« first day (3458 days earlier)      last day (76 days later) » 

1:33 AM
I was wondering - would the power set of the set {{$1,2,3$}} contain $2^1$, $2^1∗2^3$ or just $2^3$ elements?
None of the above? What's the power set of {a}? Of {{a}}?
I mean, it's been a long time for me, but it's just the power set of a one-element set, right? That the one element is, itself, a set of three elements doesn't matter.
The power set of {$a$} would be {{}, {a}}, therefore contain $2^1$ elements, I reckon?
Me, too.
(I missed your first 2^1 on first reading--hence the strikethrough.)
I see that makes sense. So, it would be essentially equivalent to finding the power set of the set {$a$}?
1:38 AM
No problem ;D thanks!
2 hours later…
3:15 AM
I am fairly fresh in set theory, and I was wondering why is [$0.5, 7.5$] $∩ N$ ≠ $[1, 7]$?
Is it due to the reason that $0.5$ and $7.5$ are $\notin$ the set $[1, 7]$ and for [$0.5, 7.5$] $∩ N$ = $[1, 7]$ to be true, both sets must be subsets of each other which isn't the case here? Or is there a different reason?
3:52 AM
@Abwatts Can you enumerate the elements of $[0.5, 7.5] \cap \mathbb{N}$? (I think so.) Can you enumerate the elements of $[1,7]$? (I think not.)
(Assuming you're using the square brackets for continuous intervals in $\mathbb{R}$, at least.)
That makes sense. So, basically, because the interval [$1, 7$] is not restricted to only natural numbers, we will have some rationals in that interval too, which are obviously not natural numbers?
3 hours later…
7:02 AM
7:21 AM
A: coin toss and conditional probability

Demetrios Papakostasyes you are correct. P(EDGE)=0, so you cannot condition on that. This is just a trick question

@EdwardEvans wanna play?
2 hours later…
9:27 AM
Composing music is very hard when you have a anti-motif bias
You end up with numerous chunks of completely distinct melodies and you struggle to glue them together into a song
Basically, most western type and video game music are ABA, ABCA, or some variation of this to ensure loopability
and so it usually involve 1-2 motifs and the whole song will be just variation of these motifs
My issue is that I often end up with music like the figure shown above
where there are like 10 pieces of melodies, none of them pairwise equal
This means it is usually hard for me to compose a full music piece since it tends to birfrucate in the middle and produce a new melody
and hence you end up with motif-less music
Maths chat people only: This makes me wonder, is there a mathematical formula that detect motifs
10:02 AM
@Leaky I was learning a generalisation of the Dirichlet unit theorem lol
a very small generalisation but cool nonetheless
in particular there's an exact sequence

$$0 \to \mathcal{O}_K^\times \to \mathcal{O}_{K,S}^\times \to \bigoplus_{\mathfrak{p} \in S} K^\times/(\mathcal{O}_K)_{\mathfrak{p}}^\times \to \operatorname{Cl}(K) \to \operatorname{Cl}_S(K) \to 0$$

where $S$ is a finite set of primes of $K$ and this guy gives you $\mathcal{O}_{K,S}^{\times} \cong \mu(K) \oplus \Bbb Z^{\lvert S \rvert + r_1 + r_2 -1 }$
where the usual notations are the usual notations
2 hours later…
12:20 PM
Does anyone here study non-commutative algebra?
12:53 PM
@SolvingProblems I do
Hello Lukas, I posted a question on main.
Q: Finding the rank of the endomorphism algebra of a projective module over a non-commutative algebra

SolvingProblemsLet $A$ be a commutative ring, $B$ be a non-commutative $A$-algebra, $P$ be a finitely generated projective left $B$-module. Let $E=\text{End}_B(P)$ as a left $A$-module (by restricting from $B$). If I were to know that $E$ is in fact a free $A$-module (since $A$, and the map $f:A\to B$ defining...

Perhaps to rephrase the question:

Let $A$ be a commutative ring, $B$ be a non-commutative $A$-algebra, and $P$ be a finitely generated $B$-module, $P\subset B^n$ with known generators. I want to take $\text{End}_B(P)$ and restrict it to be an $A$-algebra, which I'll denote by $E$, where in my case it is now free as an $A$-module. I want to find the rank of $E$.
2 hours later…
3:25 PM
I'm interested in the simplest case of the edge-of-the-wedge theorem: If $f\colon\mathbb{C}\rightarrow\mathbb{C}$ is continuous and holomorphic on both half-planes, it is holomorphic everywhere.
It is claimed that this follows from Morera's theorem by splitting any contour up along the real axis into parts in the upper and lower half-plane, where they vanish by hypothesis. However, I don't understand how that works. because one still necessarily has parts of the contour along the real axis and I don't think the Cauchy Integral theorem necessarily holds on the boundary. What am I missing?
3:36 PM
> > We asked [composer Gyorgy Ligeti] about his musical "influences." We had read—in the New York Times—that he had been influenced by the new field of fractal geometry "whereby physical objects are conceived as consisting of smaller and smaller identical shapes." He said:
> "Yes, fractals are what I want to find in my music. They are the most complex of ornaments in the arts, like small sea horses, like the Alhambra where the walls are decorated with geometric ornaments of great minuteness and intricacy, or like the Irish Book of Kells, those marvelously decorated borders and capitals.
> The most complicated ornaments—perhaps not art, perhaps geometry. It is a very complex music, difficult to describe. I only want to give a metaphysic for my music. After all, music is not a science."
@Secret Thought you might find that interesting
I think this is the quote from the New York Times they're referring to:
> ''I do not use direct mathematical translation into my music, like Xenakis,'' Mr. Ligeti said. ''The influence is poetic: fractals are the most complex ornaments ever, in all the arts, like the Book of Kells or the Alhambra. They provide exactly what I want to discover in my own music, a kind of organic development.''
Hello, I'm learning pde and I'm trying to find an example for a function u that holds: for every parallelogram ABCD with parallel lines to the lines y=+-x, u(A)-u(B)+u(C)-u(D)=0 and u does not hold u_tt=4u_xx
@Secret I can relate to the problem of having lots of chunks of music and no coherent large song
When I'm entertaining myself on the piano I often just play them in a random order with strange transitions
(some of which are more natural than others)
(Well, random-ish, influenced by how natural the transition would be)
@Eran Is u a function of x and y, or of x and t?
@AkivaWeinberger: A function of x and t
4:00 PM
I will try posting this code here in chat to get myself motivated:
Clear[a, b, nn];
nn = 60;
a[n_] := Total[MoebiusMu[Divisors[n]]*Divisors[n]];
Monitor[a1 =
Table[Sum[Sum[a[GCD[m, r]], {m, r, n}]/r, {r, 2, n}], {n, 1,
nn}];, n]
g1 = ListLinePlot[a1, PlotStyle -> {Red, Thick}];
Monitor[a2 =
If[Sum[-Abs[Sum[a[GCD[m, k]], {m, k, n}]], {k, 2,
r}] >= -(n - 1), -Abs[Sum[a[GCD[m, r]], {m, r, n}]]/r,
0], {r, 2, n}], {n, 1, nn}];, n]
g2 = ListLinePlot[a2, PlotStyle -> {Thick}];
Show[g2, g1]
Marilyn vos Savant said that learning is boring. Therefore don't multitask. Multitasking is just a way of easing the feeling of boredom to make it feel like you are progressing although you actually are not.
Prove that the blue curve times a constant is greater than the red curve.
In other words/formulas:
$$a(n)=\sum\limits_{d \mid n} \mu(d)d$$
Show that:
$$\sum\limits_{r=2}^{n} \frac{\sum\limits_{m=r}^{n} a(\gcd (m,r))}{r} \geq c\underset{\sum\limits_{k=2}^{r} -\left|\sum\limits_{m=k}^{n} a(\gcd (m,k))\right|\geq -(n-1)}{\sum _{r=2}^n} -\frac{\left|\sum\limits_{m=r}^{n} a(\gcd (m,r))\right|}{r}$$
where $c$ is the constant.
The following is not well formulated, at least not yet. Under what conditions/variable bounds is a sequence that sums to 1 and consists of positive and negative terms bounded by c*Floor(Sqrt(Length(sequence)))?
4:18 PM
@Thorgott Take part(s) of your path parallel to, and arbitrarily close to, the real axis. You do know continuity (hence uniform continuity on a compact subset).
Solve that and then determine into what category, greater than or less than c*Floor(Sqrt(Length(sequence))), the right hand side falls into.
That is what I am going to do next, I am going to find those mentioned conditions/variable bounds.
I'm trying to prove that u:R^n-->R that holds the spherical mean value property also holds the ball mean value property. I understand the intuition but I'm having trouble proving it formally.
I thought about change of variables comparing the integral on the sphere to the integral on the ball
4:42 PM
So, say we are (wlog) looking a triangle $T$ with vertices $x,y,w$, the former two lying on the real axis and the latter lying in the upper half-plane. Then we look at the triangles $T_n$ with vertices $x+\frac{i}{n},y+\frac{i}{n},w$.
We have $\int_{[x+i/n,w]}f(z)dz=\int_0^1f(w+t(x+\frac{i}{n}-w))(x+\frac{i}{n}-w)dt\rightarrow\int_0^1f(w+t(x-w))(x-w)dt=\int_{[x,w]}f(z)dz$ by uniform continuity on some compact rectangle containing all these triangles.
But $\int_{[x+i/n,w]}f+\int_{[w,y+i/n]}f+\int_{[y+i/n,x+i/n]}f=\int_{T_n}f=0$ by Cauchy, hence $\int_Tf=\int_{[x,w]}f+\int_{[w,y]}f+\int_{[y,x
Ok, that works. Thanks.
5:00 PM
@Eran Are there such functions which are not harmonic?
5:57 PM
Polar coordinates work for that problem, but I feel like there should be a more direct argument for why integrating over the ball is the same as integrating over all the spheres of smaller radius and then the radius. Can't think of one right now though.
6:37 PM
I like your polar coordinates+Fubini approach
@AlessandroCodenotti @Thorgott Thanks!
6:53 PM
It feels like the relationship between integration on balls and integration on spheres (geometrically speaking) should be more intrinsic than an explicit parametrization, idk
Ted will probably say something about Stokes soon
Probably, but I want there to be a direct way with CoV/Fubini
Isn't Stokes's theorem such a way?
Buonasera @Alessandro
Ciao Lukas
7:01 PM
The only thing about Stoke's theorem is that it generalizes the FTC, so I have no clue, really
oh, right
see, that's how little I know about it
Ritter von Stokes
7:04 PM
hey @Edward still working through cyclotomic fields?
Hey @Lukas, I finished the chapter on cyclotomic fields and on localisations, gotta go back over it tho :P
There was a slightly generalised form of Dirichlet's Unit Theorem at the end which I liked
and posted above lol
ah, the S-unit theorem
Right, where the rank of the S-units is |S| + r_1 + r_2 - 1
a friend of mone wrote his bachelor thesis on something with S-units
7:06 PM
yeah, in the S-units there are only finitely many solutions to the equation $u+v=1$
ha wtf
and you can reduce some other diophantine finiteness results to that
Nice :)
also, did the cyclotomic proof of quadratic reciprocity
ah, my favorite proof
and the 2nd Ergänzungssatz was the Übungsaufgabe
but yeah it's very cool
7:08 PM
you need to work with $\Bbb Q(\zeta_8)$ for that iirc
Hmm maybe I did it wrong but I worked with $\Bbb Q(\zeta_p)$
as in, for (2/p)
and look at the decomposition of $2$ in that guy
there are probably many ways to do it
Yeah right
You look at the field $K := \Bbb Q\left(\sqrt{(-1)^{\frac{p-1}{2}}p} \right)$ and then the splitting number of $2$ is even iff $K \subset \Bbb Q(\zeta_p)^{Z_2}$ which is equivalent to $2$ being split completely in $K$, which happens when $d_K \equiv 1 \bmod 8$ and then that's equivalent to $p \equiv 1$ or $7 \bmod 8$
and that happens to be equivalent to $(2/p) = 1$
@Thorgott How would you prove it with polar coordinates?
the proof with $K=\Bbb Q(\zeta_8)$ goes like this. Every odd prime $p$ is unramified in $K$, so we have a Frobenius element in the Galois group which is just given by $\zeta_8 \mapsto \zeta_8^p$. We have a subgroup in the Galois group $\mathrm{Gal}(K/\Bbb Q) \cong (\Bbb Z/8)^\times$ given by $\overline{\pm 1}$, the fixed field $L$ of that is generated by $\zeta_8+\zeta_8^{-1} =\sqrt{2}$. Now the Frobenius element at $p$ fixes $L$ iff $p$ splits completely in $L=\Bbb Q(\sqrt{2})$
but $p$ splits splits completely iff the minimal polynomial of $\sqrt{2}$, i.e. $x^2-2$ splits over $\Bbb F_p$
thus $x^2 \equiv 2 \pmod{p}$ has a solution iff $p \equiv \pm 1 \pmod{8}$
I like both solutions :) @Edward
7:18 PM
Brilliant :D
and then our other exercise was to prove FLT for n=4 lol
which was just a couple of successive pythagorean triples and coprimality arguments
@EdwardEvans in L-functions we defined Artin L-functions today
so we're doing the more Galois/ANT-heavy stuff right now
bleh nice
Should've stayed in it but ah well
Define them for me if you want
can't do everything
Pretty much just write it down. E.g. in two dimensions you have $\int_{|x|<r}f=\int_0^rs\int_{-\pi}^{\pi}f(r\cos\varphi,r\sin\varphi)d\varphi ds=\int_0^r\int_{|x|=s}f$
It feels backwards though, I'm not satisfied with it
7:26 PM
okay so Rösner did motivate them really well. Start with the cyclotomic field $\Bbb Q(\zeta_n)/\Bbb Q$. The Galois group of that is isomorphic to $(\Bbb Z/n\Bbb Z)^\times$, so a one-dimensional representation of the Galois group corresponds in this case simply to a Dirichlet character. We should expect that the Artin L-function for such a representation is a Dirichlet L-function
we can start by looking at Dirichlet L-functions and reinterpreting everything in ANT/Galois terms to try and generalize it to a more general setting
if $\chi:(\Bbb Z/n\Bbb Z)^\times$ is a Dirichlet character, then the L function $L(\chi,s)$ has an Euler product $$L(\chi,s)=\prod_{p \not \mid n}(1-\chi(p)p^s)$$
now for the reinterpretation: viewing $(\Bbb Z/n\Bbb Z)^\times$ as the Galois group of $\Bbb Q(\zeta_n)/\Bbb Q$, we can interpret the primes in $\Bbb Z$ as prime ideals $(p)$, the condition $p \not \mid n$ is of course the fact that $(p)$ is unramified. For these unramified prime ideals, $\chi(p)$ is the evaluation of the representation at the Frobenius element of the prime ideal $(p)$
Now suppose that $L/K$ is a Galois extension of number fields and $\rho:G=\mathrm{Gal}(L/K) \to \mathrm{GL}_n(\Bbb C)$ is a representation. Then for all unramified prime ideals $\mathfrak{p}$ of $K$ we can do the following: choose any prime ideal $\mathfrak{P}$ above $\mathfrak{p}$ and evaluate $\rho$ at the Frobenius element $\mathrm{Frob}_{\mathfrak{P}/\mathfrak{p}}$, then the factor at $\mathfrak{p}$ is $\det(\mathrm{Id}-\rho(\mathrm{Frob}_{\mathfrak{P}/\mathfrak{p}})N(\mathfrak{p}))$,
i.e. the characteristic polynomial of $\rho(\mathrm{Frob}_{\mathfrak{P}/\mathfrak{p}})$ evaluated at th
If $\mathfrak{p}$ is ramified, then there's a slight modification we can make: we can still take a lift of the Frobenius in the residue Galois group, but that will only be well-defined up to an element in the inertia group. That's no problem if the inertia group acts trivially. We can now force the inertia group to act trivially by simply looking only at the invariants $V^{I_{\mathfrak{P}}$.
As the inertia group is normal in the decomposition group, this subspace carries an action by the decomposition group, so we take the characteistic polynomial of the action of the Frobenius on $V^{I_\ma
and then the product over all those local factors is the Artin L-function
7:41 PM
Nice man
Sounds super interesting
I can just read it all in the back of Neukirch instead I guess
we're proving Cebotarev next week
Nice, I only briefly glanced over that in Milne's notes
just says the density of split primes is $1/[L:K]$ or smth right?
or completely split primes?
completely split
Cool :P
the more general statement is that for any conjugacy class of size $n$ in a Galois group of an extension of number fields $L/K$, the density of those prime ideals whose Frobenius is in that conjugacy class is $\frac{n}{[L:K]}$
and you can recover the statement about completely split primes by looking at the conjugacy class of the idenitity
8:17 PM
Hi @Balarka
9:13 PM
@Lukas hey that's also cool
I think I've seen that under another Name's Density Theorem
in Milne's CFT
9:53 PM
Hi everyone!
I'm in a dilemma, unfortunately my question (math.stackexchange.com/questions/3518491/…) has been closed as a duplicate :( It is quite different from the other, though..
10:03 PM
on the face of it, "Is X true?" is really not that different from "How can X possibly be true?"
but all it comes down IMO to is that they're equal because 0.999... is ultimately just a different name for the same thing as 1.
With finite decimal expansions, you can't have the same number represented in two ways. But why must infinite expansions play by the same rule?
I want to check the existence of the below limits using the definition:

1. $\displaystyle{\lim_{x\rightarrow 0}\frac{x}{x}}$
2. $\displaystyle{\lim_{x\rightarrow 0}\frac{x}{|x|}}$

For that do we use the epsilon-delta definition, or the left and right limit at $0$ ?
It's akin to asking why "0" is the same as "what I get if I divide 1 by 10 an infinite number of times"
It's the same number, just reached in different ways.
@MaryStar You should use whichever is easier for you
10:19 PM
@Thorgott Isn't there a way that is more correctly for such exercises?
Well, mathematically speaking, there isn't. Both definitions are equivalent. Sometimes, one might be easier to use, at another time, the other might be easier to use, but neither is "more correct". However, if that's an exercise for a course and they explicitly ask you to use the definition and you have not yet shown the equivalence of these definitions, then, of course, you should stick to the one given to you.
10:37 PM
@Thorgott my textbook says that for a parametrization along a sphere of radius R centred along the origin and with an osculating circle of radius $\rho$ gives $\rho = R\cos\theta$, without proof (that I can see). I need some help finding out where this fact comes from if you've got a sec.
for some reason I'm getting $\rho = \frac{1}{R\cos\theta}$ which makes no sense
might have figured it out...
10:53 PM
For the second limit I think the definition with the left and right sided limit should be used. Which definition would you suggest at the first limit? @Thorgott
Hi team, following question might seem kinda dumb, but I'm honestly not computing this totally right. Basically I am implementing a program and need to round up numbers, pretty simple, right? So I need to somehow keep good track record on the integers but fail to do so. Here is the scenario:
a) 1/4 * 7 = 1.4 ~1
b) 4/5 * 7 = 5.6 ~6
So if I round up (5,6,7,8,9) and round down(1,2,3,4), then 1 plus 6 equals 7, so far so good.

Problem comes in when I need o split again 1.4 & 5.6.

So $1.4/2 = 0.7 , but if I round up 0.7 to 1 and then sum the two parts I get 2. Let's say we round first 1.4
Please apologize if this is some kids math problem, but thing is that I end up with 8 if I put up everything together and not 7 which was the original number. Am I rounding wrong?
11:20 PM
Hmmm should 5 be always rounded up or down?
11:46 PM
@MaryStar If you don't have any of the sequence-based definitions at your disposal, you can use epsilon-delta for the first one. Then using the first one, you can solve the second one with one-sided limits.
@psa Sorry, I don't know anything about osculating circles
no worries
11:58 PM
Say I want a discrete set of points on the sphere with a given maximum distance from anywhere on the sphere to the set and a given minimum distance between different points in the set
(So the discs of radius max cover the sphere, and the discs of radius min/2 are disjoint)
Given n (the number of points in the set), how would I randomly select a configuration of points uniformly from the space all possible configurations satisfying the above?

« first day (3458 days earlier)      last day (76 days later) »