« first day (1653 days earlier)      last day (3663 days later) » 

17:04
Hello @DanielFischer!!!
Suppose that we have $x=(\overline{x_k}), y=(\overline{y_k}) \in \mathbb{Z}_p \setminus{\{0\}}$, there are $n,m \in \mathbb{N}$ such that $x_n \not\equiv 0 \mod{p^{n+1}}$, $y_m \not\equiv 0 \mod{p^{m+1}}$. This means that $x_n=p^{n'}u_0, u_0 \in \mathbb{Z}_p^{\star}, n' \leq n$, $y_m=p^{m'}v_0, v_0 \in \mathbb{Z}_p^{\star}, m' \leq m$. Then we have that $x \cdot y=p^{n'+m'}u_0v_0$. We set $l=n+m+1$. We have that $x_l \not\equiv 0 \mod{p^{n+1}}$ and $y_l \not\equiv 0 \mod{p^{m+1}}$. How do we deduce from this that that the greatest power of p that divides $x_l$ is $p^
@evinda We don't deduce it from that, we deduce it from $x_n \not\equiv 0 \pmod{p^{n+1}}$ and $x_l \equiv x_n \pmod{p^{n+1}}$.
Morning, @Daniel
And similar for $y$ of course.
Morning @Mike.
@DanielFischer An exercise I had on my PDE homework was to find a smooth function, not bounded by any polynomial, that defines a tempered distribution. I spent a long time trying to find one that grows just slightly faster than polynomial, but eventually found $e^x\cos(e^x)$, which is the derivative of $\sin(e^x)$, which is bounded, so a tempered distribution; so the former defines a tempered distribution. This isn't very satisfying.
Do you know if there is an example if we also place the demand that it's monotonic?
@MikeMiller I don't know, but looks unlikely. Rapidly decreasing functions can decrease rather slowly.
17:14
That was my expectation after $e^{-\log x \log log x}$ turned out to be rapidly decreasing.
@MikeMiller After fudging a bit at $0$.
Sure, mash it against something supported in $(3,\infty)$, and 1 for $x \geq 4$.
17:26
"mash it"? what is this, the south?
btw, good night, @Mike
Morning, @Ted, and happy birthday.
thanks, @Mike
@DanielFischer Can't we conclude from $x \cdot y=p^{n'+m'}u_0v_0$ that the product is different from 0?
Hi all!
@TedShifrin I need you!! :D
@TedShifrin Is it your birthday? Many happy returns.
17:35
Yup, @DanielF. Vielen Dank :)
When one speaks of a radially symmetric solution, does one really mean rotationally symmetric?
@TedShifrin Joyeux anniversaire!
@TedShifrin Happy Birthday!!!
@evinda You can conclude it from that indeed.
@DanielFischer How could we justify it?
17:36
i.e., constant on spheres about the origin of a given radius
Thanks, all. Now if you'd just grade my exams for me ...
@evinda Justify what precisely?
@TedShifrin I don't undersand Derek's answer here.
@TedShifrin you've asked too much.
@DanielFischer That $x \cdot y \neq 0$
17:37
Yes, @Mike, function of $r$ only.
Silly name.
For me when we have $G/Z(G)$ isomorphic to $G$ then $Z(G)$ is always the kernel so it's trivial by infectivity no?
@evinda Well, you note that $p^{n'+m'} \neq 0$, and that the product of units is a unit again.
@Gato The isomorphism need not be the canonical map. If $Z(G)$ isn't trivial, it cannot be the canonical map.
@Gato: Why are you asking me algebra?
Righto to what @DanielF just said.
@DanielFischer So when we said that $Z(G)$ is the kernel of the map it's only for the canonical map?
17:40
Someone I don't know pinged me a couple days ago about some hard analysis I know nothing about, @Ted.
@TedShifrin Why not?
@DanielFischer $p^{n'+m'} \neq 0$ because $n',m' \in \mathbb{N}$, right?
@Gato Not necessarily only for the canonical map, there can be more maps with that kernel.
@Gato Yes. The point is that - in this case - the projection $G \to G/Z(G)$ is an isomorphism. So it must have trivial kernel, but the kernel of the projection is $Z(G)$.
But it can be perfectly fine that $Z(G)$ is nontrivial but that $G$ and $G/Z(G)$ are isomorphic. It's just that the isomorphism can't be the projection map.
Hello everyone- one more modeling question, if you don't mind
17:42
@MikeMiller@DanielFischer Okay! Many thanks both of you.
@evinda Because $n' + m'\in \mathbb{N}$. Which follows from $n',m'\in\mathbb{N}$.
@daOnlyBG This isn't a fashion show.
har
cough har
Now a completely different question: how can I prove that $f:\Bbb{R}\rightarrow \Bbb{R^n}$ we have $\Vert \int f\Vert\le \int \Vert f\Vert$?
lol, I have a set of data points $W_i, l_i, g_i$. I have to find the best fit for $W=klg^2$
@DanielFischer So can we justify as follows that $x \cdot y \neq 0$?

$x=(\overline{x_k}), y=(\overline{y_k}) \in \mathbb{Z}_p \setminus{\{0\}}$, so there are $n,m \in \mathbb{N}$ such that $x_n \not\equiv 0 \mod{p^{n+1}}$, $y_m \not\equiv 0 \mod{p^{m+1}}$. This means that $x_n=p^{n'}u_0, u_0 \in \mathbb{Z}_p^{\star}, n' \leq n$, $y_m=p^{m'}v_0, v_0 \in \mathbb{Z}_p^{\star}, m' \leq m$. Then we have that $x \cdot y=p^{n'+m'}u_0v_0$.

Since $n' + m'\in \mathbb{N}$, which follows from $n',m'\in\mathbb{N}$ and the product of units is a unit again we conclude that $x \cdot y \neq 0$.
17:44
@Gato Which norm?
what I tried doing was something pretty basic
$\ln W = \ln k + \ln(lg^2)$
sadly, I ended up getting $k=0$
I'm not sure that's the solution
@DanielFischer It's not true for all norm?
This problem asks me to prove some properties of the radially symmetric fundamental solns to the Laplacian. They seem to be trivial if I actually know what the fundamental solutions are, so I have to wonder if I was supposed to prove these just knowing that they're radially symmetric etc
@Gato Some norms allow nicer proofs than others.
Right, @DanielF @Gato. With the usual norm you might consider dot product with $v\in\Bbb R^n$. Otherwise, you might have to resort to approximating integrals by sums.
17:48
@DanielFischer we used this in differential equations with no specific norm..
@DanielFischer I am a little confused now.. we have that $x_n=p^{n'}u_0, n' \leq n, u_0 \in \mathbb{Z}_p^{\star}$ and $y_m=p^{m'}v_0, m' \leq m, v_0 \in \mathbb{Z}_p^{\star}$. How do we deduce something for $x \cdot y$ ?
Happy birthday Proessor @TedShifrin as a present for yourself you may get the "Minds of Math" app at the Apple store for your iPad :-)
Are you a walking advertisement, skull?
@Gato Then you can go through the approximate-integral-by-sums route, or you can throw Hahn-Banach at it. Let $\lambda$ be a linear form, then $$\lambda\left(\int f\,d\mu\right) = \int \lambda\circ f\,d\mu \leqslant \int \lvert \lambda(f(x))\rvert\,d\mu(x) \leqslant \int \lVert\lambda\rVert\cdot \lVert f(x)\rVert\,d\mu(x),$$ and then take the supremum over all $\lambda$ of norm $1$.
Do you have it @MikeMiller?
17:53
Skull, what is that app?
A time line of the greatest minds of math
Only for the iPad
well, if I'm not included on the timeline, I'm not interested! :D
:D
It's free.
OK, back to grading exams before office hours ...
@evinda That doesn't make sense. You may have $x = p^{n'}\cdot u_0$ with $u_0\in \mathbb{Z}_p^\times$ or $x_n = p^{n'}\cdot u$ with $p \nmid u$ or $\overline{x_n} = \overline{p}^{n'}\cdot u_1$ with $u_1 \in (\mathbb{Z}/p^{n+1}\mathbb{Z})^\times$.
17:59
@Ted Happy birthday
@DanielFischer Nice!
@LeGrandDODOM Salut l'ami!
@DanielFischer So is it right like that? :/
$x=(\overline{x_k}), y=(\overline{y_k}) \in \mathbb{Z}_p \setminus{\{0\}}$, so there are $n,m \in \mathbb{N}$ such that $x_n \not\equiv 0 \mod{p^{n+1}}$, $y_m \not\equiv 0 \mod{p^{m+1}}$.

This means that $x=p^{n'}u_0, u_0 \in \mathbb{Z}_p^{\star}, n' \leq n$.
If we had $n'>n$ then $x_n=p^{n'}u_0 \mod{p^{n+1}}=p^{n+1+(n'-n-1)}u_0 \mod{p^{n+1}}$ and $n'-n-1 \geq 0$, so $x_n \equiv 0 \mod{p^{n+1}}$, which is a contradiction.

In the same way, $y=p^{m'}v_0, v_0 \in \mathbb{Z}_p^{\star}, m' \leq m$. Then we have that $x \cdot y=p^{n'+m'}u_0v_0$.
@Gato Salut, quoi de neuf ?
@LeGrandDODOM J'ai vu un exo de l'ENS qui demande de montrer que pour un groupe fini G de cardinal une puissance d'un nombre premier $p$ alors il existe un élément de $G$ non réduit à l'élément neutre qui commute avec tout les autres.
avec les actions de groupes c'est pas difficile mais je me demande comment un MP peut faire ça lol
Salut, @leGrandDodo !!
18:14
Hi, if anyone would like a detailed version of my question, it's right here: math.stackexchange.com/questions/1145369/…
Hi @DanielFischer can I please ask one question.
@Gato j'ai pas de solution en tête, mais je commencerais l'oral en disant que cela revient à montrer que son centre est de cardinal $\geq 2$. Mais considérer juste le centre ne mène à rien. Donc je passerais au centralisateur de chaque élément du groupe...
@TedShifrin Salut
@DanielFischer Let $\Omega$ be open bounded in $\mathbb{R}^{N}$. Consider Caratheodory function $c: \Omega \times \mathbb{R} \rightarrow \mathbb{R}$. Also, assume $u_{n}(x) \rightarrow u(x)~~a.e.~x \in \Omega$, then by the continuity of the Caratheodory function $c$, it follows that $c(x,u_{n}) \rightarrow c(x,u)~~a.e. x \in \Omega$.

Also assume that $|c(x,u_{n})| \leq k$, for some $k > 0$.

Can you see how it follows by the Dominated Convergence Theorem that results in
$$\int_{\Omega}c(x,u_{n})dx \rightarrow \int_{\Omega}c(x,u)dx$$
Huy
Huy
Happy Birthday, @TedShifrin.
18:32
@TedShifrin
18:52
@LeGrandDODOM Ah vous faite les actions de groupes en MP? Dans ce cas oui c'est pas difficile. La solution que je connais utilise l'équation des classes..
@TedShifrin Sir , can you look into my question once . mathoverflow.net/questions/196294/…
MLM
MLM
19:16
When drawing a Bezier curve, do you get a more accurate representation using a uniform sampling or just going with t which is non-linear? I made a demo, but I am still not sure: jsfiddle.net/MadLittleMods/thcstaay
19:46
Hey, quick question. Consider a bijection $f: A \to B$ whose inverse is $g: B \to A$, and suppose that $C \subseteq A$. Is it true that the preimage $g^{-1}(C)$ is equivalent to the image $f(C)$? I don't want to confuse the notation for inverses ($f = g^{-1}$) with the notation for preimages of sets ($g^{-1}(C)$). Basically, I'm wondering if $(g^{-1})^{-1}(C) = C$.
@DanielFischer Would it work to consider sequence $$f_{n}(x) := c(x,u_{n}(x))$$ and then use the DCM as usual?
@JohnDoe Referring to the above? $\Omega$ is bounded, so the constant function $x \mapsto k$ is integrable. Thus you can use the dominated convergence theorem like that.
@DanielFischer Yes so I can dominate it with the constant function. So I can use it as I said, by defining the new sequence of functions and then using the DCM as usual?
20:01
@DanielFischer Is that what I have written right?
@DanielFischer Thanks.
@evinda Yes. But the last line looks like you're repeating things I wrote without really understanding why that is the case.
Heya @OldJohn. Walked any nice hills lately?
@DanielFischer One last quick question, does the following hold for any integrable function, by the fundamental theorem of calculus: $\psi(1) - \psi(0) = \displaystyle\int_{0}^{1}\frac{d\psi(t)}{dt}dt$
@DanielFischer A few, yes -had a great walk recently over hills covered in snow - on a beautiful sunny day :)
@JohnDoe It holds for absolutely continuous $\psi$.
@OldJohn I envy you. Although I hardly remember what "a sunny day" is.
20:06
Oh yeah that's right.
@DanielFischer How could I improve it?
@DanielFischer Very few round here too at the moment. Where are you now?
@OldJohn Northern Germany. Flat and cloudy 11 and a half months of the year.
@DanielFischer darn! I was In Berlin year ago, and that was very grey and flat too!
Westerner that I am, I'm tempted to say that Berlin lies in Poland ;)
20:13
@DanielFischer certainly cold enough at times :)
@evinda Say it in your own words.
Hi @OldJohn.
@ABeautifulMind Hi JL
@OldJohn I sent Jonas the email I sent you too, but he did not say anything. I think maybe he wanted to say bad things.
@ABeautifulMind I don't think he would want to do that - he maybe just missed the email, or intended to reply and never got round to it, ...
20:16
@OldJohn OK. Do you have any life wishes to fulfil now?
@ABeautifulMind Several! I aim to climb all the 214 hills/mountains in the Lake District (known as "Wainwrights") - and I want to play Bach on the piano, ... and several others :)
@DanielFischer Do you mean this proposition?
Then $x \cdot y=p^{n'+m'} \cdot (u_0 \cdot v_0), v_0 \in \mathbb{Z}_p^{\star} \neq 0$, since $u_0 \cdot v_0$ is a unit, as a product of two units and $n'+m' \in \mathbb{N}$, that comes from the fact that $n',m' \in \mathbb{N}$
Everything, @evinda. If you can say it in your own words, that helps understanding.
@DanielFischer Do you think it would be correct to say "characteristic functions of dyadic cubes" or "characteristic functions on dyadic cubes"?
@JohnDoe That depends on what you mean. Probably it's the first, $\chi_A$ is the characteristic function of $A$. You'd use "on" if the dyadic cube is the domain, and it's the characteristic function of some subset of the cube.
20:26
@DanielFischer Understood.
@DanielFischer So if $\chi_{A}$ is the characteristic function of $A$ and $A$ is in $B$, you could say $\chi_{A}$ is the characterisitic functions of $A$ in $B$?
Doesn't sound good actually.
@JohnDoe You could say that, but it would be unusual. Normally, the domain is not mentioned at all, since it is inferred from the context.
20:41
@DanielFischer in single variable calculus if $G'(s) =g(s) $ then $G(x) = \int_a^x g(s) ds$ , what about its extension to multiple calculus sir ?
@Theorem The generalisation of the fundamental theorem of calculus is Stokes' theorem (the version by Cartan, for differential forms).
21:00
@DanielFischer Can you check mathoverflow.net/questions/196294/… , if it makes sense . :)
@DanielFischer In addition to that paper I mentioned, the Maplesoft website also claims that $W_{0}(z)$ and $W_{-1}(z)$ share a branch point at $z= - \frac{1}{e}$ with $W_{1}(z)$. But I still don't see how that could be possible considering that by choice of closure, $W_{1}(-\frac{1}{e}) \ne -1$.
24
Q: What's a less offensive substitute for "Rep-Whores"

KyleMitThis is a frequently thrown around term on Internet forums in general and Stack Exchange specifically. Although it conveys a lot of meaning, I'd much prefer a phrase with a less offensive origin. Urban Dictionary defines a "rep-whore" as: A person who is obsessed with their status on an int...

21:24
@RandomVariable I don't see it either. $h\colon w \mapsto w e^w$ attains the value $-e^{-1}$ infinitely often. Say in $w_k,\; k \in \mathbb{N}$, where $w_0 = -1$. Then $h'(w_k) \neq 0$ for $k\geqslant 1$, and so $h$ is invertible in a neighbourhood of $w_k$ for $k > 0$, and that gives branches of $W$ defined in neighbourhoods of $-e^{-1}$.
21:39
@DanielFischer I noticed that the "shrink wrap" issue with MathJax still hasn't been fixed. One of the MathJax developers seemed to say he knew how to fix it. But it must be of very low priority.
@RandomVariable I thought the status was that Davide didn't know how to fix it in MathJax, it would need SE to set some width in the HTML or so. But I may have misunderstood, or missed later developments.
Hi!
I have a question about cryptography
is there any use of matrix reduction in cryptography?
@DanielFischer Yeah, I think the fix would be just for the profile page, not a general fix. Maybe it didn't work.
@infinitesimal You might like those quotes I think
“It’s not over until you win.” :-)
22:03
I wanted to change my name and use ಶ್ರೀನಿವಾಸ ರಾಮಾನುಜನ್. It looks like I cannot do that.
@Chris'ssis haha
22:27
Thanks @Chris'ssis :D
22:59
Can somebody think of a nice approach for calculating the last 10 digits of a arbitrary large tetration of 3s? I'm really stuck :)
23:18
@Chris'ssis no one could ping you unless they were replying to another comment
@robjohn The system doesn't even allow you to use it.
@Chris'ssis that's a +1 for the system :-p
@Hippalectryon J'étais déjà parti..:)
23:50
@commitingtothechallenge did I get the 200000000th post on this chat?
@CommitingtoaChallenge

« first day (1653 days earlier)      last day (3663 days later) »