« first day (1852 days earlier)      last day (3180 days later) » 

12:16 AM
Am I alone
First week of classes... taking set theory. We are establishing axioms. So far we have extensionality, pairing, union and restricted comprehension. Are these enough to show that V is not a set?

If they are enough, please don't spoil the fun. I just have a hunch that they might not be sufficient.
 
 
2 hours later…
2:18 AM
@MikeMiller blowing bubbles?
 
No, just some of my six eyes opening wider in shock.
 
@MikeMiller ah, spextacles.
 
@Mike
do you know if such a thing is possible?
 
Six eyes?
I suppose there are some very strange medical conditions, and some insects have far more.
 
haha :)
I mean showing that V is not a set
given only 4 axioms
5 I guess
 
2:27 AM
Ah, I don't know any set theory.
 
@Mike topologically, the hopf band is the same as (homeomorphic to) an untwisted band, just twisted in how it's embedded right?
 
neither do I its week one!
 
is V supposed to refer to something specific that set theorists know about?
 
what's a hopf band
 
band with two half-twists (where a mobius band has one half-twist)
 
2:29 AM
then yeah
homeomorphism type is determined by parity of the number of twists
 
yeah. Define $V_\alpha = \bigcup_{\beta<\alpha} \mathcal{P}(V_\beta)$ with $V_0=\lbrace\rbrace$

then $V=\bigcup_{\lambda\in\text{ORD}}V_\lambda$
 
because you can unzip it along a line segment, untwist it fully to something flat, and then zip it back up to get a cuff
 
oh, i get why it's called the hopf band
 
@Prototank wow, that's big
 
Basically I think the proof to show that $V$ isn't a set involves showing that $V=\mathcal{P}(V)$
but I'm not certain
 
2:32 AM
well, ORD isn't really a set is it?
 
right but I don't think that we have that
maybe we do?
 
maybe you can show that if every ordinal embeds (as a set) into your set, then your set is not really a set
but again iunno
 
I don't know what you mean by "have that." have what? in set theory does it make sense to union over something that isn't a set?
 
right I think that is a working thing
I'm really sorry anon and Mike for not being clear
I'm in a class and we've only established 5 axioms so far
the one's I've mentioned
 
no need to apologize, also i didn't notice any lack of clarity
 
2:34 AM
being "really sorry" should be reserved for situations where you cause trouble
 
i think that any set that's too big to fail is too big to exist
 
so when I say "have that", I mean "I don't know for certain that our 5 axioms guarantee that ORD isn't a set"
 
trying to make a joke about class actions and banks. am failing.
ah
 
@anon the idea of such a joke made me smile anyway
 
user147690
3:16 AM
So if $|G|=30=2\times3\times 5$ we get:

$$n_2=1,3,5,15,\quad n_2\equiv 1\pmod{2}$$
$$n_3=1,2,5,10,\quad n_3\equiv 1\pmod{3}$$
$$n_5=1,2,3,6,\quad n_5\equiv 1\pmod{5}$$

And so by the modulo equivalences we eliminate it to:

$$n_2=1,3,5,15,\quad n_2\equiv 1\pmod{2}$$
$$n_3=1,10,\quad n_3\equiv 1\pmod{3}$$
$$n_5=1,6,\quad n_5\equiv 1\pmod{5}$$

Since 3,5-Sylow groups are Cyclic, we have that they intersect trivially
 
user147690
So then if $n_3=10$ we have $20$ distinct elements, and if $n_5=6$ we have $24$ distinct elements, which means that either $n_3$ or $n_5$ is equal to $1$
 
yikes
 
user147690
Oops didn't have chatjax on and editting is over
 
user147690
Why is there no consideration of 2-Sylow groups?
 
can you start from the top and say precisely what the question is with no $n_j$s
 
user147690
3:21 AM
If $|G|=30$, $G$ cannot be simple
 
there's no consideration of 2-sylow subgroups because there doesn't need to be
since you're trying to get too many elements from groups that aren't normal, $n_2 \geq 3$ is not particularly helpful - it says there are at least 3 extra elements
 
uhh
from earlier
is it possible to show that $\mathcal{P}(V)\subset V$
 
user147690
I don't understand
 
because of it being large
 
user147690
@MikeMiller Since 2-sylow groups are cyclic, $n_2=k$ means gives $k$ distinct elements, so then I compare the number of distinct elements given from different values of $n_3,n_5$?
 
user147690
3:29 AM
Did they do this implicitly?
 
user147690
Oh I think I see, thanks @Mike
 
4:34 AM
@Chris'ssistheartist how do we evaluate $$\int {e}^ {\sum_ {y=80} ^{x+v} b+c (y-80)}dv$$ ?
 
4:58 AM
Hello
 
ADG
5:35 AM
hello
 
user147690
5:53 AM
@BalarkaSen So what you are saying is that the Sylow theorems state that $p^am$ must admit a $p-$sylow group, so I really only need to consider the nicer sylow $7-$subgroups? (Which are nicer because intersection of primes is the trivial subgroup, whereas intersection of my sylow 2-subgroups is inteserction of order $8$ subgroups, i.e. by lagranges theorem for distinct $H,K$ sylow 2-subgroups $|H\cap K|=1,2,4$)
 
user147690
And then all it is is $8$ subgroups with $6$ non-trivial elements each, giving $48$ non-trivial elements, which forces the remaining $7$ non-trivial elements to come from my $2$-sylow, which does indeed force $n_2=1$ Which is normal
 
user147690
6:34 AM
@Rigor You change your name too much :P
 
@AlexClark that's the only thing constant :-)
 
user147690
Hopefully the user is too :P
 
:D
in The h Bar, Aug 22 at 19:59, by ACuriousMind
Hmmm...When I'm an evil overlord, I should call my servant Rigor, not Igor, as would be custom.
 
6:51 AM
@AlexClark right, precisely.
there's actually not much reason to not consider the 2-sylow subgroup. you just gotta try different p's and see which works best.
 
user147690
Thanks, that makes sense
 
@TheArtist How do you define the exponent as a continuous function?
 
i chose 7 because having 8 7-sylows would take up a lot of elements in the group
and that indicates that the 2-sylows would be near the verge of extinction
 
user147690
I am just doing some Cardano's thing right now, and afterwards I need to work out that if $|G|=112$ and $G$ is simple, that we have a monomorphism from $G\to S_7$.

Can I have a hint for that?
 
factorize 112 for me, it's too hard
 
user147690
6:55 AM
2^4*7
 
consider the 2-sylow groups. how many can be there?
 
user147690
1 or 7
 
not 1, as G is simple
 
user147690
If 1 it's not simple
 
so there are 7 of them
 
user147690
6:56 AM
Yep
 
now use Sylow's 2nd theorem to let G act on 7 of those sylow groups by conjugation
so you have a morphism, G --> S_7
can you carry on with this hint?
 
user147690
I hope so, I'll give it a good go after I finish this Cardano thing and I'll tell you how I do
 
do let me know!
hmm, i am surprised i haven't forgotten sylow theory. i guess it's a result of paying attention to these last year when i last did it.
 
user147690
Yeah it's actually pretty nice
 
sylow theory is super-nice
 
 
2 hours later…
8:56 AM
Has anybody here given a presentation in beamer use crane color theme? I want to use it, but I'm afraid it might be too bright.
 
9:35 AM
@TheArtist geometric series and then simple integration
Trying to watch $X+Y$ movie.
2
 
10:05 AM
(original title)
This^ reminds me of what the publisher told Hawking about putting equations in A Brief History of Time.
 
@r9m did you receive what I wrote to you? :-)
 
10:24 AM
Is it that a homoeomorphism takes measurable sets to measurable sets?
 
@Remember Measurable sets are just pairs $(X, \Sigma)$, where $X$ is an arbitrary set and $\Sigma$ is a $\sigma$-algebra of subsets of $X$. What does it even mean to say homeomorphisms in that context? (Note that my measure theory is limited to the very basic definitions)
 
I mean lets say if X is a topological space and Y is another topological space and they have a homeomorphism between them?
 
I don't understand what you are saying. What is the $\sigma$-algebra you have in mind for $X$, $Y$?
 
And we have $(A,\Sigma)$ where $A\subseteq X$ where $\Sigma$ is the $\sigma$-algebra of subsets of $X$
 
That doesn't answer my question.
 
10:32 AM
So my question now is will a homeomorphism map this measurable set to another measurable set , lets say $P$ in Y?
 
You have to specify a $\sigma$-algebra on $X$, $Y$.
@Rememberme Of course that is false for arbitrary $\sigma$-algebras. Why should my homeomorphism care about whatever $\sigma$-algebra I have specified on my $X$, $Y$?
You have to specify a $\sigma$-algebra compatible with the topology in some sense.
The thing I have in mind is the Borel $\sigma$-algebra, in which case I am not sure what you are saying is true or not.
I guess it is.
 
Well that is one more constraint if I think about only Borel $\sigma$ -algebra
 
For arbitrary $\sigma$-algebra, what you are saying should be blatantly false.
 
But If it does go with Borel then why not any $\sigma$- algebra?
 
Do you know what a Borel $\sigma$-algebra is?
 
10:37 AM
Yes . Wait let me give you the definition
 
It's the $\sigma$-algebra consisting of open subsets of $X$.
 
Yes
 
It's defined in a way so that it's compatible with the topology of $X$.
Why should an arbitrary $\sigma$-algebra be compatible with whatever topology on $X$?
For example, consider the boring $\sigma$-algebra consisting of all subsets of $X$, that is, $\mathcal{P}(X)$
That's my point.
 
Okay. So is just the Borel $\sigma$- algebra or are there more examples which will behave nicely under a homeomorphism . By that I mean take measurable sets to measurable sets.
 
I don't know.
Ask Daniel Fischer.
 
10:41 AM
Okay let me ping him
 
It's obviously true for Borel $\sigma$-algebras, as homeomorphisms are open maps.
 
Yes got that ...
But wait.
No no .. fine
@DanielFischer Hi! I have a question.
Is it just the Borel $\sigma$-algebra or are there more examples of $\sigma$-algebras which will behave nicely under a homeomorphism. By that I mean to say that it will take measurable sets to measurable sets.
 
Anyway, I am not the right person to talk measure theory. I know the definition, thanks to the ergodic number theory lecture I have been to, but nothing other than that. So take anything I say as a grain of salt.
 
Okay. So doing any fun stuff in commutative algebra?
 
yes
 
10:48 AM
What is it about. Btw, Yesterday I was leafing through the pages of A-M
 
it's a bit technical, i don't think you'd understand if i just talk about it
you don't know ring theory yet
 
Is it about ring homomorphisms ?
 
lol. since when did commutative algebra started studying ring homomorphisms.
 
:p . Well first chapter if I am not wrong of AM had ring homomorphisms ..
 
ch. 1 is basic ring theory, intended as a revision
 
10:51 AM
Okay. nvm
 
@Rememberme Any $\sigma$-algebras generated by a family of sets that is respected by homeomorphisms does that. Apart from the Borel $\sigma$-algebra, the only one I know off the top of my head that is used somewhat frequently is the $\sigma$-algebra generated by the compact sets. If memory serves, that's called the Baire $\sigma$-algebra.
Of course, for many spaces, that is in fact the Borel $\sigma$-algebra.
 
A very nice problem (repet, a very nice problem) from the IMO shortlisted problems
Let $a>2$ be given, and define recursively

$$a_0=1, \ \ \ a_1=a, \ \ \ a_{n+1}=\left(\frac{a_n^2}{a_{n-1}^2}-2\right)a_n.$$

Show that for all $k\in \mathbb{N}$, we have

$$\frac{1}{a_0}+\frac{1}{a_1}+\frac{1}{a_2}+\cdots+\frac{1}{a_k}<\frac{1}{2}(2+a-\sqrt{a^2-4}).$$
 
Hi, @DanielFischer
 
But you can for example take the $\sigma$-algebra generated by (path-)connected sets. I don't think that will be very useful, though, @Rememberme.
Hi @BalarkaSen.
 
Oh. we have $\sigma$- algebra generated by compact sets also..
Oh. So is it just the Borel algebra which leads to amazing results?
 
10:55 AM
@Rememberme Not only, but mostly.
 
okay.
 
@DanielFischer How's life?
 
@BalarkaSen Rainy. And a little too cold, but not yet cold enough to turn on the heating.
 
Ah. The weather's been cloudy around here too.
 
It isn't hot here too . The temp here is 20 degrees
 
11:16 AM
I just took a break from watching $X+Y$, and I'm absolutely sure I wouldn't like to be such a freak like Nathan in movie. I think being brilliant doesn't necessarily mean to be so lacked of social skills and behave like a mega freak.
I hope things improve later in the movie (they pushed things a bit too far I think).
Back to movie.
 
@DanielFischer What all can we say about the infinitely countable product of a set X whose subset A is measurable ? $A\subseteq X$
 
@Rememberme Sorry, unclear what you're asking. Can you clarify?
 
Sure.
Let $X$ be a Hausdorff space (sorry I wrote a set there). Let $A$ be a subset of $X$ with $\Sigma$ as the $\sigma$ - algebra of subsets of $X$. What all can we say about the infinitely countable product of the set $A$ in the infinitely countable product of the Hausdorff space $X$ ? @DanielFischer
 
@Rememberme You mean $\Sigma = \mathcal{P}(X)$? And you want to know something about $A^{\mathbb{N}} \subset X^{\mathbb{N}}$? Are you interested in topological things, or measure-theoretic?
 
11:32 AM
Yes and both topological things(but only which lead to measure theoretic things) and measure theoretic things @DanielFischer
 
Well, if $A$ is closed, then $A^{\mathbb{N}}$ is closed. If $A$ is connected, so is the product. But if $A$ is open, then $A^{\mathbb{N}}$ is open only if $A = \varnothing$ or $A = X$. Measure-theoretically, $A^{\mathbb{N}}$ is measurable (with respect to the product $\sigma$-algebra, which usually is much smaller than $\mathcal{P}(X^{\mathbb{N}})$).
 
That means the infinite product of any measurable set is measurable?
 
@Rememberme countable. If we have uncountably many factors, things are different.
 
Different like?
And why
 
The product $\sigma$-algebra is generated by sets of the form $$M_\alpha \times \prod_{\beta \neq \alpha} X_\beta.$$ If the index set is countable, every product of measurable sets is a countable intersection of sets of that form and hence belongs to the product $\sigma$-algebra. If you have uncountably many factors, you need an uncountable intersection, and that isn't guaranteed to remain in the $\sigma$-algebra.
 
11:42 AM
Okay.
 
@Chris'ssistheartist how do we apply geometric series to that? :/
 
@TheArtist How can you write this one? $$\int {e}^ {\sum_ {y=80} ^{x+v} b+c (y-80)}dv$$
$$\sqrt{\frac{\pi }{2 c}} e^{\large -\frac{(c-2 b)^2}{8 c}} \text{erfi}\left(\frac{2 b+c (2 v+2 x-159)}{2 \sqrt{2 c}}\right)$$
Back to movie.
@TheArtist ^^^
and one more thing, there is a difference between $$\sum_ {y=80} ^{x+v} b+c (y-80)$$ and $$\sum_ {y=80} ^{x+v} (b+c (y-80))$$
So, I assumed the second version since the first one was trivial.
Back to movie.
 
12:15 PM
@MikeMiller Taking the fact that $H^n(X; G)$ is set-isomorphic to $[X, K(G, n)]$ completely as a black-box, the natural question that comes to mind is if we can transport the structure of abelian group of $H^n(X; G)$ to $[X, K(G, n)]$. I mean, of course we can do it (just label and do addition on labels), but the question is if the resulting operation geometric enough. I haven't thought about a group structure yet, but I seem to have some kind of analogue for the cup product structure :
take two classes of maps $f : X \to K(G, n)$, $g : X \to K(G, m)$. We want to construct a map $X \to K(G, n + m)$ out of these. OK, first consider the map $f \times g : X \to K(G, n) \times K(G, m)$ obtained from sending $X$ to the first copy by $f$ and to the second copy by $g$. $K(G, n) \times K(G, m)$ has the CW-structure consisting of a $0$-cell, an $n$-cell, an $m$-cell, an $m + n$ cell and all the highers cells which kills the higher homotopy groups.
To eliminate $\pi_n$, glue a $(n+1)$-cell along a meridian of $K(G, n) \times K(G, m)$ (I'm visualizing at a torus, to clarify). Thus, any map from $S^n$ can now be slided to that meridian and contracted through that ball. Similarly, glue an $(m + 1)$-cell along the latitude. This kills $\pi_m$. The resulting space, I think, should be a $K(G, m+n)$, but I can't prove this rigorously. Note that the modified space is just $K(G, n) \wedge K(G, m)$ (wedge product).
I probably won't be able to prove this, as my homotopy theory is weak. I am also trying to get the abelian group structure to work, but I don't have any interesting geometric way to endow $[X, K(G, n)]$ with a group structure. I am not even sure if there is one.
The thing I am ultimately gunning for is to prove that $[X, K(G, n)]$ is a functor, and in fact a cohomology theory, so that the classification of cohomology theories would hand me in a isomorphism $[X, K(G, n)] \cong H^n(X; G)$ in a formal way.
 
Again, taking a break from the movie, and I dare to say that being mathematically very smart, brilliant (if I can say that) is not enough at all in my opinion. Behaving like a freak a whole life I don't know if this makes you happier.
Excelling in social skills I might say is at least as much important or far more important, but well, that's my personal view finally and many of the authors of the books I read about personal development agree with me.
I highly recommend some reading by Dr. Walter Doyle Staples (for anyone interested in the matter).
 
12:31 PM
I see I've been sort of vague at a few sentences, so here're those spelled out : (1) I am giving $K(G, n)$ a CW-complex structure by attaching a few $n$-cells to a $0$-cells, attaching a few $(n+1)$-cells for the relators, and attaching higher cells to eliminate higher htpys. (2) The modification I did on $K(G, n) \times K(G, m)$ kills $\pi_n$, $\pi_m$. The only nontrivial homotopy is at the $\pi_{n+m}$ level now. The issue is what happens higher above.
(3) The modified space is nor exactly the same as $K(G, n) \wedge K(G, m)$. It's homotopy equivalent to it, by sliding the whole thing along the two attached balls.
That is, the subspace $A$ consisting of wedge of the two balls is contractible, and thus can be pinched without any issue. $X \to X/A$ is a homotopy equivalence.
 
12:55 PM
ah, hell, I wrote down wedge product when I really meant smash product.
@MikeMiller oh, actually, I don't need a homeomorphism between $K(G, n) \wedge K(G, m)$ and $K(G, m+n)$. Indeed, that's not in fact true : Take $G = \Bbb Z$, $m = n = 1$. What should be true, on the brighter side, is that there exists a map $K(G, n) \wedge K(G, m) \to K(G, m + n)$. For $G = \Bbb Z, m = n = 1$, the obvious map takes $S^2$ to the $2$-skeleton of $\Bbb CP^\infty$.
Indeed, this is sufficient to give a pairing $[X, K(G, n)] \times [X, K(G, m)] \to [X, K(G, m + n)]$.
 
1:16 PM
The end of movie, and the last part was somewhat nicer. Overall I spent almost 2 hours watching a mediocre movie, I could have done more work in the research area. That's it.
(the truth is that I also thought of some problems during the movie - I wasn't totally unproductive)
 
The best policy is to not watch movies at all, @Chris'ssis. That way, no time would ever be wasted.
:P
 
@BalarkaSen Most of the time I regret after watching movies. :-)
 
Hi @MatsGranvik
 
@evinda hi, what is up?
 
Fine, thanks. How are you doing? :) @MatsGranvik
 
1:20 PM
Fine, I found that the argument of this expression is zero:
$$\arg \left(\sum _{n=1}^{k} \frac{1}{n}\text{Total}\left[\mu (\text{Divisors}[n]) \text{Divisors}[n]^{1-(x+i y)}\right] \text{Total}\left[\mu (\text{Divisors}[n]) \text{Divisors}[n]^{x+i y}\right]\right)$$
 
@MatsGranvik Interesting.. :)
 
Yes interesting and disappointing at the same time.
 
Yes, that's true. @MatsGranvik
Didn't you work as a chemical engineer? @MatsGranvik
 
@evinda Yes, or more precisely as a laboratory worker.
 
@MatsGranvik Nice... Do you still like your job?
 
1:26 PM
@evinda It is okay. I don't complain.
 
@MatsGranvik Nice...
 
@MatsGranvik What does Total[...] mean? I am not familiar with mathematica codes.
 
@BalarkaSen Just the sum over the divisors.
 
And $\mu$ is our friendly mobius function, right?
 
@BalarkaSen yes
 
1:30 PM
@BalarkaSen Do you have vacations now?
 
@MatsGranvik Eh. Sum of what over divisors of what?
 
Wait I will latex it for you. But anyways this is most likely trivial.
 
@evinda Somewhat. Exams ended about 2 weeks ago.
 
I see... When will the next school year begin? @BalarkaSen
 
Oh, it was not a final or anything. Just a midterm (if that's what it's called).
 
1:32 PM
@BalarkaSen Ah, I see..
 
@BalarkaSen $$\sum\limits_{n=1}^{n=k}\frac{1}{n}\left(\sum\limits_{d|n} \mu(d)d^{(1-s)}\right)\left(\sum\limits_{d|n} \mu(d)d^{(s)}\right)$$
 
ah, I see
Why do you find this interesting/disappointing? I'm curious. (I don't know how to prove that thing is zero off the top of my head)
 
@BalarkaSen Go to the number theory room and look at the plots. They are variants of Franca LeClair's functions
This expression is the symmetrized expression that anon gave me two days ago. The interesting part is that a single möbius sum over the divisors (times the Riemann zeta function) accentuates the zeta zeros a little like a Fourier series. But when taking the argument of the symmetrized functional equation (the chi function) this part above just becomes zero.
 
Sounds interesting. Post it as a question in MSE when you find something interesting! I really like the weird numerical coincidences you come upon while tweaking with the codes.
k, I have to run.
 
1:48 PM
Good morning chat!
 
hi
 
morning, @Fargle. doing galois theory much?
 
School has started back up for me, so I've had to put some of my self-study on the backburner. I should use my free time for that, though. What've you been up to, @Balarka?
 
not much. learning multivariable calculus, commutative algebra, singular cohomology.
 
Gummornin
 
1:59 PM
morning.
 
Sounds like fun.
 
singular cohomology is not very fun so far. way too algebraic for my taste.
if I can make out to the end of the cup product section, then the fun things will come, e.g., Poincare duality.
 
I have a cup for drinking water out of
 
2:35 PM
One day I'll want to visit India, the birthplace of Ramanujan, where he lived, learned, played, and so on.
 
3:11 PM
@Chris'ssistheartist the thing i'm immediately tempted to do with that problem (mostly b/c i'm a fan of nice parametrizations) is to write $a=2\cosh\tau$ for some positive $\tau$. then the upper bound takes the form of $1+\cosh\tau-\sinh \tau = 1+e^{-\tau}$, which is rather cute looking.
(no guarantees it's actually useful, to be sure, since one still has to deal with than nonlinear recurrence relation)
ooh, but mathematica seems to say that then $a_n = \dfrac{\sinh(2^n \tau)}{\sinh \tau}$
which is really quite nice
 
shit that's a lot of words @Balarka
 
lol
 
(which further suggests a link with Chebyshev polynomials of the second kind, but i don't see a reason to go down that route)
 
sorry, @Mike. read it when you get time.
 
Yes, $K(G,n) \smash K(G,m)$ is almost $K(G,n+m)$. You then add higher cells to make the smash product actually $K(G,n+m)$.
 
3:23 PM
ah. and then the map is just inclusion into an appropriate skeleton.
 
Well, collapsing the wedge product first.
There's also a way to make $K(G,n)$ an H-space so that the isomorphism $H^n = [X;K(G,n)]$ is a group isomorphism with the induces group structure on the latter. I don't remember how, but it's an exercise in Hatcher, so it can't be too hard.
The fact that they're isomorphisms as functors to set is the thing that really gets my goat anyway.
 
@MikeMiller oh, nice. I am going to try to find the $H$-space structure.
@MikeMiller well, once one can show that $[X, K(G, n)]$ is a cohomology theory, that is immediate.
the only drawback being that being a cohomology theory might not entirely be a trivial fact
 
i'm not gonna click on that.
 
given the context of that line in the movie..."get a room, you two!"
 
3:27 PM
mm, i guess he meant it when he said it here, which is perhaps not my intent
@BalarkaSen this is a very trivial sort of "that is immediate", given that yes, to do so, you need to show it has a group structure etc.
 
i have no idea what both of you are talking about. i guess it must be something about that stupid video Mike linked me.
 
yes, it's immediate that a cohomology theory has a group structure, by definition. which is why you have to prove that it has a group structure...
 
@MikeMiller right.
 
r9m
@Chris'ssistheartist ya .. I got it :) Thanks!!!! :D
 
@MikeMiller wut? if $K(G, n)$ has an $H$-space structure, so does $K(G, 1)$ by specializing to $n = 1$. $\pi_1(K(G, 1)) = G$ might not be abelian, contradiction.
sorry, I think what you said is false
 
3:34 PM
"except for $K(G,1)$ for nonabelian $G$"
lol
in the case that $G$ is nonabelian $H^1(X;G)$ is not a group, but one can still make sense of it as a set, and in that case we still have $H^1(X;G) = [X;K(G,1)]$
 
"in the case $G$ is nonabelian $H^1(X;G)$ is not a group" ah, indeed!
should have figured that implicit assumption
ok, i am really going to try to find the $H$-space structure now.
 
oh, got it. it's not terribly explicit but it's cute
 
3:57 PM
@MikeMiller isn't this stupidly easy? $G \times G \to G$ is given by addition on $G$. apply $K(-, n)$ to get $K(G \times G, n) \to K(G, n)$. now $K(G \times G, n) \cong K(G, n) \times K(G, n)$
so there, you have a map $K(G, n) \times K(G, n) \to K(G, n)$
 
Sure. Why is it homotopy associative, homotopy commutative, homotopy has an inverse?
 
Hello all! I just asked a question here and I think I've been able to answer it. Could someone verify my logic? Thanks!
 
@MikeMiller i thought we just needed left and right identities (modulo homotopy)?
we need an H-space structure, not a topological group structure
 
A topological group structure is not up to homotopy.
I guess you're right. Whatever. Do it anyway :)
You probably need that to get a group structure on $[X, K(G,n)]$.
 
yeah, i think so too.
but this should be a good exercise one way or another. i have never proved the functorial properties of $K(-, n)$ before. (as a functor from Grp to hTop)
 
4:04 PM
All of my brackets should have been $\langle$, as a side note.
 
what's the difference?
 
Pointed maps.
 
ok, right.
i use $[\cdots]$ interchangeably to write both.
 
They're different, of course.
 
yeah, that i know.
 
4:05 PM
I tend not to make a big deal out of the difference because I always mean pointed maps unless otherwise specified.
 
i just didn't know there was this concise notation.
i guess Hatcher uses it, but i just thought that was meant as a stylish replacement of $[\cdots]$.
 
4:19 PM
Do you like lamingtons, @chamington?
 
4:37 PM
lol
 
4:50 PM
@BalarkaSen: I assume you proved that the map above is homotopy commutative etc and moved on?
 
Sorry, @Mike, I haven't tried it yet. Pondering on theory than doing concrete exercises, as usual! It's turning into a bit of a bad habit.
Let me try it.
 
and then don't forget to get some calculus in
 
I was doing calculus this morning, in fact.
 
ok, homotopy commutativity is not hard. $K(G \times H, n)$ is homotopy equivalent to $K(G, n) \times K(H, n)$, that's all you have to note.
 
5:02 PM
No, you want to say something about $G$ being commutative.
In fact I'm not sure where the $H$ shows up here.
 
yeah, I mean, the maps $G \times G \to G$ given by $(x, y) \mapsto x + y$ and $(x, y) \mapsto y + x$ are the same thing. so by that fact, they descend down to the same maps at K(G, n)-level modulo homotopy.
 
Yeah.
 
ok, essentially the same thing about homotopy associativity. the maps $G \times G \times G \to G$ given by $(x, y, z) \mapsto (x + y) + z$ and $(x, y, z) \mapsto x + (y + z)$ are the same, so same at the level of K(G, n) modulo homotopy.
 
mhm
 
erm, what does homotopy inverse mean?
 
5:09 PM
there is a function $i: X \to X$ such that $\mu(x,i(x)) \simeq \mu(i(x),x) \simeq e$, where $e$ is your 'homotopy identity'
 
ah. i was confused, as $K(G, n)$ is not like a group or something.
 
it's an inverse map up to homotopy, just like $e$ is an identity up to homotopy
 
yeah, understood.
it's precisely the map $K(G, n) \to K(G, n)$ induced from $G \to G$ defined by $x \mapsto x^{-1}$.
 
yeah
(which is another place you used that $G$ is abelian.)
 
right. i don't think there was anything else which i had to prove?
 
5:18 PM
nop
 
ok, so this ensures that $\langle X, K(G, n)\rangle$ is an abelian group.
 
mhm
 
ok, i am going to ponder a little more on the theory and get back to calculus.
 
I need to answer some questions with a higher profile. My answers have been sleeping.
 
5:54 PM
@robjohn: If a couple good new questions show up in my tags today I'll hit the rep cap for the first time in a while...
 
@Semiclassical Yeap. :-)
@Semiclassical ^^^ it's a raw closed-form
Anyway, it's far too long in this form.
 
6:13 PM
@MikeMiller congratulations! I was lucky enough to cap yesterday. Likewise for me, it has been a while.
 
6:24 PM
There's a cap?
 
200 rep a day, ignoring the rep you get from accepted answers.
 
Oh, I see.
 
What are you up to?
 
Differential geometry now. Struggling with Lee. Favourite course this semester is definitely algtop.
How about you?
 
Putting off some reading I should do today, driving to the video store to return some movies.
How're those courses going? What are you doing?
 
6:31 PM
We haven't got very far yet; still on Riemannian metrics, although Lee uses notation that seems to presuppose mathematical maturity I don't have.
Otherwise I am doing cohomology, algtop, complex analysis and commutative algebra this semester.
 
This is Lee's Riemannian book?
Cohomology is a separate course than algtop? What do they cover?
 
0
Q: A triple integral dancing in the unit cube

Chris's sis the artistStraight integration seems pretty tedious and difficult, and I suppose that the symmetry might possibly open some new ways of which I'm not aware. What would your idea be? $$\int_0^1 \int_0^1 \int_0^1 \frac{x^2}{\sqrt{x^2+1} \left(x^2-y^2\right) \left(x^2-z^2\right)}+\frac{y^2}{\sqrt{y^2+1} \lef...

 
Yes, they split the courses, cohomology gives only half the credits of algtop, it was simply too big for one course. CW-complexes, fundamental group, homology and then cohomology for the full thing.
 
I see.
 
As I do not know the material yet its hard to give a more accurate description. And yes: Lee's Riemannian book.
 
6:34 PM
Ok.
 
I find some explanations inadequate (nothing wrong with that, the book is not intended for undergraduates.) Currently writing a translation to undergradian.
What were you supposed to be reading?
 
I'm not very good at Riemannian geometry but feel free to ask questions. It'd be good for me. :P
A paper by Michael Hutchings relevant to my research proposal. I should read it before I meet with my advisor tomorrow.
It's not too technical so it shouldn't be too bad.
 
Ohh, advisormeetings are scary, especially if unprepared.
 
@AndrewThompson They really should not be (but then, you should not be unprepared for them)
 
Ah, I have plenty to say. I just found out about this paper yesterday and it would be good to know what's in it, too.
 
6:39 PM
I'll take you up on that: given a Riemannian metric $g$, it is claimed that one may write it locally as $$g = g_{ij}\varphi^i \otimes \varphi^j$$ where $g_{ij}$ is the matrix $\langle E_i, E_j \rangle$ for a local frame $(E_1, \dots, E_n)$ and $\{\varphi^j\}$ a dual coframe. So I'm assuming the claim is simply $$g_p(X, Y) = \langle E_i|_p, E_j|_p \rangle \varphi^i_p(X)\varphi_p^j(Y)$$ with summation by the Einstein convention, right?
Ah, good!
the dual coframe, I'm sorry.
 
Yeah, that's precisely it.
 
Thanks. Its not routine yet for me, so I think I'll write it out in childish detail at some point.
 
I recommend that!
Childish detail helps me a lot.
 
Yes, I have learned that the hard way. We have very active student participation in our classes (in practice mandatory to show solutions on blackboard etc), and oftentimes I have proudly presented a solution and the prof goes
"Well, this all well and fine, but why is that map smooth?"
And I'm stuck.
 
Good on the prof for doing that!
 
6:47 PM
Yup, he's superpedagogical.
 
Are these homework problems or problems he just decided to ask people to do in class?
 

« first day (1852 days earlier)      last day (3180 days later) »