« first day (2592 days earlier)      last day (2724 days later) » 

04:00
Ok, so there is a closed form, but I am not sure what's the abstract reason that the series manipulation works. More investigation needed:
Let $H_0=0$
\begin{align}
\sum_{n=1}^{k}\frac{H_n}{n} & = \sum_{n=1}^k\frac{\left(H_{n-1}+\frac{1}{n}\right)}{n}\\
& = \sum_{n=1}^k\frac{H_{n-1}}{n}+\frac{1}{n^2}\\
& = \sum_{n=1}^k\frac{H_{n-1}}{n}+\sum_{n=1}^k\frac{1}{n^2}\\
& = \sum_{n=1}^k\frac{H_{n-1}}{n}+H_{k,2}\\
& = \sum_{n=1}^{k-1}\frac{(H_k-H_n)}{n}+H_{k,2}\\
& = \sum_{n=1}^{k-1}\frac{H_k}{n}-\sum_{n=1}^{k-1}\frac{H_n}{n}+H_{k,2}\\
& = H_kH_{k-1}-\sum_{n=1}^{k}\frac{H_n}{n}+\frac{H_k}{k}+H_{k,2}\\
\therefore 2\sum_{n=1}^{k}\frac{H_n}{n} & = H_kH_{k-1}+\frac{H_k}{k}+H_{k,2}\\
Secret allways posts this random stuff that I dont understand =p
@Ted Don't worry, my sleep schedule was bad going in!
I totally ignore it.
It would be better if somehow this notion of summing partial sums by rows and summing partial sums by columns can be represented geometrically
Well, you're a whole other kettle of fish, Demonark.
04:01
Also Ted, I dunno why but the voice in my head when reading your messages is different from your real one
LOL ... you shouldn't be listening to voices in your head.
Oh I always rely on them for advice... That's why my life choices have been stellar!
rolls 10 3/4 eyes
prove that every group of order 4 is isomorphic to either z/2 x z/2 or z/4 , by analysing the multiplication table
so is this a proof by constraction?
Not even a natural number of eyes anymore...
04:04
I wouldn't say so. I would say: Two cases — is there an element of order 4 or not, @Kasmir.
it's a proof by "how many ways can I make the op table for this group without something going wrong"
z/2 x z/2 is {(0,0) , (0,1) , (1,0) , (1,1) } right?
each element appears exactly once in every row/column, and you know what's in the identity row and identity column
not a lot of options left
Yes.
I guess Semiclassic has a brute force argument, too.
well, he said to do it by analyzing the multiplication table
04:06
I find that most proofs of abstract algebra so far are easy to understand but hard to write down in a good way
like I dont know how to start
Semiclassic, do you think I listen to him?
omg :o
where did this came from ? :o
never in my life expect this from Ted -.-
I'm still bleeding, so ... expect the worst.
04:08
hmm that is not good at all
you did not get medication for that tooth?
Best way to do this is to find subgroups of $S_4$ :P
quiet, you.
I mean in principle that's not too bad a way of doing things... But this exercise is about multiplication tables so don't use Cayley
(a,b) x (c ,d ) = (ac, bd) ?
04:10
that's one possibility.
Yes, @Kasmir: You defined the product group in an exercise 9 hours ago.
ehm then how is this a group
(0,0) * (a,b) = (0,0) allways
There is no 0. There's e.
You were writing things with addition for your $\Bbb Z_2\times\Bbb Z_2$.
Be careful about what your group operation is.
oups right
><
This is what happens sleepless at 6 AM.
04:12
especially when you're having to do isomorphism stuff
okay :D
very important that you have different notation for different operations
so the group z/2 x z/2 is bascilly each element is its own identity
Own inverse?
i meant
04:13
Every element other than the identity has order 2.
order of each element is at most 2
><
right. and that means every element is its own inverse
If $x^2=e$ and $x\ne e$, then what is $x^{-1}$?
I'm not sure you need x != e for that. I'd count e^{-1}=e.
(though not as an element of order 2, of course)
04:15
Yeah, yeah, yeah. I was writing down order 2, thank you very much.
It was a big deal with my students when I taught algebra to remember that you had to say $x^k = e$ and $x^\ell \ne e$ for $0<\ell<k$.
Okay iv written down both tables
Are you sure there can be only two?
so far no =p am trying to do things with the knowdge we have
i mean what we went thru in class
04:18
if your answer is "the problem said there's only two" then you haven't actually justified it
Yes I know =p
I need to show why that is the case
well the order of the elements must be less than or equal to 4
Oh actually now that we're talking about algebra
cant be any element that is 3
but am not allowed to assume that =p
part of what makes this tricky: if your elements are labelled as e,a,b,c then you can swap the labels on a,b without changing the group structure
Right. But you can prove it easily enough.
04:21
so when they say there's only one such table, it's true only up to isomorphism
So there's the sign of a permutation that you can define using that $\prod_{i < j} \frac{\pi(i) - \pi(j)}{i-j}$, I'm wondering how to show that this matches up with the transpositions thing
I've never seen anything like that, Demonark.
So is there only one non-trivial homomorphism from $S_n$ into $\{-1,1\}$?
yeah, that's bizarre to me
for one because it doesn't seem like it's +-1, so why would one call it a sign?
I would use determinant of the permutation matrix. Why in the world is that it?
04:23
Is proof by contradiction an option here?
Yes, Demonark, but that takes work to show. You need to know $A_n$ is simple for $n\ge 5$. And I suppose we need to work on $n=3,4$.
For what, Kasmir?
I mean like i try to show that assume ord (a) = 3 and arrive at contradicton
Oh, for that.
because it will break rules from group structure
So how would you do that?
04:25
I mean questions like this drive me crazy, it is easy to see why they r true but hard to come up with something to write
by constraction
but thats not the idea here I assume
@Semi I think it's that in this product you'd cancel out each $i-j$ with some $\pi(m) - \pi(n)$ up to a sign later
It's only easy to see once you know Lagrange's Theorem later. :)
Yeah for $m = \pi^{-1}(i)$ and $n = \pi^{-1}(j)$ (or swapped) that'd do it
So you need another element $b$ that's not in the list $e,a,a^2$.
That hardly defines a unique $\pi$, Demonark, not that I'm paying attention.
Oh, wait, $\pi$ is fixed.
Oh wait okay I think I can prove that $A_n$ is the only subgroup of index 2
Maybe
04:29
That's not easy.
@TedShifrin @Semiclassical So the first row and colum is allways predetermned
I think I see an argument that doesn't involve simplicity
I have 3x3 table to fill
Transpositions are conjugate, right?
04:30
both groups are abelian one is cyclic and other is not
Things of identical disjoint cycle structure form a conjugacy class, Demonark.
How do we know they're abelian and that it's a both?
Z/n is cyclic and abelian for all n
Sure, so what?
You're still assuming the result.
Ehm yes just thinking outloud ><
If there's an element of order 4, we get the cyclic group. But if not ...
04:34
if not a^2 = e
that forces a=a'
Why in the world can you say that?
because if order of a were 3
a^3 =e
a*a^3 = a
but order of a = 4
so ... ?
WTF?
no no thinking in wrong direction
a^3 =e
means order of a <= 3
well, let's assume 3.
(although you should be able to show that if $a\ne e$)
04:37
\begin{align}
\sum_{n=1}^{\infty}\frac{H_n}{n} & = \lim_{k\to \infty}\frac{1}{2}\left(H_kH_{k-1}+\frac{H_k}{k}+H_{k,2}\right)\\
& = \frac{1}{2}\left(\lim_{k\to \infty}\left(H_{k-1}^2+\frac{H_{k-1}}{k}\right)+\lim_{k\to \infty}\frac{H_k}{k}+\lim_{k\to \infty}H_{k,2}\right)\\
& = \frac{1}{2}\left(\lim_{k\to \infty}\left((\gamma+\ln (k-1))^2+\frac{\gamma+\ln (k-1)}{k}\right)+\lim_{k\to \infty}\frac{\gamma+\ln k}{k}+\zeta(2)\right)\\
& = \frac{1}{2}\left(\lim_{k\to \infty}(\gamma+\ln (k-1))^2+0+0+\zeta(2)\right)\\
Secret: This really is annoying. Must you do it?
If you're talking to someone, so be it, but just to spam us all ?
I truly will have to put you on ignore.
Ok fine, I will do it in my room
Thank you.
hmm
assume ord(a) =3
a^3 =e
let b =a^2
ab =e
I already gave you a hint up there ^^^^^ ... That won't get you anywhere.
04:43
ae = a, a^2 =b , ab =e , this forces ac=c
in order for that row not to have same element twice
but if ac=c
a=e
OK, I see. This is equivalent to what I was suggesting.
So you can't have an element of order 3. So everybody has order 2 (other than the identity).
yes
a^2 =b^2=c^2 =e
ij =k
04:45
i meant to say that
product of 2 different elements yields the third
that is the only way to have a group structure
if we assume order of a = 4
then G = <a>
Then why is that isomorphic to $\Bbb Z_2\times\Bbb Z_2$?
Yeah, element of order 4 is done immediately.
hmm because we can find a function that is 1-1 and surjective
You know that's not good enough.
04:47
f(xy)= f(x) f(y)
but need to show this for all elements
they are not that many but
How are you writing $\Bbb Z_2\times\Bbb Z_2$, the way you did earlier? Remember you used addition, not multiplication.
yes i did it right =p
I mean the structure is the same
3 elements of order 2
But you've been talking multiplication in our discussion.
So one group is with multiplication and the other is with addition.
Yes I know we are allowed to use multiplicative notation allways =p
both addition
04:50
Z/4 not z/4 *
I'm not worrying about that one.
okay
what did you mean by multiplication? =p
one group is with multiplication
You have $\{e,a,b,c\}$ and $\{(0,0),(1,0),(0,1),(1,1)\}$, multiplication in the first and addition in the second.
oh
You mean it that way =p
Oh.
Maybe you should switch to multiplicative notation for the second one?
I dunno what you want to do to finish.
04:53
Am gonna ask my teacher if we are allowed to use theorems we did not do in class
Anyhow, I'm leaving. Get some sleep.
because doing it this way when there are good theorems availble seems not good idea =><
Okay that might be a good idea
Ill figure this out tomorrow:D
@TedShifrin Thanks alot and good night! :)
You're gone now @Ted but the idea is that transpositions should map identically under the homomorphism, but since they generate S_n their images have to generate {-1,1}. So they map to -1
Anyway good night!
 
1 hour later…
06:06
To prove that each finite semigroup contains an element $x$ s.t. $x^2=x$ can we say that identity element is enough?
sorry then it would be a monoid
06:36
@RE60K Consider powers of some given element.
Yo @Tobias!
How's it going?
@Daminark Good. Doing the final reading through of my paper before putting it on the arXiv
Also as good as done with my application. I just need to go borrow a Windows computer to fill out some parts of it.
Awesome!
Why Windows though?
Well, Mac would work too. It needs to have a browser that supports the Adobe pdf plugin
06:55
Can someone here make sure I'm not high?

https://i.imgur.com/9gGJjFS.jpg

That second line should say b-qa-a right?
Yeah it should
Okay good, thanks.
I mean ofc you may still be high but you're also right
also note r = b - qa
so it's gotta be a typo
07:02
True, that should've made it clear that it was a typo.
@Daminark There. Paper submitted.
Found a new formula: $$\sum_{n=1}^k \frac{\sum_{m=1}^na_m}{b_n}=\sum_{n=1}^k \frac{a_n}{b_n}$$
Typo:
$$\sum_{n=1}^k \frac{\sum_{m=1}^na_m}{b_n}=\sum_{n=1}^k a_n(B_k-B_n)$$ where $B_k=\sum_{n=1}^k \frac{1}{b_n}$
This formula allow summation in terms of rows to be converted to columns
Proof: Write out the series in rows n, columns m, and note the terms form a upper right triangular structure
NB: May not work for infinite series
07:41
@KasmirKhaan I've scrolled through your conversation above regarding Z/4 and Z/2 x Z/2... I will refrain from leaving any comment.
08:00
Given two complex number $z_1 = \rho e^{i\theta_1}, z_2 = \rho e^{i\theta_2}$
consider the complex number $z(\alpha) = (1-\alpha)z_1 + \alpha z_2$
where $\alpha \in [0,1]$
what can I say about the magnitude of $z(\alpha)$?
can I find an explicit function?
I believe the function should have the following features
calling $\rho(\alpha) = |z(\alpha)|$
I'd say $\rho(0) = \rho(1) = \rho$ and for all $\alpha \in [0,1] - \frac{1}{2}$ we have $\rho \left( \frac{1}{2} \right) > \rho(\alpha)$
hi chat
hello all
how does one turn on mathjax on chat?
08:17
it never worked for me :/
hi @MAGO
thanks for confirming my observations
@skullpatrol I unsuccesfully already tried this
odd :-/
Its origin is very puzzling :^)
@SteamyRoot hi ! how are you?
Alive :P
You?
Made it so far too :P
Good thing I am not doing this at the last minute. Turns out the deadline for my application is a day before I thought it was
for your Phd paper?
09:09
@Liad For a Marie Curie fellowship
09:32
is it possible to invert the function $y = x + kx^2 + hx^4$ where $x \in [0,1]$ and $k > h > 0$?
I mean finding the inverse function
Try using that a function is invertible if and only if it is injective?
Guys, is the order of a differential equation $\geq0$ or strictly greater than 0?
It depends of the definition of differential equation you are managing.
my teacher doesn't really give any. I was wondering if there is a generally accepted def
I don't think it matters?
A "0-th order differential equation" doesn't involve any differentials
09:38
Usually it is $\geq 0$.
So whether that term exists or not, it's not interesting (in a course on differential equations)
uh, I was just curious, so it matters to me :P
Well, unless you are studying pseudodifferential operators it doesn't matter.
What is the meaning of $x/t\to r$ as $t\searrow 0$?
I found it in some notes without explanation about the $\searrow$ meaning.
it is the same than $t\to 0$? OO
I think it means the right-hand limit
at least, that's the notation we use
09:42
i.e: $t \to 0^+$?
Uhm okay, makes sense, thanks.
There's too many notations for that >.<
$$\lim_{x \to 0^+}, \lim_{\substack{x \to 0\\>}}, \lim_{x \searrow 0}, \lim_{x \downarrow 0}$$
10:15
@SteamyRoot although I have seen the last one in some places, predominantly, I see the first one.
It's likely to be somewhat "geographically" determined
I'd only ever seen the second one until, like, my 3rd year as a Bachelor student
10:28
I think the first one is usually read as "from the positive side...
Hi @robjohn
@SteamyRoot It must be because I've never seen the second form
@skullpatrol hey there. what's up?
Not too much @robjohn how are you doing?
 
1 hour later…
11:48
@mago what was the problem? It seems to work on most platforms.
@skullpatrol busy, but hoping that I can include more time to be back here.
12:07
Who wants to hear a joke?
sure @KrisWalker
What's the integral of $\frac{1}{\text{cabin}}$ with respect to $\text{cabin}$?
sorry, I've heard that one
Damn haha
I just found it and I want to try it on someone but nobody in my direct vicinity will know what I'm going on about.
Log cabin + C
12:13
Yup, you remembered the C :)
How do you tell an extrovert mathematician from an introvert mathematician?
If an introvert mathematician is talking to you, he'll be staring at his shoes.
If an extrovert mathematician is talking to you, he'll be staring at your shoes.
12:33
Hi, $$Q,P\in\{0,1\}[x], \text{gcd}(P(3),Q(3))=1\\\text{Is it true that : gcd}(P(x),Q(x))=1 ? $$
p
So, wait
$P$ and $Q$ are polynomials over, say, the integers, but just happen to have all coefficients $0$ or $1$ ?
Or are they polynomials in $\mathbb{Z}_2[x]$ ?
No, with all coefficients egal to 1 or 0.
$$P,Q \in\mathbb Z[x]$$
13:10
@Dattier a very interesting question.
13:20
@LeakyNun Hello !
:D
@LeakyNun you gonna be here for a while ? =p
yees :D
Gonna go out for 20 -30 mins
then I kinna need your help with some isomorphism stuff :D
I solved 2 exercices but
Dont feel good about how I solved them =p
@KasmirKhaan are you done with the Z4 and Z2 x Z2 question?
@KasmirKhaan send me via email
13:23
haha how did you know that :D
Did not put them on texstudio yet
6 hours ago, by Leaky Nun
@KasmirKhaan I've scrolled through your conversation above regarding Z/4 and Z/2 x Z/2... I will refrain from leaving any comment.
@KasmirKhaan then just send me what you wrote
just solved them on papper =P
@LeakyNun All rithty brb 30 mins max! :)
13:42
$$f (x) = x^5 + x^3 + x, \text{ calculate the derivative 8 ith in 0 of the reciprocal function of }f $$
@Dattier my conjecture is yes
@Dattier what is "derivative 8 ith in 0 of"?
Yes, I am agree with you
@LeakyNun
$$g^{(8)}(0), \text{ with } g^{(1)}(0)=g'(0)$$
do you mean the 8th derivative of the reciprocal function of $f$ evaluated at $0$?
but the reciprocal of $f$ is undefined at $0$
13:47
why ?
because $f(0) = 0$
f is croissant
but f'(0) not equal to 0
@Dattier ???
f'(0)=1
@Dattier tu as dit l'inverse de $f$
1/f
13:49
oui
f est localement inversible en 0 car la dérivée y est non nulle
deplus f est croissante donc inversible sur les réels
c'est quoi croissant?
increasing
oh, croitre
13:52
yes
oui
@Dattier ... do you mean 1/f or f^-1?
f^-1 n'est pas reciprocal... it's inverse
$$g=f^{-1}, g(f(x))=x$$
reciprocal is 1/f
sorry
Question: $f: \Omega \to \Bbb C$ is holomorphic s.t $f(\Omega) \subset \{\alpha z_0 : \alpha \in \Bbb R \}$ for some $z_0 \in \Bbb C$ .
i need to prove that $f$ is constant.
if $z=1$ , then if $f = u + v i $ , we have $u(z) + iv(z) = \alpha$ so by the Cauchy–Riemann equations we get that $v' = 0$ which implies $u' = 0$ so $f$ is constant .

in the general case, we have $u(z) = \alpha(z) x_0 , v(z) = \alpha(z) y_0$ where $z_0 = x_0 + y_0 i$. im not sure how to apply Cauchy–Riemann equations here. someone can help ?
13:56
@LeakyNun in french la réciproque is the invervse, and l'inverse is the reciprocal
@Dattier alors parlez en francais
$f(x)=x^5+x^3+x$ calculer la dérivée 8 e en 0 de la fonction réciproque de $f$.
On indiquera la méthode employée.

« first day (2592 days earlier)      last day (2724 days later) »