« first day (3692 days earlier)      last day (1322 days later) » 

3:20 AM
Hmm ... It doesn't have rotational symmetry about the $z$-axis, so it's only a rotating shape.
 
4:07 AM
Can anybody explain how the normalization coefficients calculated?
I have recently given a paper which had 12 shifts in september and 8 shifts in january....I found
[this](https://jeemain.nta.nic.in/WebInfo/Handler/FileHandler.ashx?i=File&ii=97&iii=Y) . But this looks like BS....it doesnt give substantial information about comparing candidates across different shifts
 
4:38 AM
bounty 200 rep^
 
 
2 hours later…
6:31 AM
I need help in solving this question !image
 
7:26 AM
@SarGe use the formula to express sin(3x) in terms of sin(x) and (sin(x))^3 and you get a telescopic series
 
@TedShifrin $x^{2/3}+y^{2/3}+z^{2/3}=1$
 
@MartinSleziak what I said exactly matches with the answer on that page
 
Thank you @TheTestosteroneFanatic and @MartinSleziak for your concern.
 
@TheTestosteroneFanatic yes, it details your suggestion, but SarGe was given that link in another forum is what Martin was trying to point out.
 
@SarGe As a side note, you can use Approach0 and many other methods to search among the posts on Mathematics.
Relevant post on meta: How to search on this site?
 
8:25 AM
Is there a coordinate free way to show that if $(V, \omega)$ is a vector space with a skew symmetric non degenerate bilinear form, then $\wedge^{n} \omega \neq 0$ ?
 
@MartinSleziak Actually, I'm using mobile site and Approach0 isn't supported on mobile. I tried searching on Google but didn't find anything constructive.
 
That's a bit unfortunate. Still, I guess you'll be able to get to desktop eventually. Can SearchOnMath be used on mobile devices?
 
@SayanChattopadhyay Yes but the coordinate free definition of wedge is just as much work.
Start by writing down a formula for $\omega^n(X_1, Y_1, \cdots, X_n, Y_n)$
It's best to choose $X_1, Y_1, \cdots, X_n, Y_n$ to be a symplectic basis for $\omega$, in which case the end result will be Pfaffian of the matrix of $\omega$ wrt this basis.
That is, the matrix $(\omega(X_i, Y_j))_{i, j}$
And Pfaffian is square-root of the determinant, which is nonzero by nondegeneracy, and thus...
 
Anyway, further discussion of Approach0 (and searching in general) might be more suitable in the searching chatroom. (I have asked there some details on how things work on mobile devices - I have never used any mobile device, so I do not have any experience with that.)
 
@BalarkaSen Okay, let me do that, makes sense. And also on the half lives half die thing, I think I have figured out this commutative diagram using PD, once I am free from this category theory intensive symplectic geometry HW, I will write it down
 
8:38 AM
OK cool
 
Lmfao this guy is 5head
brilliant
 
9:00 AM
Can anybody explain how the normalization coefficients calculated?
I have recently given a paper which had 12 shifts in september and 8 shifts in january....I found
[this](https://jeemain.nta.nic.in/WebInfo/Handler/FileHandler.ashx?i=File&ii=97&iii=Y) . But this looks like BS....it doesnt give substantial information about comparing candidates across different shifts
$$\Huge{\mathrm{HELP}}$$
 
9:38 AM
@LeakyNun That's not a wooden shield
 
Lol
 
Is complex geometry as boring as my prof makes is sound ?
 
What are they teaching in complex geometry
 
For now not much, we went over multidimensional holomorphic functions and the Cauchy formula
 
Cool. I actually don't know much about this
Multivariable holomorphicity is quite different from one-variable holomorphicity
 
9:44 AM
Is it ?
 
Yeah, let's see if I can tell you why
$f : \Bbb C^2 \setminus \{(0, 0)\} \to \Bbb C$ be a holomorphic function of two variables. Then I claim $(0, 0)$ is a removable singularity, $f$ always extends to a holomorphic function on all of $\Bbb C^2$.
 
What about $z, z' \mapsto 1/zz'$ ?
 
so the singularities can only be codimension 1?
@Astyx it's singular on x-axis and y-axis
 
@Astyx That has singularities along $z = 0$ or $z' = 0$
@LeakyNun Yes, that's correct
 
Indeed
 
9:47 AM
But anyway let's try to prove this
 
why did I say x-axis and y-axis
@BalarkaSen do or do not, there is no try
 
Lol
Ok, so $f(z, w)$ is a function of two variables. Fix some $w \neq 0$ a complex number, and consider $g(z) = f(z, w)$, the restriction of $f$ to a "horizontal" slice.
This is of course also holomorphic as a function $g : \Bbb C \to \Bbb C$
Hm I am not sure I want this, give me a moment.
Yeah OK. I can write $g(z) = \sum_{n = 0}^\infty a_n z^n$ by Taylor expanding near $z = 0$ (because $w \neq 0$ was fixed). Here the Taylor coefficients $a_n = a_n(w)$ only depends on $w$.
Explicitly, $a_n(w) = \frac{1}{2\pi i} \int_{|z| = 1} f(z, w)/z^{n+1} \, dz$, yes?
 
Is there a simple example of a coalgebra which is not a bialgebra? I only have an example of an algebra which is not a coalgebra
Actually what I'm looking for is a coalgebra without group-like elements (or a proof that no such things exist), and a coalgebra which is not a bialgebra seems like a good place to start from
 
10:04 AM
@BalarkaSen Yep
 
@Astyx I realized this doesn't work, but some modification does. Suppose that $f : \Bbb C^2 \setminus \Bbb D_r(0) \times \Bbb D_r(0) \to \Bbb C$ was a holomorphic function, where $r > 0$ is anything.
 
@BalarkaSen why doesn't it work
 
So instead of removing a point I am removing a polydisk nbhd of the point.
 
@AlessandroCodenotti $K[X]$?
 
@LeakyNun I mean, my argument wasn't working. The fact I am trying to prove still is; my solution is to prove a more general fact.
 
10:11 AM
@LeakyNun Is this an example for the first or the second of my questions? In any case I don't see why it works
 
why doesn't your argument work @BalarkaSen
 
What is my argument
I didn't write it yet
 
your taylor series lol
 
I don't understand
 
it has a taylor series therefore it is holomorphic
 
10:13 AM
Sorry, I do not understand. What is $a_n(0)$?
 
$\frac1{2\pi i} \oint f(z,0) / z \ dz$
 
Why does that exist (you meant $z^{n+1}$)
 
because it only invokes $f$ for values outside the origin
 
Ok, because you're integrating over $|z| = 1$.
You still need to prove $a_n(w)$ is holomorphic, but that's probably easy by Morera's theorem.
$\int_{|z| = 1, |w| = 1} f(z, w)/z^{n+1} dz dw$
Yeah, that's zero because $f(z, w)/z^{n+1}$ is holomorphic in $w$ for all $z$ along $|z| = 1$, so the integral over $w$ vanishes.
 
@AlessandroCodenotti what's an algebra that's not a coalgebra?
 
10:19 AM
@LeakyNun Ok, thanks. So $\sum_{n = 0}^\infty a_n(w) z^n$ does seem to define an analytic function on $\Bbb C^2$ which agrees with $f$ on $\Bbb C^2 \setminus \{(0, 0)\}$
 
great
 
@Astyx So there you go.
 
Which coalgebra/algebra structure do you have in mind for $K[X]$? Because with the usual algebra structure and the coalgebra given by $\Delta(X^n)=\sum X^i\otimes X^{n-i}$ and $\varepsilon(X^n)=1$ iff $n=0$ and $0$ otherwise it looks like a bialgebra to me
 
Yeah I never remember this proof, I don't know why.
 
@LeakyNun $M_n(D)$ where $D$ is a nonzero division ring which is not a field
 
10:20 AM
@AlessandroCodenotti yeah forget my example
 
Cheers, that makes sense
 
But actually note that this same proof works to show that any holomorphic function $f : \Bbb C^2 \setminus \Bbb D_r(0) \times \Bbb D_r(0) \to \Bbb C$ extends to $f : \Bbb C^2 \to \Bbb C$
 
oh no
all my rings are commutative
 
You do the same argument but with Laurent series, and argue that the negative terms vanish.
Super surprising stuff. This is called Hartog's extension if you want a name to look up
 
@LeakyNun rip
It's easy to see that this is not a coalgebra because if you try to build a counit $\varepsilon\colon M_n(D)\to K$ you run into issues by looking at its kernel
 
10:26 AM
@AlessandroCodenotti so basically if $C$ is a coalgebra then $\operatorname{Hom}(C,V)$ is a monoid for every $V$?
 
@Alessandro Can't you always dualize a finite-dimensional algebra? If $A$ is an algebra over a field $K$, look at $\text{Hom}(A, K)$.
 
@BalarkaSen Yes and that's also a coalgebra but I don't see your point
 
That's automatically a bialgebra is it? Because you can take product of two functionals... Hm.
 
@BalarkaSen yes
@LeakyNun I guess so for nonzero $V$ because you should be able to use the counit of $C$ to build the identity of the Hom set
 
why do you need nonzero
 
10:32 AM
Ok I guess a monoid with one element is still a monoid
 
@Alessandro I have an example, but it might be overkill
 
Go for it
 
Actually I am not sure anymore. Consider the coalgebra defined as follows. $A = \Bbb Q[1] \oplus \Bbb Q[x]$, with co-unit map projecting to the first factor, and co-product defined by $\Delta(x) = 1 \otimes x + x \otimes 1$.
This is a valid coalgebra right? Does it have a compatible algebra structure?
 
Wait $\Bbb Q[1]$ is $\Bbb Q$ or am I drunk?
 
Yeah, I am specifying the generators in every factor.
 
10:39 AM
Oh ok
Let me think about it then
 
It's just $\Bbb Q \oplus \Bbb Q$
$1$ and $x$ are the generators in the respective grades
I didn't mean polynomial rings
 
And the 2020 award for the worst possible notation goes to... :P
 
Sorry :(
I am saying because $\Bbb Q[x]/(x^2)$ feels like the natural algebra structure but it's not compatible with $\Delta$
 
Hm ok let me see
@BalarkaSen Wait what's $\Delta(1)$?
 
That has to be $1$, right?
I could be dead wrong about this example, I am getting confused.
 
10:47 AM
I'm getting confused by the notation I think
No ok it has to be $1$
 
Yeah because projection to $1$ is the counit. It's very confusing to write.
 
So $\varepsilon(x)=x$?
I'm a bit suspicious because in finite dimension the dual of a coalgebra is an algebra (whose dual is the starting coalgebra), and I thought that by identifying the coalgebra with its dual you get a bialgebra in this way, but now that I think about it I don't know if the operations are compatible
 
I think $\varepsilon(a + bx) = a$. Does that work? Let me see, one moment.
So $\varepsilon(x) = 0$
$\Delta(x) = 1 \otimes x + x \otimes 1$, $\text{id} \otimes \varepsilon$ of that is $0 + x = x$, like we want
Yeah seems to work
$\varepsilon : \Bbb Q_1 \oplus \Bbb Q_x \to \Bbb Q_1$ is projection to first factor, like I said.
 
Can we write $1$ as $(1,0)$ and $x$ as $(0,1)$ please
 
Ok haha sorry
That gets too cumbersome. I think the reason nobody cares about coalgebras is how its equally natural as algebras but so much harder to write compared to them
 
10:57 AM
"nobody cares about coalgebras" doesn't sound very accurate (unfortunately)
Anyway why do you say your example doesn't have an algebra structure?
 
So I don't know, under an algebra structure, what would $x^2$ be? $\Delta(x^2) = (1 \otimes x + x \otimes 1)^2 = (1 \otimes x)(1 \otimes x) + (1 \otimes x)(x \otimes 1) + (x \otimes 1)(1 \otimes x) + (x \otimes 1)^2$ and $(a \otimes b)(c \otimes d) = (ac) \otimes (bd)$ is the forced algebra structure on $A \otimes A$, right?
So I get $\Delta(x^2) = 1 \otimes x^2 + 2x \otimes x + x^2 \otimes 1$
 
Could you suggest anything to search which is (maybe even slightly) related to "symmetries of a space determines the space"
 
Note that $x^2 = 0$, the natural algebra structure on $\Bbb Q_1 \oplus \Bbb Q_x = \Bbb Q[x]/(x^2)$, does not work because $x \otimes x$ is not a torsion element.
What are the possibilities?
We have to try $x^2 = a + bx$ in general
I can always cook up a $y$ such that $y^2 = c$ some constant, by completing the square
Eg $y = x - b/2$
Then $\Delta(y^2) = \Delta(c) = c$ whereas $\Delta(y^2) = 1 \otimes c + 2y \otimes y + c \otimes 1$?
Um
No what is $\Delta(1)$ again? $1 \otimes 1$, right?
That's what it has to be
 
@BalarkaSen Hm wait I'm not convinced here
Why do you have $\Delta(x^2)=(\Delta(x))^2$
 
Because we assumed the algebra structure is compatible with the coalgebra structure, no?
Isn't that what it means
 
11:08 AM
We want $\Delta(xy)=\sum x_1y_1\otimes x_2y_2$
Where $\Delta(x)=x_1\otimes x_2$ is the awful notation everybody uses for $\Delta(x)=\sum_{i=0}^n x_{i1}\otimes x_{i2}$
 
Comultiplication $A \to A \otimes A$ is a homomorphism of algebras, that's what you're saying, yeah?
That implies $\Delta(x^2) = (\Delta(x))^2$, no?
 
Hm ok I guess it does
Makes sense
 
Or put $x = y$ in your notation
$x_{01} = 1, x_{11} = x, x_{02} = x, x_{12} = 1$
So basically you have to prove that $x^2 = c$ and $\Delta(x^2) = 1 \otimes x^2 + 2x\otimes x + x^2 \otimes 1$ are not consistent conditions
Because the only algebra structures on our beast are $\Bbb Q[x]/(x^2 - c)$, upto isomorphism.
 
Hm ok, your example does seem to work to me now, but I'm still suspicious for some reason
 
I claim that my example is, give or take, the reason that $S^2$ is not a $H$-space :)
 
11:17 AM
So if we take this coalgebra and put the standard algebra structure on its dual (and then identify it with the dual), what do we get?
 
I was confused at the beginning because graded bialgebra is different from just bialgebra (and the graded info is what distinguishes $S^2$ from $S^3$ say)
But this is ok
 
@BalarkaSen Curiously I believe there are places where the coalgebra is much easier to work with
 
@AlessandroCodenotti Thats a good question. So we need to understand where $x^2$ goes to in the dual
@MikeMiller I'd listen if you have examples off the top of your head
I find it very difficult to write coalgebras
 
Ok so $x$ goes to $x^\ast$ as usual
 
So let's call $A = (\Bbb Q_1 \oplus \Bbb Q_x, \varepsilon, \Delta)$ as we defined it. $m : A^* \times A^* \to (A \times A)^* \to A^*$ gives $m(x^*, x^*)(a) = (x, x)^*(\Delta(a))$
for any $a \in A$
 
11:26 AM
Right
Well this is a thing in $\Bbb Q\otimes\Bbb Q$ as written to be very pedantic but then we just multiply the two pieces
 
So if $a = \alpha + \beta x$, $\Delta(a) = 1 \otimes (\alpha + \beta x) + (\alpha + \beta x) \otimes 1 = 1\otimes\alpha + \alpha \otimes 1 + 1 \otimes \beta x + \beta x \otimes 1$
$(x, x)^*$ of this thing is uh
 
@BalarkaSen Steenrod algebra is famously better written in terms of the coproduct. Milnor has a basis in which the dual algebra to Steenrod is polynomial
Since you usually don't want to pass to the dual algebra you instead talk about the coproduct of things in the Milnor basis
Whereas the Steenrod algebra itself is complicated
 
Pretty sure it's $2\beta$ or something like this, and $x^*(a) = \beta$, so $(x^*)^2 = 2x^*$
Something of this sort
 
Yeah that seems sensible
 
11:31 AM
@MikeMiller Hrm ok, I am not familiar with this though.
 
Unfortunately this example still has a group-like element
@Mike do you happen to know off the top of your head whether any coalgebra has a group-like element? This is obvious for bialgebras, so I was trying to look at examples of coalgebras that are not bialgebras, but with little success so far
 
@BalarkaSen Just think of it like this. The Steenrod algebra has all these Adem relations with binomials etc. Seems pretty complicated, even if you write it in terms of the generators Sq^{2^i}. But surprisingly the dual of this complicated algebra is free, so the coproduct is easy to write down (in the right basis).
@AlessandroCodenotti Why is this obvious for bialgebras?
 
Isn't the unit of the algebra always group like?
 
Oh sure
Maybe try a cofree coalgebra
 
11:46 AM
Uhm I need to look up what that is, let me see
 
As an example of something which isn't a bialgebra I mean.
It probably has grouplikes
You could also look in Sweedler's book
 
So if I'm understanding correctly the comultiplication of the cofree coalgebra extends that of the tensor coalgebra? In that case there is at least a group like
@MikeMiller The one just called "Hopf Algebras"?
 
@MikeMiller Oh ok bizarre
 
I have to leave for lunch, be back soon
 
@AlessandroCodenotti Yeah but I thought you wanted examples of coalgebras which aren't bialgebras
 
11:53 AM
I gave him one for that already I think
He wants no grouplike elements
To which, shrug
 
Yeah the thing is that I want an example of a coalgebra with no group-like elements (or a proof that this is impossible), so coalgebras which are not bialgebras seemed like a natural place to start from, but the two examples I've seen so far both have group-like elements
 
I remember the time coalgebras really clicked for me is when I was thinking about topology of divisor complements in $E \times E$ where $E$ is an elliptic curve
Something silly, but yeah
 
Meh/10
 
 
3 hours later…
2:38 PM
is the following true
Let $M$ be a Riemannian manifold
Let $\omega \in \Omega^1(M)$
Then there exists a unique smooth vector field $Y$ on $M$
sucht that for any smooth vector field $X$
we get $\omega(X)=<X,Y>$
How can I prove it without using coordinates
 
"The natural-medicine guru Andrew Weil claims that plant extracts are effective because they operate on multiple systems..." apparently Weil's american counterpart dabbles in natural-medicine and not in algebraic geometry
 
Smoothness you should check in coordinates
Existence of a unique (possibly discontinuous) vector field Y with that property follows from linear algebra
If you know that a vector field is smooth iff <X, Y> is smooth for all smooth X, then this follows from the fact that omega(X) is smooth for all smooth X
 
@EdwardEvans Curing germs by sheaves of stalks
 
ha
actually this is about the effectiveness of Saint-John's-Wort
 
wort is that
 
2:51 PM
some natural remedy to depressive symptoms
also nice puns
 
Hm I see
 
ANT 2 and Modular Forms will be running concurrently, but ANT 2 will be asynchronously run as an online lecture course, while Modular Forms will be run at university with standard lectures
which is the dream
change of topic
maybe this time I won't fail modular forms rofl
 
@EdwardEvans You better not fail them this time. Those forms count on you.
 
ha
The lecturer this year has some weird space of modular forms named after him and proved some stuff with Gross and Zagier
 
3:21 PM
@MikeMiller how would you prove the converse of a vector field $Y$ is smooth on $M$ if and only if $<X,Y>$ is smooth for all $X$
 
In charts
In the end a map is smooth iff it's smooth in every chart. That's the definition. So you shouldn't be surprised that you have to talk about charts somewhere to check something is smooth...
 
 
2 hours later…
5:46 PM
if $f,g:M\rightarrow \mathbb{R}$ and $f$ is smooth, then $g$ is smooth
right?
 
you forgot to tell us how $f$ and $g$ are related
 
Sorry, I meant to add that $fg=c$, where c is a constant
if f is non-zero then surely it holds
 
If $c=0$ it's false
 
If $c\neq0$, then $f$ is everywhere non-zero and $g=c/f$ is smooth
 
G'day, @Mike, @Thor, @orientablesurface.
 
5:51 PM
hi
 
hi Ted
 
okay, the second question is: given a riemannian manifold $(M,g)$, then expressing the metric tensor locally i.e. $g_{ij}$, we may invert the matrix $[g_{ij}]^{-1}=[g^{ij}]$, each $g^{ij}:M\rightarrow \mathbb{R}$ is smooth, right?
hello @TedShifrin
 
@orientablesurface: How do you show the entries of $A^{-1}$ are smooth functions in the entries of $A$?
 
$g_{ij}$ is not a function on $M$
but this is a detail you can fix
 
$g_{ij}$ is a function on $M$
 
5:54 PM
Not literally, @orientablesurface. What's the correct statement?
 
"expressing the metric tensor locally" doesn't take place on M
 
Huh? @MikeM
 
$g_{ij}(p)= <\frac{\partial}{\partial{x}^i|_p},\frac{partial}{\partial{x}^j}|_p>=$ $<\frac{\partial}{\partial{x}^i},\frac{partial}{\partial{x}^j}>(p)$
 
what are $x^i,x^j$?
 
coordinates on a chart
 
5:58 PM
and those coordinate functions naturally live on R^n, not M. :)
 
so their domain is?
 
open subset of a manifold
@Thorgott
 
whatever
 
yeah, so it only makes sense to talk about $g_{ij}$ on that open subset as well
 
I can't compete with Mike's pedantry.
 
5:59 PM
oh yeah, ofcourse
would $g^{ij}$ be a function $U\rightarrow \mathbb{R}$ as well?
 
Well, Mike's point is that once you choose a chart, everything is a function on an open subset of $\Bbb R^n$, not the open subset of the manifold. :D
 
oh, I see
Write a book on Riemannian Geometry, @TedShifrin please :P
 
Sorry, @orientablesurface ... just the curves and surfaces book. I'm not about to make my lecture notes for the graduate stuff accessible to the public.
 
ahh I see, was worth a try anyways :P
 
I always thought the $x_i$'s are functions from an open subset of $M$
 
6:04 PM
Right.
 
They are @LeakyNun
 
$x^i$ are
 
Hi, a @Balarka!
 
Hi @Ted
Gotta get dinner, be back soon
 
 
3 hours later…
8:48 PM
quick question: I know in English $\chi'$ is pronounced "chi prime". But what about $\chi''$? Is it just "chi two prime" or something else? I'm not a native speaker.
 
1. it's "primed" not "prime"; 2. I would say "chi double primed" or "chi primed primed"
 
chi tick tick
 
@LeakyNun I have always said "prime"
 
Me too
I think I alternate between primed and prime
Lol
 
The prime symbol ′, double prime symbol ″, triple prime symbol ‴, and quadruple prime symbol ⁗ are used to designate units and for other purposes in mathematics, science, linguistics and music. Although the characters differ little in appearance from those of the apostrophe (or the single or double quotation mark), the appropriate uses of the prime symbol are different. However, an apostrophe is often used in place of the prime (and a double quote in place of the double prime), due to the lack of the prime and double prime symbols either on commodity keyboards. These substitutions would not normally...
I stand corrected
but now you make me curious: where did I get this "primed" from if you also sometimes say "primed"?
 
8:55 PM
I have never said primed
But I assume it comes from thinking of it as the symbol being primed
(not that this actually makes much sense in this context)
 
and what sense does "prime" make?
> The name "prime" is something of a metonymy. Through the early part of the 20th century, the notation x′ was read as "x prime" not because it was an x followed by a "prime symbol", but because it was the first in the series that continued with x″ ("x second") and x‴ ("x third"). It was only later, in the 1950s and 1960s, that the term "prime" began to be applied to the apostrophe-like symbol itself. Although it is now more common to pronounce x″ and x‴ as "x double prime" and "x triple prime", these are still sometimes pronounced in the old manner as "x second" and "x third".[citation nee
 
9:29 PM
thanks for the feedback!
 
9:54 PM
I've never heard of ”primed” in a mathematical application. Painting, yes.
 

« first day (3692 days earlier)      last day (1322 days later) »