« first day (2357 days earlier)      last day (2958 days later) » 

18:00
arbitrary = $\forall$
I get what they were trying to say
But, still... :P
MSE just made my browser crash
Loading too many questions; or too much mathjax?
None of that
How, then?
It just decided : "I'm done with this"
18:04
Oh :P
And I got an error popup telling me to stop procrastinating on this chat
But here I am again
Never give in to the errors!
Stop procrastinating, @Astyx.
Errors would even have prevented me from annoying @Ted
Hi @Ted
How are you @Ted ?
Salut, Astyx
18:09
Hi @Ted
Hi @Balarka
@Astyx: Any progress on that linear algebra question?
Yes, I found out I needed to read the question again to understand it, which I will do now
@TedShifrin Been reading a thing or two on moving frames.
Are you sure it's not three?
18:17
Hah, right, three.
OK ... any comments, impressions?
There are 7 portable mini suites (a.k.a. cages) in a row at the Paws and Claws Holiday Pet Resort. They are neatly labeled with their guests' names. There are 3 poodles and 4 tabbies. How many ways can the "suites" be arranged if:
a) there are no restrictions.


b) cats and dogs must alternate.


c) dogs must be next to each other.


d) dogs must be next to each other and cats must be next to each other.
hi@TedShifrin
Okay, so you consider both the row subspace and line subspace to be in $\Bbb R^n$ (ie not in $M_{n,1}(\Bbb R)$ and $M_{1,m}(\Bbb R)$ respectively)
@Kasmir: I'll let you do that with other people.
@TedShifrin iv done part a -c just d) is unclear =p
and thanks :D
18:20
Ah, right, the row space is a subspace of $\Bbb R^m$ and the column space is a subspace of $\Bbb R^m$, @Astyx. [We think of the row vectors as vectors, not transposed vectors.]
(Recall that I said you could think of it as the image of the transpose map.)
@Kasmir: Think of the three dogs as one clump and the four cats as another clump. Where can you put the clumps?
Cats are evil.
1 cat in the middle and the other groups will be on left or right so thats 2 ways
<--- likes cats
@TedShifrin Haven't progressed a lot; got a fever. But I started reading, so that counts for something. I can define $\omega_i$ and $\omega_{ij}$, currently reading ahead.
Sounds like you are packaging the Ist and the IInd fundamental form.
but the other way of thinking can be translated as this 1100011
where 0 is dog
1 is cat
so the dogs are next to each other and so is the cats
18:23
Well, must all cats be together, or only pairs, @Kasmir? The question is ambiguous.
@Balarka: You always get a fever when you do math :P
thats what i dont know
@Balarka: The $\omega_{ij}$ have way more than just second fundamental form. That comes from the $\omega_{i3}$ only.
If you died, they would eat your corpse.
@TedShifrin Fever started yesterday.
in that case Ted would the answer be 4*3! * 3!
18:24
And the implication should go the other way
1111000 and 0001111
1 is cat and 0 is dog
I'm not sure, @Balarka :)
If the cat died, you would eat its corpse?
@Kasmir: Can you distinguish the cats from one another? If so, I say the answer is $2\cdot 3!\cdot 4!$.
btw @TedShifrin why in your lecture when you say "ask yourself" they all say self
18:26
Well, if I say, ask Kasmir why this works, they'd say, "Kasmir, why does this work?" So instead they say "self." :P It's a silly joke, but they like it.
La nuit, tous les chats sont gris
Ah, yeah, agreed that only $\omega_{i3}$ carry the information of the shape operator.
Non, @Astyx, il y en a qui sont noirs.
Are you trying to imply I'd be healthy if I stay away from math?
ah okay =p
18:27
But, @Balarka, you agree that this is very geometric: you're seeing how one basis element twists towards another as you move in a certain direction.
Right.
@Balarka: I think you're taking a converse of what I said.
Wikipedia seems to be overly accurate on this topic @Ted link
1111000 and 0001111
1 is cat and 0 is dog
cat must be in the middle since there are 4 of them, so 4 ways to chose witch cat in the middle and 3! ways to arrange to rest of cats and 3 ! ways to arrange the dogs
Eh bien, @Astyx. C'est bien intéressant.
18:28
so it is 4*3! * 3! why did you get 4! *3! *2!
You said I get fever whenever I do math. And I am sure you want my general well-being, so...
So I said math $\implies$ fever. Your statement up a few lines was the converse :P
@Kasmir: Are we putting all 4 cats together? I'm still confused by your words.
Well, that means not doing math is a necessary, albeit not sufficient, condition for not getting a fever.
I'm lost, @Balarka. Back to geometry.
18:30
the way that all the cats and the dogs must be in a row is where to either first put all cats followed by dogs
or all dogs followed by cats
1111000 and 0001111
1 is cat and 0 is dog
like that
so the middle one must be cat
OK, so if the cats go first, you get $4!\cdot 3!$. If the dogs go first you get $3!\cdot 4!$. So I get $2\cdot 3!\cdot 4!$.
Forget about the middle one.
okay ._. thank you
i dont know why my thinking of middle place ruined the answer
It should come out the same.
We've started reduction btw @Ted
You forgot to double because of the 0-1 switch, @Kasmir.
18:32
Hi chat.
Not so interesting, but OK, @Astyx.
yepp thats what i missed :D
hi @Semiclassical
So you understand, @Astyx, that I'm asking you to determine when those two isomorphisms are inverses of one another.
Hi @Semiclassic.
I do (at least I hope so)
@Semi Hi
First, as I said yesterday, you should understand why they're isomorphisms.
18:34
Time to figure out if there's anything interesting about the characteristic polynomial of the correct matrix.
I saw your comment earlier, @Semiclassic, although of course I have no idea what that changes.
Is it immediately obvious from the definition that $\omega_{13}, \omega_{23}$ live in $T^*M$, without thinking of them as $\Bbb{II}(-, \mathbf{e}_i)$?
I'll give this some thinking tomorrow, I have a lot to do right now :)
Well, it changes what quartic polynomial I'm interested in. (Rather drastically, it seems.)
Sure, @Astyx. It's not my fault you're always requesting problems :P
18:36
Plus, one thing I know for sure about the general 4-by-4 matrix considered in the paper is that it should only be 'solvable' when certain conditions are fulfilled. So the hope is to find something interesting about the characteristic polynomial, considered as a Riemann surface in an appropriate sense, in these special cases.
Sure, @Balarka. All the $\omega_{ij}$ are $1$-forms; they're just components of $de_i$ with respect to the orthonormal basis. (There's officially a tensor product going on here, of course.)
OK, @Semiclassic.
But, uh, I need to actually find something special first :/
But not all $\omega_{ij}$'s live in $TM$, do they? They are elements of $T^*\Bbb R^3$ along $M$.
Ok, you are saying it's $-de_3 \cdot e_i$, which obv lives in T^*M
No, the $e_i$ are all functions on $M$, so these are all $1$-forms on $M$.
(For abstract Riemannian manifolds, we need to alter the discussion, but let's let you master this first. :) )
Ah, right. I was a bit confuzzled about the embedding.
18:40
Here's a question I should be able to answer, though, just based on linear algebra.
You're allowed the occasional confuzlement, @Balarka.
Thanks for being patient with me.
I can be dopey sometimes
If $f(t)=\det(A+Bt)$ for matrices $A,B$ with $B$ diagonal, what can be said about $A,B$ if $f(t)$ is even?
Obviously one has $\det(A+Bt)=\det(A-Bt)$ but I'm not sure if there's anything further.
@Semiclassic: Let's first take the case that $B$ is invertible and turn it into the identity by multiplying by $B^{-1}$ (not $1/B$!).
So then what do you know about symmetric functions of the eigenvalues of $B^{-1}A$?
Pretty sure $B$ is always invertible here (and diagonal) so that's easy enough.
Not a lot, though I probably should.
Something something symmetric polynomials?
18:44
You can tell me what the various coefficients of the characteristic polynomial are.
Ahah. Yeah, I think I'm seeing it.
$\det (A-tI) = t^4 - t^3 (...) + t^2 (...) - t (...) + \det A$.
So $f'(0)=0$ iff the trace is zero.
Nope.
Trace is the coefficient of $-t^3$.
derp.
yeah. coefficient of -t is the sum of the triple products.
18:46
Right.
Key thing is that one needs to compute $B^{-1}A$ and look at the eigenvalues.
That gives the coefficients explicitly, and in particular the conditions for their vanishing.
Okay, thanks.
I'm not sure there's a nice way to see the eigenvalues of $B^{-1}A$ unless $B$ is a scalar multiple of the identity.
Also, note we have a sign change from what you started with.
(Doesn't matter since the char poly will turn out even.)
Right.
Yeah. $B$ is diagonal here, but that's not quite enough.
However, $B$ turns out to be block diagonal in this case as well. So that helps.
It would be nice if you knew $A$ and $B$ commuted :P
You mean $A$ is block diagonal?
If they commuted, the problem would be trivial.
Bah, yeah. I got mixed up on which one I'm calling the t-coefficient.
18:49
Gotcha.
Anyhow, go think and ignore me.
@TedShifrin So the point is that your embedding has trivial normal bundle, otherwise you couldn't have done this.
If [A,B]=0, then $i\dfrac{d}{dt}\Psi=(A+Bt)\Psi$ could be done just from the matrix exponential since the matrix on the RHS would commute with itself at all times.
Couldn't have done what?
@Balarka: In general, if not, the second fundamental form is normal-bundle-valued.
Ohhh, sure, @Semiclassic.
So the non-commutativity of A,B is essential to the story, for better or worse.
Make those forms on $M$, I meant.
Ahh, makes sense
18:50
What I'm doing here is trading $i\frac{d}{dt}$ for $\epsilon$ and studying the resulting eigenvalue equation.
A coin is tossed 8 times.
a) How many different outcomes are possible? 2^8


b) How many different outcomes have exactly 5 heads? what they mean by that ?
No, this is all local, @Balarka, anyhow. You don't get a global $II$ unless you make it normal-bundle valued.
We're taking a local trivialization of the frame bundle.
The idea, roughly speaking, is that one can compute things about the full time-dependent equation by knowing the characteristic polynomial for the adiabatic problem (the above replacement).
Yeah, sure. We don't need to do things globally to compute stuff like curvature, so yeah.
@Kasmir: Out of the $2^8$ possible outcomes, how many of them have precisely 5 heads and 3 tails (in some order)?
18:51
I suppose these forms would in general be bundle-valued?
isint this probability ?
this combinatoric stuff is driving me crazy
And, appropriately, the adiabatic approximation is to time and energy what the semiclassical approximation is to position and momentum :)
No, no, @Balarka. But you would get the second fundamental form by taking $\sum \omega_{i3}\otimes\omega_i\otimes e_3$. (Similarly for higher codimension.)
(name drop)
No probability yet. Just counting, @Kasmir.
18:53
OK, got it. I'll get back to the concrete stuff now.
Probability is just a certain kind of counting problem, really
I have no objection to your trying to understand it carefully, @Balarka.
Ok i need more thinking :D
Hey everyone! I have a question that didn't get attention (probably because it links to another question) It's really stupid and it's driving me nuts, I'm desperate can anyone help? It's: math.stackexchange.com/questions/2100482/…
well, discrete probability anyways.
18:54
@John11: Isn't there a hypothesis that the sequence is decreasing (or non-increasing)?
Of course, the word 'just' in such a statement is always a bit problematic :)
@TedShifrin do you mean the sequence of terms or the sequence of partial sums?
The sequence of $a_n$'s you start with.
Also you say there are $n$ terms, but there are clearly $2n$ terms.
yes the sequence of terms is decreasing
So $a_{2n}\le a_i$ for all $i=1,2,\dots,2n$. Therefore, adding ...
18:55
from n to 2n there are n terms I meant
well n+1 actually but that's what I don't get
Gotcha. I missed that.
You should be doing the sum $i=n+1$ to $2n$, anyhow.
You're right. Basically you want to double the series (and adjust the first term).
oh you mean I should take epsilon/2 instead of epsilon and move the 2 factor?
Right.
awesome! Thank you so much! I have all my confidence shaken just before the exams and just assume I'm wrong
Chill out, @John11 :)
18:59
Hi @Ted
Hi @Alessandro
What is a principal minor of a martix ?
Determinant of a $k\times k$ submatrix where you use the same rows & columns.
haha @TedShifrin I'm trying! :D
I'm sure you'll do fine, @John11. Good luck!
19:00
don't you "cut" it out from the upper left corner if it's a principal minor?
Nope, @Alessandro, not necessarily.
I've just spend an hour or so trying to solve a question because it has bad terminology
Maybe you have a different definition in Italy (like saying $0\in\Bbb N$ ... ugh).
I love my life
@TedShifrin Thank you!
19:01
Relax, Astyx.
I'm pretty we sure we always assumed that principal minors are those you get starting from the upper left corner, for example we'd say that a matrix is positive definite iff it's principal minors are positive
I am perfectly calm :p
must just be a different notation
It's just that now the question sums up to a rather trivial induction proof
Anyway, supper awaits me
Later
A coin is tossed 8 times.
How many different outcomes have at least 2 heads ? 2^8 - ( 1+8 )
is this logical ?
19:03
@Alessandro: But most people say that the coefficients of the characteristic polynomial are the sums of the principal minors, and this includes all the ones I described.
Bon appétit, @Astyx.
i counted how many had 8 tails and 7 tails instead and deleted from total
You're counting by counting the complement, @Kasmir.
Correct.
:D
i think am getting a grip on this stuff :D
thanks Ted !
How many different outcomes have at most 4 heads?
1+8+28+56+70
Seems right.
= 163 but i feel like this should be half of 256
or what is am missing
19:08
Why should it be half?
if we we think of complemnts
0 pairs with 8
For that to be true, you'd need the number of different outcomes with at most 4 heads to be the same as the number of different outcomes with at least 4 heads.
1 -7 , 2-6 , 3-5 , 4-4
So whom does 4 pair with?
Your intuition works fine when 8 is odd.
ohhhhhh
yes yes :D
I see my mistake :)
@Semiclassical did not understand that :(
19:11
Can I have a sanity check?
@TedShifrin Hm, I see, I'm not sure how would we call those
You fail, @Ali.
Suppose $\Bbb Z_p$ is an integral domain, then $p$ is prime
I said this statement is false
The number of different outcomes with at most 4 heads is the same as the number of different outcomes with at least 4 tails. (If you've got 8 coins and you flip 3 heads, then the rest are tails.)
but it is supposedly true
19:11
it's true
You're wrong
I am confused, what if p is a prime power?
@Alessandro: See Wiki. I'm right, you're wrong. Yours are called leading principal minors.
try an example. Z/4
2x2 = 0
generalize
regarding my question from yesterday I noticed that actually the trivial topology is enough to make addition and multiplication continuous functions so probably the correct question to ask is whether the euclidean is the coarsest topology with which $\Bbb R$ is a topological field @Ted
19:12
Z/4 is a field however so is it not an integral domain?
@Semiclassical got it ! thanks :) its easy to get lost on this topic , i was more fine doing calculus
But the probability shouldn't depend on heads versus tails. The two outcomes are equally likely. So therefore switching tails to heads in that what I just said shouldn't change a thing.
Z/4 is not a field
i just gave you a zero divisor
Oh ugh @Alessandro :P
@Ali: $\Bbb F_4 \ne \Bbb Z/4$.
oh shit
19:13
Uh huh.
it's not a field @Ali, $GF(p^n)$ is not $\mathbb{Z}/p^n\Bbb Z$ unless $n=1$
Maybe wait a bit for algebraic geometry :D
the field of order 4 actually is obtained as a degree 2 extension of F2.
Welcome to today's derp club.
19:14
I am the president
Also known as the derp-ident.
Thank you
(of the club, not of the united states; lest people identify me with Trump)
@BalarkaSen I was just about to say something like that.
Although I think Trump is the derp-ident.
I'd call Trump the derp-ident, but I associate that with people who can actually realize they're wrong.
5
19:16
He's far, far worse than derp.
Termp.
This is not random starring.
That was pure gold right there.
@Alessandro: You see my link on minors?
yep, I'm confused now, I'll have to check the terminology in my linear algebra book because I'm pretty sure we call principal the leading principal and I have no idea how we call the principal
Glad to have helped.
19:19
English is a weird language
or maybe Italian is, but I'm biased :P
We just need to agree on the mathematics.
Thanks @Ted
The fact that $\mathbb{F}_q$ is unique up to isomorphism just gets weirder as you do more field theory
It does?
Yeah, when I first saw it, it was just "Ah, okay, cool"
But the implications are massive
19:25
It totally just meants finite fields are boring ;)
+1
Every field has characteristic 0.
Quite the contrary I would say
pops popcorn
You're an algebraist, your opinion doesn't count
Lol
19:26
what about infinite fields with nonzero characteristic @Balarka
@BalarkaSen :(
Meh, @Alessandro
I'll set up my own chatroom for algebraist, with blackjack and hookers
Everything of that sort is like $\overline{\Bbb F_p}$, isn't it?
@Krijn and fields of characteristic 2!
19:27
I'm also an algebraist
@BalarkaSen Function fields?
I just prefer groups
I don't really know why but it seems that everyone hates characteristic 2...
Well positive and negative opinions become the same
@AlessandroCodenotti It's always a massive pain in the ass when primes are even
19:27
@Krijn Oh yeah, right.
This was like the third result when you google "blackjack and hookers"
3
A boy has 3 red , 3 yellow and 2 green marbles. In how many ways can the boy arrange the marbles in a line if:
a) Marbles of the same color are indistinguishable?

11! / ( 3!*3!*2!)

b) All marbles have different sizes?
where they have different sizes they are all different now right ?
so the color wont matter anymore?
Knowing how Futurama's staff is full of mathematicians, maybe they came up with that in a similar situation as ours...
i've lost my marbles
This boy has too many existential questions
19:29
So it would seem, @Kasmir.
LOL @Ali
isn't there also a theorem called Futurama Theorem?
@AliCaglayan -.-
@AlessandroCodenotti yes
thank you Ted i hope there were more of you Ted :D
Most of the world thinks one is more than desired.
19:30
^
@TedShifrin how comes that?
Too many alarm clocks for me
what? Do you actually sleep once in a while?
Alarm clocks? Moving frames set off alarms, @Balarka?
Does the number of Teds necessarily have to be integral...?
19:32
@Steamy: If only you knew my Hausdorff dimension!
@Alessandro Sometimes
Any one know if @robjohn come's around these parts once in awhile?
not very often, but he does
He does
Occasionally he does. None of Pedro, robjohn, anon is here as often as yore.
19:34
I've seen robjohn quite often lately
Hey all, hey all :-)
hi Kari
@TedShifrin I've been trying to find him, as he doesn't frequent the Mod's chat. I need to ask him if he can attach a "wedge of cheese" to the top of my gravatar... He's very skilled at doing such things. (It's a Green Bay Packers thing).
One of my best friends is a cheesehead.
How's it hangin, @TedShifrin?
19:37
This quesiton came into my mind : What is the goal of MSE ? Is it only to make it easy for people to discuss maths, or is it to build some kind of huge pile of questions/answers to make it easy for people to answer questions they might have ?
Both and more
Both, @Astyx.
OK, Kari, and you?
A week into my last semester of lectures so I'm doing alright
Congrats, Kari. Then you'll be all grown up and ... :D
and ... please let me stay in uni forever
:-)
19:39
Are we having fun?
Balarka is making fun of me :(
You are, across the pond, @MikeM.
Because a lot of questions have a lot (too much) in common. For instance I've seen dozens of diophantine equations that are not all the same, but the method for solving them is the same. It seems a bit silly to keep all those questions since all it does imho is drown more relevant questions/answers in a flood of questions that are fundamentaly the same
@Balarka is learning geometry at long last :P
@Astyx: I used to get annoyed when people would interrupt my hint-discussion with the OP and post complete solutions. Now I just don't care.
Does anyone know a reference of a proof that a connected polytope (not necessarily convex) in R^n has a finite number of vertices?
19:40
@MikeMiller yes
I don't do well at the computations in geometry.
I'm also looking for a geometric definition of vertices
that does not rely on convexity
You should like this approach way better for surfaces, @Balarka.
Of course, sometimes the first fundamental form isn't diagonal ... but ...
@TedShifrin sorry for full answers sometimes
$$\huge\overbrace{\left(\ddot{\stackrel{\quad>}{\frown}}\right)}_{\begin{align}--- \ \ \end{align}}$$
@Astyx Look for meta math posts about "abstract duplicates". Indeed, you have good company with respect to your question, and so "a duplicate" has come to mean "not necessarily identical to"....
19:41
I've given my share of full answers, @SimpleArt. I'm complaining particularly about a situation where I'm giving hints and the OP is working it out.
oh
Well ye who is posting full solution probably does not see you
@TedShifrin Let's see. You're assuming that I am good with forms, which I doubt :)
Will do, thanks for your answers
Practice makes ... pseudo-perfect.
@Astyx your welcome
19:42
@barto Let's say it is defined as a finite union of convex polytopes. The problem is I don't know how to formalize the geometric intuition about vertices.
Sometimes I think I should have just done the crash course in forms when I had the chance... but then I didn't know multivariable calculus, so, well.
@Ted A little. Depends on the minute. I had some good conversations with Gordon and Roberts.
@Astyx Glad to help. And yes, it does get tiring to read the same-old same old as answers to the same-old same hold, especially when there are six duplicate questions with answers, listed to the right of the question/answers.
Exactly
Are there some sort of reading groups for mathematics?
19:46
@amWhy Lmao, true
@Krijn Let's make one!
I has question. @TedShifrin Do you think its appropriate to guide the OP all the way through the problem using only comments?
Appropriate?
Yeah
As in, good etiquette.
According to my definition of good teaching, yes
Ok, just wondering
Often such discussions get moved to chat though
19:49
Let the OP write the solution and I'll check it.
@TedShifrin :D I shall try to start doing that
Here's a weird proof of a simple fact I came up with.
The Euler characteristic of the n-torus is 0, because the n-torus is a Lagrangian submanifold of $\Bbb R^2n$ and $\Bbb R^2n$ has trivial cohomology.
Say, does anyone have introductory problems that are easy enough to be interesting to a broad range of people?
but with a lesson to be learned in the problem?
Introductory to what ?
That sounds very, very broad
19:59
Idk
Introductory to anything that doesn't require a lot of pre-reqs
For example, I asked this question:

« first day (2357 days earlier)      last day (2958 days later) »