« first day (1945 days earlier)      last day (3080 days later) » 

8:00 PM
one thing I have learned is that math is a social game, in every aspect. you can't confine yourself in a dungeon and do a breakthrough.
 
@TobiasKildetoft Sorry, I'm not familiar with that show.
 
lol, "show"
 
@happyEddie vixra is like arXiv, except without even the minimal requirements to upload something
 
ah, not a show, I read that carelessly, sorry
@TobiasKildetoft OK thanks.
 
Huy
@happyEddie different people are different. some people can rather accurately evaluate themselves, others tend to completely under- or overestimate themselves repeatedly. maybe experience helps, but I don't know
 
8:02 PM
it is sort of better than entertainment shows in TV's.
so it'd'nt be a complete lie to call it a show.
 
@Huy And the worse someone is at something like math, the more a tendency they have to over-estimate their skill (the Dunning-Kruger effect)
 
Huy
is that so?
 
@BalarkaSen True, there is much entertainment to be had reading stuff there
 
lol, Dunning-Kruger effect
 
Huy
at my uni I feel like most people very accurately evaluate themselves
 
8:04 PM
that thing even has a name
 
Huy
then again everyone's a genius, except for me
 
@BalarkaSen Just to return to your earlier comment that you can't confine yourself. I think there is a counterexample: Andrew Wiles and his work in proving the Fermat thm.
 
@BalarkaSen Ever seen the site with notable reviews from MathSciNet?
@happyEddie He was not at all confined apart from the last few years
 
no, @happyEddie, Wiles didn't confine himself. sorry.
 
And he only confined himself as regarded his work on Taniyama-Shimura (as in, he didn't announce far and wide that he was working on it)
 
8:05 PM
that's very contrary to popular belief. he just didn't share his work/progress with anyone
 
@BalarkaSen Or do you even have access to MathSciNet btw?
 
he studied and worked and taught and talked with mathematicians his entire life about Galois representations and elliptic curves. his work stands on what people did earlier.
@TobiasKildetoft Nope and nope.
 
@BalarkaSen Ahh, the list is with links to MathSciNet, so not much use without access
 
aw. what is it about?
 
@TobiasKildetoft @BalarkaSen Really, that's surprising. Well thanks for correcting my misconception.
 
8:07 PM
@BalarkaSen It is a list of reviews that are worth reading for various reasons
mainly because the author of the review is unusually honest
 
oh?
 
@BalarkaSen I might do the polygonal theorem aswell
as application of Ulam Borsuk theorem
 
Trying to self-study some algebraic geometry and I'm reading and working through the exercises in Algebraic Curves by Fulton. There is this exercise: find the irreducible components of V(X^2+Y^2-1, X^2-Z^2-1) in C^3. Any hints how you would start to solve this? I'm quite stuck here.
 
Ok. It's Borsuk-Ulam, by the way.
 
checking in to see if @Balarka is doing his homework :D
 
8:19 PM
Noticed that (Y-iZ)(Y+iZ)=Y^2+Z^2 is in the corresponding ideal but Y-iZ and Y+iZ are not so it should mean that the set is not irreducible and will decompose.
 
Not as much as I'd want to, @TedShifrin, schoolwork is taking up my time. I am not doing Mike's exercises either :( I am going to fix a date when I'll read Lagrange maximization, do that exercise you gave me, and send e-mails.
 
Maybe you need to emulate Mike and spend less time in chat, Balarka :)
 
see that is also what is happening to me
I didn't also do star convex problem you gave me
because of school
 
You're doing math, @L33ter
I am studying... nonsense :(
very sad face
 
@TedShifrin
Suppose we have the following
 
8:21 PM
@TedShifrin How about 4th? I'll read the stuff in 3rd, and then e-mail at 4th. Would that be ok to you?
By read, I mean, read/do exercise/study
 
Balarka: It's not my role here to set deadlines.
 
$f(z) = \Sigma_{n = 1}^{n = \infty} \frac{(log n)^k}{n^z}$
 
You do what you want to. But you do waste lots of time here.
 
I want to prove that this will be analytic
 
@BalarkaSen Speaking of Wiles, here is one of them that is short enough to post here: It is a review of the paper The Euclidean character of the Fermat's last theorem by Malvina Baica. The reviewer is Andrew Granville who writes
 
8:22 PM
I'm just asking if you'd be free :)
 
On what domain, Karim?
 
okay so consider the riemann zeta function
 
"Herein the author states "her genuine concern'' about Wiles's purported proof of Fermat's last theorem [A. Wiles, Ann. of Math. (2) 141 (1995), no. 3, 443–551; MR1333035 (96d:11071)] which, after all, appeared in an "in-house publication in the Annals of Mathematics at Princeton''. Baica's concerns seem to stem from a worry that results concerning elliptic curves "may not be equivalent to the result in the Euclidean geometry''.
She backs up her concerns by noting that "there is a need to provide Galois' connection from category theory''. Of course, she has no such worries about the validity of her own, Euclidean-algorithm-inspired, proof of Fermat's last theorem."
 
on domain $\theta = \{ z \in C : Re(Z) > 1\}$
 
Um, I have no specific plans, @Balarka.
 
8:23 PM
@TedShifrin My chat here has drastically decreased. You haven't been here much, I'm online only at my bedtime.
 
consider the riemann zeta function
 
I haven't been here much for semi-obvious reasons.
 
Oh? What might that be?
 
Yes, Karim, and how does one show it for the Riemann zeta function?
 
which is defined as $g(z) = \Sigma_{n = 1}^{n = \infty} 1/n^z$
so I used Weistrass M-test to show that the riemann zeta function converges uniformly
 
8:24 PM
say things carefully, Karim ...
that series converges uniformly everywhere on its domain?
 
Anyone familiar with quantitative finance here?
 
i.e that the Riemann zeta function is analytic on its domain
 
Yikes, not that I've ever seen, @misheeko.
No, Karim.
You're talking in circles.
 
Can you not use Morera's theorem up there and be done?
 
Not quite so glibly as that, @Balarka.
 
8:26 PM
well, yeah I showed that the riemann zeta function converges on arbitrarily closed disc in the domain.
 
@TobiasKildetoft That seems like an odd review.
 
But I love applications of Morera's Theorem. When did you learn this?
 
@BalarkaSen Why?
 
Yes, Karim, the correct statement is that the series converges uniformly on compact subsets of its domain.
hi @Tobias
 
yeah
 
8:26 PM
Hi @TedShifrin
 
OK, so can you make the same argument in this exercise, Karim?
 
@Ted: I should follow my own example.
 
@TedShifrin From Titchmarsh.
 
@MikeM: You mostly do.
 
so I was thinking if I take the k-derivative of the riemann zeta function
 
8:27 PM
I didn't ask where, @Balarka. :)
 
When I was learning complex analysis to understand the proof of PNT.
Right, just noticed.
 
@Balarka: You still haven't learned what line integrals are, btw.
 
right
because if we substitute
k = 1
 
I know what line integrals are, but I have forgotten most of complex analysis because I didn't study carefully :)
 
and take the first derivative
 
8:28 PM
I have to restudy it.
 
apostol is nice
 
It'll make more sense once you know about differential forms, Stokes's Theorem, etc., @Balarka.
 
Ah.
 
@MikeM: I thought you already took it.
 
I remember MikeM saying that he is taught complex analysis before
 
8:29 PM
So you're telling me that this series can be obtained by differentiating term-by-term? Can you justify when that works?
 
@DanielFischer Ah I see... But does this imply that $\inf \{ ||x-y||: y \in Y\}=0$ ?
 
yes I have that idea
 
@BalarkaSen On another review of a paper by that same author, the reviewer concludes "The reviewer regrets that this paper is unfaithful to its title." (the paper is titled "Solution of Goldbach's conjecture").
 
I will try it now
 
I want to stop feeling guilty about learning chapter 1 Guillemin-Pollack by learning the proof of implicit + inverse function theorem first, and I'm still stuck at Lagrange maximization :(
@TobiasKildetoft Haha.
 
8:30 PM
OK, Karim, that's not an unreasonable idea. Do you have to work that hard, however? Can you just repeat exactly the same proof as for Riemann zeta?
Plus you're supposed to learn some integrals and differential forms stuff, @Balarka, although you don't need that until the last chapter of G&P.
But you do need it for your education :P
 
Yeah, I do want to study forms, but implicit and inverse function theorems are priority.
 
Took what?
 
BTW, @Balarka, although it's not necessarily your style, when I taught Guillemin & Pollack numerous times, I usually saved the proofs of the transversality theorems in Chapter 2 until after I did the various mod-2 intersection theory applications, just so students would be motivated. You're already motivated.
@MikeMiller Your own advice.
 
@TedShifrin this excerise is for my class
 
Mostly.
 
8:32 PM
I understand, Karim.
 
@BalarkaSen it is amazing your doing alot of topology and stuff and you didn't learn analysis
 
@TedShifrin A cool exercise from G-P was to prove that the stability theorem fails for noncompact domains.
 
but same with me I didn't take a course in analysis yet atleast formally
 
It had some very complicated ctrexample, but I found my own.
 
BTW, @Balarka, there are piles of typos in G&P. I used to have a list of them on my website. I'm not sure if it's still there.
 
8:34 PM
Consider R^2. Consider the submanifolds R and another copy of R which is like some damped sine curve which intersects R bit by bit as it goes, yet tangential to it. Let the homotopy be obtained from translating above by $t$.
 
They build character.
 
Many of them really confuse students, @MikeM.
Remember, I taught sophomores, juniors, seniors, not graduate students (except for one or two).
 
You'll never be able to make them transeverse at any level of the homotopy.
 
what is a manifold.
 
I have no idea what that means, @Balarka.
 
8:35 PM
@TedShifrin How else are you supposed to teach them?
 
@TedShifrin Yikes, that's a problem.
 
That doesn't seem like a counterexample to stability, @Balarka. That seems like a counterexample to genericity.
Don't worry, @PVAL, I managed to be challenging/confusing enough without typos in the text/exercises.
 
did a student ever ace your higher level classes @TedShifrin ?
 
Many of them were pointed out to me by former students, though. :)
 
8:36 PM
Although, surprisingly, he never defines a manifold in the whole lecture...
 
Yes, Karim, quite several did. @MikeM knows one who did very well.
 
@PVAL: About to email you.
 
@Balarka: I still say that's a counterexample to something else.
 
@MikeMiller Um. Which part?
It's a damped sine function intersecting $\Bbb R$.
So it takes values below $0$.
@TedShifrin Sorry, but what is genericity again?
 
You'd better ask yourself that question, @Balarka.
 
8:38 PM
Stability says if $f : M \to N$ is transeverse to $W \subset N$, then for every homotopy $f_t$ there is an $\epsilon$ such that for all $t < \epsilon$, $f_t$ is transeverse.
 
Right.
 
skd
hi
bben
any one here?
 
Tran sever se sounds like a Harry Potter spell.
 
It sounds gruesome.
 
But there is always a $\delta > 0 $ such that lifting that sine curve would make it tangential to $\Bbb R$.
 
8:40 PM
@Balarka: The thing where you said "This can't be made transverse by any level of homotopy". Those are words, for sure, but I don't see how they fit into a coherent sentence.
 
Stop and think, @Balarka. You're being very sloppy, I think.
 
I mean the homotopy is obtained from shifting the sine curve by some factor of $t$.
Maybe I should stop and think. Hmm.
 
Shifting vertically. But, Mike's complaints notwithstanding, mine has priority.
I'm taking a lunch break. BBL.
 
Ok, ok. $M$ be the submanifold of $\Bbb R^2$ which is the damped sine function. And the homotopy of $M \hookrightarrow\Bbb R^2$ is $f_t$ where $f_t$ shifts the curve vertically by $t$.
Hmm.
And now I am claiming $f_t$ is tangential to $\Bbb R$ for arbitrarily small $t$.
Not sure what goes wrong here.
 
Well does your example satisfy the hypotheses? (are the two maps transverse?)
 
8:47 PM
$M$ is transverse to $\Bbb R$, it's the damped sine function which intersects $\Bbb R$ transversely wherever it does, right?
Note that I am working with embeddings here. So submanifolds.
 
You should write out everything explicitly.
 
Yeah, let me do it. Thanks.
 
@evinda Yes, it does.
Hi @Ted.
 
@BalarkaSen This statement "some damped sine curve which intersects R bit by bit as it goes, yet tangential to it" suggests to me that your maps are not transverse.
 
No, I mean, you start with the damped sine curve $f_0$. $f_t$ be the thing obtained from lifting by $t$. $f_0$ is transeverse, right? But as $f_0$ has arbitrarily small local minimas, $f_t$ would be tangential to $\Bbb R$ for arbitrarily small $t$ (not $0$), correct?
 
8:53 PM
Hellllllllloo.
Can anyone tell me how studying zeros of multivariate polynomials is related to geometry in the naive sense of the word?
 
@Balarka: I haven't been paying attention to what you're saying but it's going to be hard to argue with you either way when you don't ezplicitly define your notions like damped sine curve.
@Anthony: Well, study in what sense?
 
How do we deduce it? @DanielFischer
 
Ok, I apologize for being sloppy, I'll write things down.
 
@BalarkaSen That sounds like it will work. Now write out the homotopy and your candidate $f_0$ and actually show that there exists arbitrarily small $f_t$ which has non-transverse intersections with the x-axis
 
@MikeMiller I meant study in the sense that wikipedia says it's a branch of mathematics studying zeroes of multivariate polynomials. But I think I may have found an explanation sufficient for myself here: quora.com/Why-does-algebraic-geometry-have-geometry-in-its-name
 
8:58 PM
@Anthony: Sure. All I was hinting very vaguely at was that "studying" is too vague to mean anything, much less call it geometry. It's much more precise things that one is interested about these zero sets that make it geometry.
For instance, I would hardly say "Does $f$ ever output the number two?" is a geometry question, say.
 
Yeah.
I should, as always, be more careful with what I say.
:P
Thanks.
 
@TedShifrin
here?
 
@evinda Clearly, we have $\inf \{ \lVert x-y\rVert : y \in Y\} \leqslant \lVert x - y_n\rVert$ for every $n$, since $y_n \in Y$. But $\lVert x - y_n\rVert \to 0$ implies $\inf \{ \lVert x-y_n\rVert : n \in \mathbb{N}\} = 0$.
 
@DanielFischer Ah, I see... But we use the fact that $ $\inf \{ \lVert x-y\rVert : y \in Y\} =0$ in order to find a contradiction, right?
 
Hello, and good <time interval> to each of you on this fine <larger time interval>
 
9:09 PM
Urk, minimums of $\sin(x)/x$ are actually kinda hard to deal with, heh.
 
@MikeMiller Actually, are you still there?
 
Once one agrees that there are arbitrarily small minimums, it's not hard. Pick a minimum of value $-c$ such that $c < \epsilon$. Now consider $\sin(x)/x + c$.
 
Hi all.
 
@MikeMiller @PVAL @TedShifrin OK, here's everything I have to say, done formally: $f : \Bbb R \to \Bbb R^2$ be the map $f(x) = \sin(x)/x$ for $x \neq 0$ and $1$ for $x = 0$. $f$ is transverse to $\Bbb R \subset \Bbb R^2$, the x-axis. Let $f_t(x) = f(x) + t$ for $t \in [0, 1]$. For any $\epsilon > 0$, you can find a minimum $x_\epsilon$ of $f$ such that $|f(x_\epsilon)| < \epsilon$. Denote $c_\epsilon = f(x_\epsilon)$.
Then $f_{c_\epsilon}(x) = f(x) + c_\epsilon$ is tangential to $\Bbb R$ at $x_\epsilon$, as the tangent at $f_{c_\epsilon}(x_\epsilon)$ lies on $\Bbb R$.
Given any $\epsilon > 0$, you'll find $f_t$ which is not transverse to $\Bbb R$ for some $t < \epsilon$ : pick $t = c_\epsilon$.
Cruds, typo, I meant $c_\epsilon = |f(x_\epsilon)|$.
To wrap up, $f$ is transverse to $\Bbb R$, but the homotopy $f_t$ contradicts stability as you can find arbitrarily small $t$ such that $f_t$ is not transverse to $\Bbb R$.
 
Howdy.
 
9:23 PM
@evinda Contradiction to what? We use it to show that the distance is not attained if $x \notin Y$.
 
if some series converge uniformly does it mean that the negative of the series will converge uniformly aswell ?
 
@Anthony: I'm only moderately here. You're always free to ping me w questions.
 
@DanielFischer So is the following right?

$Y$ is dense in $\ell^2(\mathbb{N})$.

This means that for every $x$ there is a sequence $(y_n)$ in $Y$ such that $\lVert y_n - x\rVert \to 0$.

We have $\inf \{ \lVert x-y\rVert : y \in Y\} \leqslant \lVert x - y_n\rVert$ for every $n$, since $y_n \in Y$. But $\lVert x - y_n\rVert \to 0$ implies $\inf \{ \lVert x-y\rVert : y \in Y\} = 0$.

But it cannot be that there is a $y \in Y$ such that $||x-y||=0 \Leftrightarrow x=y$ since $x \notin Y$.
 
@MikeMiller
I have a question
if we have the riemann zeta function which uniformly convergent on every closed disc
is the negative of that also uniformly convergent
nvm it is true
every constant multiple of convergent series will be convergent
 
9:42 PM
Ah, ok, @Balarka, I misunderstood the picture. I apologize. That is a counterexample to stability. There are far easier examples, btw ... Think about an example that is transverse at $t=0$ by default.
hi @DanielF @Anthony
 
Well, I should apologize for my bad picture.
Not you.
 
BTW, @Balarka, I uploaded the errata to my webpage (on my profile) if you want it when you start doing G&P seriously.
 
Thanks!
Um, what do you mean by transeverse at $t = 0$? You mean transverse to $\{0\}$?
Not severe, damn.
 
I only heard G&P from somewhere.
 
No, I mean the original map is transverse by default.
 
9:46 PM
@evinda You should say that you take an $x \in \ell^2(\mathbb{N}) \setminus Y$ somewhere before the last line. Then it's okay.
 
I haven't seriously studied this but I wonder whether it's a must.
 
Who decides what is or is not a must, @Frank?
 
@TedShifrin I solved the problem
the complex analysis one
 
By just redoing the original proof?
 
yeah
and I did it also
 
9:47 PM
But the original map $f = f_0$ is transverse to $\Bbb R$ in my case?
 
BTW, I suggest you understand Balarka's suggestion about Morera's Theorem. That's a very powerful technique.
 
using the RZ function
 
Or am I misunderstanding you?
 
ok I don't know about morera's theorem
but will check it now
 
@L33ter Converse of Cauchy's theorem.
 
9:48 PM
@Balarka: I'm suggesting you think of a simpler counterexample. The one I have in mind, the original map is transverse BY DEFAULT. Do you know what that means?
@Balarka @L33ter: But one must assume/prove continuity!!
 
No, not really.
 
I mean, whether it's worthwhile to study G&P or just use it as a reference to look up.
 
BY DEFAULT signifies that it holds vacuously (i.e., there are no $x$ in the intersection).
 
Oh, so trivial intersection. I see.
 
9:49 PM
Depends on your interests and talents, @Frank. I would say it's arguably the best undergraduate course someone can take.
 
now I need to prove the following
$f(z) = \Sigma_{n = 1}^{n = \infty} e^{-n} sin(nz)$ is analytic on the domain $\theta = \{ z \in C : -1 < Im(z) < 1\}$
I guess
I can rewrite sin(nz)
 
Why bother?
 
$sin(nz) = \frac{1}{2i}( e^{inz} - e^{-inz})$
 
Um, no.
 
I mean
n in there
 
9:52 PM
But, still, why bother?
 
G&P is not a terribly good reference book. It's also not hard to learn from if you wanted.
 
I derived a really cool thing all on my own! :D
 
One does need the appropriate multivariable analysis/point set background, @MikeM.
 
I don't know I thought it will simplify things
 
Similar techniques can be used, I think. The deal is to make them transverse at $\infty$. I'd start with something asymptotic to $\Bbb R$ and add waves to the ends. Let me ponder.
 
9:53 PM
Frank has that.
@Balarka: This is absurdly unnecessary.
 
I don't know Frank, so that's why I found his question challenging :)
 
Since we have Milnor's book and notes, say.
 
In how many ways can you can place $n$ things into $k$ spots, if there must be $1$ thing in each spot at least and $x$ things in each spot at most?
 
@Balarka. For my idea, don't use all of $\Bbb R$ as the submanifold.
 
On differential topology
It's very thin, though.
 
9:53 PM
@Ted: That works just fine for the appropriate example.
 
Oh, if you've read Milnor's little book, you know, in a nutshell, most of what's in G&P.
 
Answer:
$${n-1 \choose k-1} - k {n-x-1 \choose k-1}$$
 
Thanks. I haven't digested it well, especially these chapters about cobordism, so I need to reread.
And there's a lecture note given by Milnor on Differential Topology and vector bundles, approximately 30 pages.
 
G&P don't do the cobordism stuff. But it's worth your reading and working exercises. There's a lot of rich stuff that's not in Milnor, and it's much more detailed. Milnor is like a telegraphic message :)
Oh, that's graduate level, but I prefer Mo Hirsch's book.
 
I guess we could use Weistrass M-test here right @TedShifrin ?
 
9:55 PM
Hirsch also does a lot of the transversality stuff, but very very sophisticatedly.
 
@MikeMiller Hmm, ok. I'm thinking.
 
Hirsch's book is a bible.
 
Yes, Karim, it should be just like the others you've done. You just need to think about the sin function.
But super super sophisticated, @MikeM.
 
You mean asymptotics are unnecessary, or do you mean the waves?
 
I always remind you I'm the one who introduced you two. :) I should get a wedding invitation.
 
9:56 PM
I guess the later.
lol.
 
@Balarka: I'm telling you there's an example that a 9th grade algebra student could understand.
 
@Ted: I remember.
 
Could somebody perhaps help me out with a proof of continuity implying uniform continuity on a closed bounded interval? (I can upload a screenshot of the proof and where I can't follow it.)
 
@Khallil: Suppose not. Negate carefully.
 
I have the negation, but the part where we choose $\delta$ as in my case $1/n$, generating two sequences $\{x_n\}$ and $\{y_n\}$ and concluding stuff after that confused me, @Ted.
 

« first day (1945 days earlier)      last day (3080 days later) »