« first day (3572 days earlier)      last day (1466 days later) » 

3:01 PM
You can find a draft on his homepage.
 
@Yai0Phah Can smell it from long distance.
I love the smell of new books.
 
Where do you live now?
 
@Yai0Phah Very embarrassing place to name.
@Yai0Phah This book is way too much more advance than ross probability.
 
If near a university, you can have a try to ask the librarian.
 
@WhyWhatWhereWhenHow I hope you don't live in Fucking
 
3:05 PM
@Yai0Phah But university library are private though it will take me hours to walk to get there.
I country is densely populated with university
 
Some libraries are public.
 
@BalarkaSen Well I don't live in Fu《king.
@Yai0Phah Then I have to go to national library of my country. Which is too far from my home. Not too far but take 1 to 2 hour by car.
At may be 60 kmph
 
Really? Flagging for naming a village in Austria?
Nuts
 
No, morally universities should be open.
 
I live close to Fucking 60 km say.
 
3:10 PM
I mean, open to all.
 
@Rudi_Birnbaum Lmao nice
 
I had to guide a visitor to there once. He was happy like a baby taking photos from the signs ...
 
@Yai0Phah I kinda don't know if I can go to my country's most prestigious university I will try it.
 
While the sense of humor of the locals is considerably less developped in that direction.
 
@Rudi_Birnbaum Why else would you visit Austria?!
 
3:12 PM
It is not necessary. You just go to the university to ask whether you are allowed to read books in the library.
 
To talk about group theory with me!!
 
You don't need to buy them. You can read there.
 
Yep. I hope to read all day.
I can only go there at weekend if available.
 
I wonder how it was when the first GI discovered the village ..
 
And read the hell out of that.
 
3:13 PM
Sometimes you can obtain a reader card.
 
@Rudi_Birnbaum Fair enough
Vienna is actually one of the best places in Europe for logic and set theory
There's the Kurt Gödel research center
 
Well I totally forget to register a reader card lmao.
 
@AlessandroCodenotti Really didn't know that, cool!
@AlessandroCodenotti any take on my last question?
on group theory
 
Which question? I'm afraid I didn't see it
@Rudi_Birnbaum Ah this one
 
You are studying set theory?
 
3:15 PM
@Yai0Phah I read books until everything gets blur. Which is kinda annoying you you have that problem?
 
@Yai0Phah yes, mostly
 
@AlessandroCodenotti yes
 
It's not true that groups are always built of smaller groups though
 
Interesting.
 
they are built from smaller groups, but in a sense vastly more general than just products or even semi-direct products
 
3:17 PM
@AlessandroCodenotti Yes sure, but thats not what I meant.
 
Sure sure, direct limit over f.g. subgroups
Sure, some quotients will be isomorphic to a subgroup, as in the case $G=H\times K$ that you mention, but there's no reason in general for a quotient to be a subgroup
 
I think I should come here when I finished my topology and by then you guys would have finished more than I did lol.
 
I cant see how $G/H$ is no-way a subgoup of G while $H$ is. and at the same time the product of the two forms G
@AlessandroCodenotti OK
Fine. Thats what I wanted to know
 
for abelian groups, quotients and subgroups are the same thing (up to isomorphy, of c ourse)
 
Bye guys I need some rest. Gonna have brunch.
 
3:19 PM
but that's just cause they have a nice product structure
 
@Thorgott Which subgroup of $\Bbb Z$ is isomorphic to $\Bbb Z/(5)$?
 
yeah but I want to understand non-abelians
 
finite*
 
I see group and assume "finite" for some reason
 
3:21 PM
Another thing: if $H$ is a normal subgroup of $G$ and there is an isomorphic subgroup of $G/H$ in $G$, is it then normal as well?
(I think thats about finite groups right?)
 
@Rudi_Birnbaum It's not true in general that $G\cong G/H\times H$, I think that's the problem here
 
@Thorgott lol
@AlessandroCodenotti Are there conditions when its true?
 
If $H$ is a normal subgroup of $G$ you have a short exact sequence $1\to H\to G\to G/H\to 1$, this sequence splits iff $G$ is a semidirect product of $H$ and $G/H$, but not all ses of groups split
 
thats an cyclic argument, right?
 
Take $S_3$, the unique order $3$-subgroup is normal and the quotient is cyclic of order $2$, but no order $2$ subgroup of $S_3$ is normal in it
 
3:23 PM
@Thorgott Ah OK!
 
@Rudi_Birnbaum What do you mean?
 
@AlessandroCodenotti you say iff $G$ is a semidirect product, I would be rather interested when ses split
 
I speak of s.e.s. only in abelian cases.
 
exactly in abelian cases?
 
Can we please stop the flagging of legitimate (but crass in other languages, I guess) on that city in Austria?
 
3:28 PM
It splts if and only if $G$ is a semidirect product of the other two, I'm not sure if it can be made any more explicit
There's the splitting lemma
 
OK I see!
 
@MikeMiller Thanks Mike
 
@Rudi_Birnbaum kconrad.math.uconn.edu/blurbs/grouptheory/splittinggp.pdf maybe you'll find those notes interesting
 
@AlessandroCodenotti Gracie mille!
 
4:04 PM
@Balarka Ok, say I have an open cover $\{U_{\alpha}\}$ of $B$ with charts $\varphi_{\alpha}$ and you give me smooth maps $\phi_{\alpha\beta}\colon U_{\alpha\beta}\rightarrow\mathrm{GL}(\mathbb{R}^k)$. We construct $\coprod_{\alpha}U_{\alpha}\times\mathbb{R}^k/\sim$ as before. Surely (read: probably (read: hopefully)) this is Hausdorff and second-countable. Now, fix indices $\alpha,\beta$, then $(\{\alpha\}\times(U_{\alpha\beta}\times\mathbb{R}^k)\cup\{\beta\}\times(U_{\alpha\beta}\times\mathbb{R}^k))/\sim$ should be an open set in $\coprod_{\alpha}U_{\alpha}\times\mathbb{R}^k/\sim$ and $(\{
I should probably be more explicit
 
Yeah that's the atlas pretty much. Please dont be more explicit
also compact subspaces of the vector bundle are fiberwise compact, yeah
not only that but also projection to the base is compact
 
4:30 PM
is that an iff?
Let $\pi\colon E\rightarrow B$ be a vector bundle. If $K\subseteq E$ is compact, $\pi(K)$ is compact as continuous image of a compact set and $K\cap\pi^{-1}(b)$ is compact as closed subset of a compact set, so that direction is clear.
 
@Thorgott Would you agree about continuity of the inverse if you restrict the codomain to just the bit lying above $U_{\alpha \beta}$ and the domain just the points that map to that
Because then the domain is just udentified with $U_{\alpha \beta} \times \Bbb R^n$
 
@Thorgott its an iff yes
 
@Mike I'm not sure what that means, if not what we already have. $\varphi_{\alpha}(U_{\alpha\beta})\times\mathbb{R}^k$ is just the bit lying above (the diffeomorphic copy of) $U_{\alpha\beta}$.
@Balarka is that non-trivial or another exercise
 
i leave it to you as an exercise :)
 
shouldntve asked
 
4:47 PM
it's a twisted version of "X x Y is compact iff X and Y are compact"
instead of $X \times Y$ you have $X \rtimes Y$
but $\rtimes$ is locally $\times$ so the same proof goes thru
proof by wobbles
either you take things as exercises or accept wobbles, you have two choices my man
 
@Thorgott I mean you have a space $E' = \sqcup_\alpha U_\alpha \times \Bbb R^k$ with a map to your bundle $E$. You have a map $p: E' \to E$ which is a continuous bijection. If you retrict to $p: p^{-1}(\pi^{-1}(U_{\alpha \beta})) \to \pi^{-1}(U_{\alpha \beta})$, do you agree that the inverse is continuous?
My claim is that the inverse in this case is not only continuous, it is something like obviously continuous.
Which gives that $p^{-1}: E' \to E$ is locally continuous whence continuous.
 
oh no, a Balarka is wobbling again.
 
Ze wobbles
(Look at the conductor's hands)
 
The ultimate handwave?
 
5:02 PM
Yup
Let's just say this piece was too much too Handel
 
Famous.
 
5:31 PM
@Mike that should just be $p\mapsto(\alpha,\varphi_{\alpha}^{-1}(p))$ followed by projection onto the quotient, which I would agree is continuous, but I don't see why one needs to restrict to $U_{\alpha\beta}$ rather than just $U_{\alpha}$ for this to work, so I may be missing something
 
Oh just $U_\alpha$ is fine
I got my notation confused
All I'm saying is that if you restrict your map to obvious opens it's clearly invertible
 
right, locally the map just looks like applying the local trivializations (up to working with equivalence classes instead, but they don't make an issue in this direction)
 
@BalarkaSen Well, it came out of "Methods of classical mechanics" by V.I. Arnold who was a mathematician, and he just said "write the differential equation" but not "solve," so that led to my suspicion that it might be a demonstration that good choice of coordinates can save you work.
Then again, I haven't done differential equations for years and some of the methods employed look like brain surgery to me.
 
Ah makes sense
Yeah Arnold can't be evil
 
@Balarka this is more subtle, the bundle locally looks like a product, but the subset doesn't necessarily look like a product of subsets
 
5:36 PM
There is no subtlety, the exact same argument goes through. You just cover segments by tubes and then refine
The compactness of the factors is used only upon projection
And locally you have a product so tube lemma is no problem
 
I don't even know what the tube lemma is lmao
 
That's what you use to prove $X, Y$ compact implies $X \times Y$ is compact!!
These are easy details you can just sit down and figure out on your own. There's no hidden subtle issue here
I would much rather you proceed with the differential topology than get bogged down in general topology
 
5:54 PM
Hello guys!!!
I am interested in knowing if this is a proof for this question:
0
Q: Prove or disprove: $\lvert\mathcal{R}\rvert=\lvert\mathcal{R}^{-1}\rvert$

manoooohProve or disprove: If $\mathcal{R}$ is a relation then $\lvert\mathcal{R}\rvert=\lvert\mathcal{R}^{-1}\rvert$. I think it is true but I do not know how to prove it. Facts: $\mathcal{R}^{-1}=\{(y,x)\mid(x,y)\in\mathcal{R}\}$. $\mathcal{R}\subseteq A\times B$. $\lvert\mathcal{R}\rvert\...

In that question some people told me that I have to construct a bijection, then prove it etcetera
But it seems that it can be solved in an "easier" way, which I present to you:
 
lol
 
huh, I just realized the proof for $X,Y$ compact implies $X\times Y$ compact we did in lecture was fallacious, yet I was buying it
 
$|R|=|\{(x,y)\in A\times B\mid x\in D_R\land y\in I_R\}|=|\{(y,x)\in B\times A\mid y\in I_R\land x\in D_R\}|=|R^{-1}|$, where $D_R$ is the domain and $I_R$ is the image of the relation $R$
 
Fucking hell
 
Do you think it is a valid proof? Thanks!!
 
6:00 PM
It's the same idea as what was suggested in the comments
your middle equality follows precisely because $(x,y)\mapsto(y,x)$ is a bijection
 
@Thorgott oh, I thought we didn't use that fact since $\land$ is commutative
 
@Thorgott Really? I don't know how to come up with a false proof of that
 
Do we?
 
well, that only tells you $|\{(x,y)\in A\times B|x\in D_R\land y\in I_R\}|=|\{(x,y)\in A\times B|y\in I_R\land x\in D_R\}|$, if we are to be pedantic
to switch the $(x,y)\in A\times B$ in the beginning to $(y,x)\in B\times A$ is precisely the same as noting $(x,y)\mapsto(y,x)$ is a bijection
 
Good to know. Thank you!!
 
6:05 PM
For any $y \in Y$, one may find an open cover of $X \times \{y\}$ and then pass to a finite subcover, giving you $V_{i,y}$, a finite family of open sets which covers the slice over $y$. Note that $\cup V_{i,y} \supset X \times U_y$ for an appropriate set $U_y$. One may thus choose a finite subcover of $U$ that covers $Y$; but that means that the corresponding finite collection of $V_{i,y}$ covers all of $X \times Y$.
It's a short enough argument I don't see the room for mistake
"Note that..." is the so-called tube lemma
 
yeah, we didn't do it that way at all
I'll have to ask the prof about this next week
 
did you use nets
dont tell me you used nets
 
nah
we're not doing nets
 
Noooo :(
tbh we're not either
nobody gives a shit
 
how the fuck did you do it?
I thought about teaching nets but then I'd have to explain nets
 
6:09 PM
theyre just like sequences yeah
just longer
put it in extra topic
 
"sequences just longer" are transfinite sequences
nets are even wilder
 
who wants maps from ordinals m8
directed sets is where its at
 
apparently the transfinite sequential closure need not agree with the closure in general
I should check out an example of that
ordinals are cool
 
define should
 
shouldnt
 
6:15 PM
$\beta \Bbb N$ should work
 
@MikeMiller just do filters
 
I'll do that when I next teach a set theory class
 
"I would say that a net is to a filter what a pseudofunctor is to a fibered category. "
 
you need to pass through filter formalism to get to Tychonoff anyway
nets just make it a whole lot less conceptually harder
 
you can show Tychonoff by nets
 
6:17 PM
ultrafilter lemma is "every net has a universal subnet" aka a subnet such that given any set A in your space the subnet either frequently enters A or X \ A
yeah its 2 lines @Thorgott
amazing power
 
well, you have to be rather terse for that
 
is it another way of saying ultrafilter
 
that there is an ultrafilter containing any filter yeah
the tail filtration of a net is a basis of a filter
 
@BalarkaSen Do you really
 
ehh i guess not really but then understanding why every net has a universal subnet, which i am pretty sure is crucial to the proof, is hard
in filter formalism thats just easy zorn's lemma
 
oh huh
i havent seen this one
Oh they use filters
just dont mention em
 
Ah, let $X,Y$ be top. spaces with $Y$ compact and $N$ open with $\{x\}\times Y\subseteq N$. For each $y\in Y$, there is a simple neighborhood $(x,y)\in U_y\times V_y\subseteq N$. Then $\{V_y\}_y$ is an open cover of $Y$, hence admits a finite subcover $V_{y_1},\dotsc,V_{y_n}$.
Then, $U=U_{y_1}\cap\dotsc\cap U_{y_n}$ is open. If $(x,y)\in U\times Y$, there is an index $i$, such that $y\in V_{y_i}$ (since the $V_{y_i}$ cover $Y$) and $x\in U_{y_i}$; so $(x,y)\in U_{y_i}\times V_{y_i}\subseteq N$, hence $U\times Y\subseteq N$.
 
Yes, it's like the ur-application of compactness
@BalarkaSen Oh
 
yeah, we used a similar argument to show disjoint compact sets admit disjoint open neighborhoods
 
Ttttttttt1
Actually you need T2
 
6:38 PM
right
for some reason we use the convention that compact sets ought to be T2 by definition
 
6:58 PM
Some people use "bicompact" for compact+t2
 
nets, filters ... I think it's time for Ted to quit this chatroom.
 
Haha hi @Ted
That's just honest topology
 
barfs
 
You don't see me encouraging them
 
I miss anon and Pedro.
 
7:02 PM
those guys were worse
clifford algebras and homotopical algebras
we're just doing filters, we dont deserve this hate
 
Nonsense.
They both talked about everything.
 
We don't talk about filters all the time
(unfortunately)
 
Speak for yourself
anon and Pedro were dirty algebraists and they had to go for a major wobbbly reform of the chat
thats how communism works
 
smacks Balarka
 
Balarka has been wobbling too much
 
7:06 PM
A space is $T_2$ when every pair of points got fairly assigned by the party a pair of disjoint nbhds
 
Oh, now they have to be assigned by some supreme power?
 
Yeah I remember those guys mainly doing algebra lol
 
I first got dragged to chat by Pedro for multivariable analysis stuff; I don't think I ever solved his question. That must have been soon after I started on MSE, whenever that was. ... Anon started with algebra but then broadened out quite a bit. I do wonder what's become of him.
I know what Pedro is up to ... in Ireland.
Hmm, 7 years 1 month ago.
 
@AlessandroCodenotti and all points are equal
 
Only homogeneous spaces are allowed?
 
7:16 PM
@TedShifrin I haven't seen anon in a while.
 
@TedShifrin Yes, he's up to alegbra. ;)
 
Yup, that he is, definitively now.
 
Is it easy to construct the differential graded Lie $\Bbb Q$-algebra whose homology computes the rational homotopy groups in a functorial manner? It seems Quillen claims something like this but I just see pages of algebra instead of a construction
I guess Ted would tell me it's the rational PL forms
 
No, that's the Sullivan model
 
i see
 
7:29 PM
I don't know the Quillen model
 
Griffiths/Sullivan/Morgan rational homotopy? I actually have a copy of that somewhere.
 
That will presumably spend most of its time on the Sullivan model
I think that's the model that's been more successful, but that might be an overstatement, idk
 
Ah found the original Sullivan paper
"infinitesimal computations in topology"
such a nice title
unlike "algebra, co-algebra and differential graded co-co-algebras in higher homotopy theory" or whatever
i made that up but that should be accurate
 
I actually have handwritten notes by some Italian graduate student (I think) from the Friedlander, Griffiths, Morgan lectures in summer 1972.
That ultimately turned into a real book.
Title: "Homotopy Theory and differential forms"
 
Nice!!
i also debunk the title "rational homotopy theory" because homotopy theorists are irrational by default
 
7:35 PM
Ah, the Springer book is just Morgan and Griffiths. springer.com/gp/book/9781461484677
 
Thanks, I'll look into these sources someday
Also I hate my dunk on algebra/homotopy theory, overdid it too much. They're both fine, just mysterious to me, until I understand it (kind of a tautological statement)
Just trying to stave off sleep, but it's been 48 hours already so I shouldn't anymore
Gonna hit the bed, cya
 
night
I'll figure out the compact subspaces till you're back
 
@Thorgott What's that
 
compact subspaces of a vector bundle are precisely those whose projection to the base is compact and which are fiberwise compact
 
7:45 PM
Are you making assumptions about the base
 
not that I'm aware of
well, it's a smooth manifold, of course
 
That's a pretty big assumption
 
I guess you can have vector bundles more generally, but I've only considered them over smooth manifolds so far
 
8:06 PM
The statement is false
Consider the union of the $x$-axis with the hyperbola $xy = 1$ and restrict to the points with $|x| \leq 1$
The projection to the base is $[-1,1]$, while the fibers are either two points or one
 
how's that a vector bundle?
 
that's $K$ sitting inside $[-1,1] \times \Bbb R$
the base is $[-1,1]$
 
ah
ok yeah, that counter-example makes sense
 
the thing I know is that closed maps with compact fibers are proper
so if you have a set with a closed map to the base whose image is compact and fibers are compact you're good
probably using hausdorff somewhere here idk
so I figured if we don't have that the map is closed we're out of luck
just wrote down an example where it's not closed
 
wait, the projection to the base isn't $[-1,1]$, it's $[-1,0)\cup(0,1]$, no?
nvm, the base is part of the set
 
8:25 PM
You can simulate a Turing machine by tiling
> This is the “origin-constrained” version of the tiling problem, in which you are required to use a certain tile (the origin tile) at least once. By making the origin tile the tape head, you obtain the unique tiling that encodes a computation.
 
9:05 PM
Hi chat!
I am reviewing complex analysis and have lost a bit of touch
arg$(z)$ is not analytic
Is it true that arg$(z)$ is not differentiable anywhere
 
9:25 PM
Complex differentiable? Yes, correct. It's a real-valued function, and the only complex-differentiable real-valued functions are constants. Remember?
 
Thanks!
So in a similar fashion Re$(z)$ is also not differentiable
 
There's another problem that the argument function is not well-defined (or not continuous if you make it well-defined).
Yes. You can check directly that the Cauchy-Riemann equations are violated. Super easy.
 
Thanks!
If $r$ is a function of bounded variation, then $r$ is continuous?
confused with
If $r$ is a function of bounded variation, then $r$ is rectifiable?
 
@CaptainAmerica16 Trying to do every exercise is surely redundant. There is a lot of repetitiveness in textbooks at the lower (university) levels. I'm not saying that you choose only the routine problems and omit the medium and challenging ones; but, on balance, probably a correctly chosen "half" is optimal. I count this as so among my books, too. Maybe even less than half. Surely you'll recall I didn't list even half the problems in Spivak for you to do!
Well, can you give me definitions, @Baymax?
 
9:42 PM
Let $I=[a,b]$ with $a<b$ and let $u:I\rightarrow\mathbb{R}$ be a function with bounded variation, i.e. $$Var_u=\sup\{\sum_{i=1}^n|u(x_i)-u(x_{i-1})|\}<\infty$$

where the supremum is taken over all partition $P=\{a=x_0<x_1<...<x_{n-1}<b=x_n\}$.
Function with bounded variation AND intermediate value property implies continuity
 
So can you give me a simple discontinuous function that is BV?
 
no idea
BV only have jump discontinutiy
so I think then BV functions are rectifiable
 
No idea? You can't just say that.
And you haven't told me the definition of rectifiable yet, so no fair switching questions.
 
That second message clearly implies you do have an idea
 
It's the lazy way out .... don't engage brain.
But it makes me angry.
 
9:53 PM
Sorry for that
 
It's OK, I'm leaving in a few minutes anyhow.
 
How about
 
I know what it means to say a curve is rectifiable. I don't know what it means to say a function is rectifiable.
 
discontinuities of monotone functions
No worries
 
 
2 hours later…

« first day (3572 days earlier)      last day (1466 days later) »