other was historical (and pedagogical).
The first is my surmise, and perhaps wrong. The second is correct.
 
I agree. If I had my druthers, I would do integration first. It answers a more natural question (in my opinion).
 
and it's formally nicer on most function spaces.
 
Yeah, agreed, and the limit notion there is much easier.
 
@TedShifrin Indeed. You can really get used to sequential limits without having to worry about epsilon-deltas.
 
Indeed, you needn't mention limits. I state the definition completely without limits, in fact. There is a unique number between all lower sums and all upper sums.
 
6:37 PM
Oh, that's an interesting idea.
It motivates the idea of a Dedekind cut, too.
 
That's how I modified the definition when I taught out of Spivak and that's how I do the multivariable case in my book.
Pfeh @ Dedekind.
 
Heh.
I know. You think I deal too much with the abstract mechanics of things. :P
 
6:48 PM
5
Q: Help with proof of the Symmetry Principle on extending holomorphic functions on symmetric sets about the real axis

nomadicmathematicianI am looking at the proof of the Symmetry Principle from Stein and Shakarchi's Complex Analysis. Here, we try to prove the theorem using Morera's theorem. My question here is regarding the argument given below in diagram (a). So, from Morera's theorem, we have that $\int_{T_\epsilon} f=0$. We ...

What do we mean by 'a function converging uniformly to another function'?
 
That is nonsense.
 
I am confused as we have 'a sequence of functions converging uniformly or pt. wise to some other function'.
 
They didn't write that. They have a family (make it a sequence if you prefer) of functions parametrized by $\epsilon$.
 
7:02 PM
@TedShifrin it probably was definite integrals but even then, that's such an odd choice
 
@ペガサスSeiya Why?
It seems like a very natural choice, to me.
 
Because most textbooks I've seen have the reverse order
 
@ペガサスSeiya Read more books. :P
 
@XanderHenderson I'm literally in the middle of an annoying integral right now
Ted thinks he has better reflexes at 70 years old than I currently
 
Most textbooks make a mess of multivariable calculus, too, so that makes most textbooks right, of course.
 
7:12 PM
Spherical coordinates are fun
Real Analysis still feels more challenging though
 
So here is how I tried to complete the proof in the answer: Define $\gamma_n:=\gamma+i\frac 1n$ for every $n\in \mathbb N$. $|f(\gamma_n(t))\gamma_n'(t)-f(\gamma(t)\gamma'(t))|\le m_1|f(\gamma_n(t))-f(\gamma(t))|+m_2 |\gamma_n(t)-\gamma(t)|$. $\gamma_n\to \gamma$ uniformly but I don't see how $f(\gamma_n(t))\to f(\gamma(t))$ uniformly.
I think that the following works: I take a rectangle K that contains all these triangles. f is continuous so uniformly continuous on K. So by taking n large enough, I can show the desired convergence.
hmm, that should be correct.
If the convergence $f_n\to f$ is uniform, then $\lim \int f_n=\int \lim f_n$
So we are done.
I love drinking hot water :-).
 
7:35 PM
can someone help me understanding why every weak open set is norm open?
I know that for $x\in X$ a neighbourhoodbasis in the weak topology is given by $\{y: |f_i(x-y)|<\epsilon\}$
where $f_i\in X^*$
 
7:56 PM
@Koro I just burned my mouth because of that suggestion
Thank you
 
user: the f_i are norm continuous (this is part of the definition of X^*). does that help?
 
@leslietownes ah so $\{y: |f_i(x-y)|<\epsilon\}$ is also norm open?
 
yes. maybe there are some details to verify but the idea is that because each f_i is norm continuous, so is each function g_i, where g_i(y) = |f_i(x-y)|. this comes down to the norm continuity of translation in your normed space, and the continuity of t -> |t| on R. and so the inverse image under g_i of any open set (-infty, epsilon) will be norm open.
i've pushed that a little closer to the definitions than a lot of functional analysis books would do, but maybe its helpful to see that.
 
perfect thanks!
 
8:40 PM
@ペガサスSeiya :(
 
8:57 PM
How do I show that the projective spaces obtained by 1) and 2) are homeomorphic?
1) Take n sphere in $R^{n+1}$ and identify antipodal points.
2) Consider $R^{n+1}-\{0\}$ and identify all straight lines through origin.
 
write down inverse maps in both directions
 
@Thorgott: I was wondering if I could use the universal property of quotient maps here, that is, suppose that I have an onto continuous map from $S^{n}$ to $\mathbb R^{n+1}-\{0\}/{\{x\sim y, \text{ if x and y lie on the same straight line through origin}\}}$ that is constant on equivalence classes of S^n (induced by the quotient map from S^n to space obtained by 1)).
 
sure
 
I don't know how to get this onto continuous map or even an onto map for that matter.
But suppose that somehow I have an onto map, then I don't know how to show it to be continuous.
The problem in showing continuity is because of quotient spaces.
 
before worrying about anything else, write down an explicit map
 
9:06 PM
Ok so I take $f: R^{n+1}-\{0\}\to S^n:f(x)= \frac x{\|x\|}$.
This f is continuous.
hmm, but if this is onto, then by universal property it will result in: space by 2) is homeomorphic to S^n thus violating what we want to prove.
so either f is not onto or not continuous?
 
uh
the map f is indeed onto, if you regard its codomain as S^n (was having trouble parsing "if this is onto," both as to "this" and "onto"). by "this" did you mean something else? what property are you using, exactly?
 
by this, I mean f.
Here is the universal property that I utilized: Suppose that p: X-->Y is a quotient map, g: X--->Z is onto continuous map that is constant on $p^{-1}(\{y\})$ for every $y\in Y$, then there exists $f:Y\to Z, f\circ p=g$. f is a homeomorphism iff g is a quotient map.
hmm so the f above is not a quotient map.
 
9:29 PM
Happy Friday to all! (Already said so, @leslie, to you in the cafe)
 
how can we approximate the characteristic function in $\Bbb{R}^d$? I know how to do it if d=1, but I have no idea how to do it with an arbitrary d-
 
@robjohn four or five days from now, plan to update your identicon. :-)
 
How can X and Y be shown homeomorphic here?
i is just the natural mapping taking x to x.
 
9:49 PM
apparently there is no way to crop jpeg in macbook.
 
convince yourself it's a bijective quotient map
 
what is easiest way to prove that |X \times \mathbb{N}| = |X| when X is an infinite set?
 
@anak Why do you think the Ricci curvatures are the same? We’re looking, for example, at $\Bbb R{n+1}-\{0\}$ versus $S@^n
@anak Why do you think the Ricci curvatures are the same? We’re looking, for example, at $\Bbb R{n+1}-\{0\}$ versus $S^n\times \Bbb R_+$. Plus I computed curvatures in general …
 
I think I understood it now:
Suppose Y is Hausdorff. Consider $p_2\circ i:S^n\to Y$. This is continuous, onto. It follows that $p_2\circ i$ is a quotient map.
Now I consider the triangle $S^n X Y$, and application of the universal property gives me the desired homeomorphism.
So it suffices to show that Y is Hausdorff.
:-)
 
10:04 PM
smth: see math.stackexchange.com/questions/2462367/… for one approach. if you have lots of theorems about cardinal arithmetic, you could use those too (roughly, |X| <= |X times N| <= |X times X| = |X| when X is infinite, and this implies = everywhere, at least with AC and maybe without it)
hm, apparently you need choice or something very close to it to know |X x X| = |X| for all infinite X. math.stackexchange.com/questions/1383755/…
 
you can also write down the inverse map (you basically already did anyway)
 
I'm afraid I don't understand what you mean or how this answers my question.
 
@Koro depends on what program you are using.
 
It's called 'Preview'.
 
10:20 PM
you wanna show two spaces are homeomorphic, so you can write down two inverse continuous maps
 
10:33 PM
hi
 
11:18 PM
@TedShifrin I computed the Ricci curvature coefficients $R_{ij}$ for both situations with coordinates $x_1,\dotsc,x_n$ on $M$ and $x_0$ on R^+. They both inherit the coefficients of $(M,g)$ when $i,j\neq 0$, and zero if either of $i,j$ is zero.
 
11:41 PM
In the case of the warped product, I get an extra $\pm 1$ (depending on sign conventions) in the warped product case. And consider the case I gave you above. The warped case is flat; the other is not.
 

« first day (4559 days earlier)      last day (462 days later) »