« first day (1360 days earlier)      last day (3666 days later) » 

12:33 AM
@KarlKronenfeld @Mike
 
Damn, I was just about to break the eerie silence, @Pedro.
 
@KarlKronenfeld I read your mind. Don't worry.
=)
I have a problem.
I think I am doing things correctly.
Suppose $0\to M'\to M\to M''\to 0$ is exact. If $M',M''$ are finitely generated, so is $M$.
 
I had first the idea of using that the above is equivalent to having an exact sequence $A^n\to M'\to 0$ and $A^m\to M''\to 0$, (we can take $m=n$) and obtain one $A^k\to M\to 0$, I am thinking I can get $k=n$:
But now I think I can just work with generators, and obtain a set of generators from $M$; like with vector spaces.
I will do it and see what I get.
 
The second is fairly straightforward, but a variant of your first idea also works.
(I wouldn't try to establish k=n)
 
12:40 AM
@KarlKronenfeld OK.
 
that's so hot
 
What did you have in mind?
 
Let $k=m+n$
 
Oh, I wasn't talking about that. But yes, I get what you're saying.
Suppose $\langle x_1,\ldots,x_n\rangle=M'$ and $\langle y_1,\ldots,y_m\rangle =M''$.
Let $x_j'\in M$ be such that $x_j'\mapsto y_j$.
Then the claim would be $\langle f(x_1),\ldots,f(x_n),x_1',\ldots,x_m'\rangle$ generate $M$.
$$M'\mathop{\longrightarrow}^f M$$
 
Hey @JasperLoy
 
12:50 AM
@KarlKronenfeld Done. =)
 
Someone! I asked this this morning, and I think I got a sufficient answer, but to show that $[0,1]$ is not of measure zero, can I do the following?:
Actually I don't know how to formalize this.
It's covering compact, so every cover has some finite subcover.
 
Guys! the arithmetic - geometric mean of $1$ and $x$ is $$\pi \div \int_{0}^{\frac{\pi}{2}}\frac{2}{\sqrt{1-\sin(t)^2\cdot(1-x^2)}}\,\mathrm{d}t$$
 
In order for a finite subcover to cover something, it would seem intuitively that the combined length of the subcovers would need to be at least as large as the interval, but I don't actually know how to say that.
 
An interval $[a,b]$ is, by definition, of measure $b-a$
 
Ugh
I don't think I have that...
I just know that it has a finite cover...
 
1:01 AM
How did you define measure if not?
 
I'd guess Anthony uses open intervals.
 
I'm in intro to real analysis
We've only talked about measure zero
And it's just as Karl said.
I think? Hold on.
In other words, for each  $\epsilon$> 0, there some open set U containing Z whose total length is less than $\epsilon$
 
ok, but what's length then?
 
I suppose I shouldn't have said length?
The question is
Show that [0; 1] is not measure zero. [Hint: recall that you know [0; 1] is covering compact.
Then show that for any finite cover of [0; 1], the sum of the lengths of the intervals is bounded
below by 1.]
 
@Anthony So that's your definition of ($Z$ having) measure 0?
 
1:06 AM
A countable cover by open intervals
 
So you know how to measure open intervals then, right?
 
What do you mean measure an open interval?
 
Aka the "length".
 
I mean, I guess. Here this is the problem.
 
ok, so you have finitely many open intervals covering $[0,1]$. There's actually a cute way to do this by replacing $1$ with an arbitrary positive real number, and inducting on the number of open intervals in your covering.
 
1:11 AM
Oh the question wasn't even including in the screenshot
I don't even know what I'm asking anymore
 
I was just going from where you left off when you initially asked your question in chat
 
Yeah-so I have [0,1]
Every cover has a finite subcover
Then what? What's the best way to proceed?
 
I don't know the best way to proceed. I offered a suggestion just a moment ago though. :P
 
Inducting on the number of open intervals?
 
There will always be an interval which covers 1. Remove that, you have one less interval.
The set covered will be smaller, hence I suggested replacing 1 with some $x$.
 
Ram
1:18 AM
I have a question regarding Connected sums. T1#T2/T1-B^0 how to prove that this is homeomorphic to T2?
 
ping @Pedro
 
lol I looked up $p$-divisible group in nLab.
 
@KarlKronenfeld , I'm having trouble following. First, our finite subcover isn't necessarily disjoint, right? Second, if I remove a subcover, I will have one less interval, and the set will be smaller, but why does that help?
 
HAI @FernandoMartin
@KarlKronenfeld
 
Instead of proving that any finite cover by open intervals of $[0,1]$ has cumulative length (that $\sum (a_i-b_i)$ thingy) of at least $1$, generalize a little and show that $[0,x]$ any finite cover by open intervals has cumulative length of at least $x$.
Let $(a_1,b_1),\dots,(a_n,b_n)$ be your open cover of $[0,x]$ and suppose without loss of generality that $b_i\ge a_i$ (remove extra intervals as necessary). Then, $(a_1,b_1),\dots,(a_{n-1},b_{n-1})$ covers $[0,b_{n-1}]$ and $(a_n,b_n)$ covers $[a_n,x]$. There is some overlap and you can find a lower bound of the cumulative length using yo
 
1:31 AM
I was trying to prove the exercise from last class @Pedro
The corollary of that result by Serre
have you thought about it?
 
I didn't pay much attention. =P
 
smacks Pedro with a commutative algebra textbook
 
$\mathcal O_0^{\rm an}$ where analytic functions in a nbhd of $0$.
 
germs of
 
Yeah, yadda yadda...
 
1:33 AM
and let's denote it $A$ please
 
And $\mathcal O_0$ where rational functions defined in a nbhd of $0$?
 
defined in 0
i.e. the denominator doesn't vanish at 0
 
Ah, OK.
 
though yeah they don't vanish in a whole nbhd
 
The claim was the former was a faithfully flat the other-module.
 
1:34 AM
faithfully
That's the result we are going to use
without proof
 
to prove the exercise
 
And the exercise was?
 
@KarlKronenfeld :(
 
let's say $f, f_1, \dots, f_m$ are rational functions such that there exist germs $h_1,\dots,h_m$ such that $f=\sum h_i f_i$ (as germs)
 
1:35 AM
OK.
 
then there exist rational functions $g_1,\dots,g_m$ such that $f=\sum g_i f_i$
 
@Anonymous What's the matter?
 
I have proved it in the case where $f=0$
 
Egasp my name changed.
 
maybe it follows from affinizing (is that even a word?) the case $f=0$
but I don't see how
 
1:40 AM
Let's write out what it means that $\mathcal O_0^{\rm an}$ is a faithfully flat $\mathcal O_0$ module.
 
It is fixed!
 
It means that the "tensoring by germs over rational functions" functor is exact and reflects exactness
 
Hi @Fernando @Pedro ... Damn, them is some fancy words you'ze slinging around, mr @Pedro.
 
Hi @Ted
 
@FernandoMartin I know what it means. I'm just saying that it might help to write things down and look at them carefully.
@TedShifrin It's a big theorem by Serre.
 
1:45 AM
Oh I though you didn't write it down
 
Our professor gave it out, and said we're going to use it.
 
Yes, I've known a few of his big theorems :P
 
@FernandoMartin I didn't.
 
I'm still a little lost despite Karl's help. To show [0,1] has nonzero measure, I don't really know how to proceed. He said to generalize, but I don't see how that helps...
 
I'm glad you're finally in a class that challenges you, @Pedro.
 
1:46 AM
 
I prefer not to stay in groups.
@Anthony: It's not a totally easy argument, as I recall.
So you're supposing you have a countable cover by open intervals of total length $<\epsilon$?
 
The route I was recommended to use, for [0,1] is to use the fact that it's covering compact.
 
@TedShifrin I can't get enough of them. =D
What is Anthony trying to prove?
 
That $[0,1]$ is not of measure zero, I take it.
 
Yep.
 
1:49 AM
The above is a group by Hall.
 
yes, @Pedro, I dun recognized it.
 
Ram
Hi, can someone help me with this connected sum problem, T1#T2/T1-B^0 how to prove that this is homeomorphic to T2?
 
OK, @Anthony, so what happens if we do that?
 
Just prove that it has measure 1 by applying the theorem that the length of an interval is its measure, lol.
 
@Anthony Oh, and you define measure by the exterior measure using open intervals?
 
1:49 AM
not helpful, @Jasper
 
@Ram NOT ENOUGH INFORMATION FOR AN ANSWER.
 
I don't think he's doing Lebesgue theory yet, @Pedro
 
How do you define measure Tony?
 
When did Anthony become Tony? LOL.
 
For example, I define measure zero in my manifolds and differential topology classes, but not measure.
 
1:51 AM
Ahhhh
All I have is measure zero, and it was defined for me as i.stack.imgur.com/oDawU.png
 
Ram
@PedroTamaroff, T1 and T2 (say closed and connected) are two surfaces, and their connected sum as usual. And by cut and paste, it is clear that T1#T2/T1-{int open Ball} is homeomorphic to T2, but how to prove it in a better way?
 
So we have a finite cover of $[0,1]$ by open intervals whose total length is $<\epsilon$? So then we should be able to show easily that the length of the interval is $<\epsilon$, right, @Anthony?
 
Ram
Compact too incase
 
@TedShifrin You mean BWOC?
 
gesundheit! say what?
 
1:53 AM
WTF is BWOC?
 
@Jasper, you're back to name #127?
 
@TedShifrin Yes. I managed to change it before 30 days was over because I created a new account on another SE site.
 
By way of contradiction, lol.
 
Claim If $[a,b]$ is covered by finitely many open intervals $I_1,\ldots,I_m$ their total length is $>b-a$.
 
Ohhh ... never seen that before, @Anthony :D Yes.
 
1:56 AM
So yes, then we would need to show each interval is $\lt \epsilon$. So....?
 
Let $(a_i,b_i)=I_i$.
*Claim 1*. We may assume that $a_0<a, a_2<b_1,b_2<a_3,\&c$.
 
Wait. Are you proceeding my way or Pedro's way, @Anthony?
 
Using this, we're done. =)
 
Waaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaat is going on. I was listening to you Ted, but now I'm reading Pedro's.
 
Make a drawing.
 
1:57 AM
@Pedro now gives Ted's advice. Impressive.
 
Ram
@PedroTamaroff, "Make a drawing" is that for me?
 
@PedroTamaroff I understand that if there are finitely many covers then each should have non trivial length, but I don't know how to say that formally.
 
@Anthony: I meant the length of the total interval they covered had to be $<\epsilon$ if all the individual ones had lengths adding up to $<\epsilon$. Pedro's doing the contrapositive, essentially.
An open interval $(a,b)$ has length $b-a$, @Anthony.
 
Always be sure your argument is independent of the drawing used to motivate it.
 
Hush up, @Jasper. :)
 
2:00 AM
@Ram Not really, no.
 
@Jasper: I'll give you a linear algebra problem that I'd never seen before (it was posted earlier as a disguised differential geometry problem).
 
Ram
ok, can you suggest me some thing?
 
@TedShifrin No, no, please don't.
 
@Ram I have no idea about what you're talking about =)
 
Ram
ohh, dont you know connected sums from Algebraic Topology?
 
2:01 AM
Suppose you have three unit vectors $x,y,z$ in $\Bbb R^2$, and $x\cdot y = \cos\theta$, $x\cdot z = A$, $y\cdot z = B$. Prove that $\sin^2\theta = A^2+B^2-2AB\cos\theta$. I've never seen this before.
 
@PedroTamaroff alicia just sent an email
 
@Ram: That's one subject Pedro hasn't studied yet.
 
@FernandoMartin Alicia or Alyssa?
 
@TedShifrin, Little confused as to what is being said.
 
@FernandoMartin Didn't get any.
 
Ram
2:02 AM
@TedShifrin :-(
 
Ahh, did you send her your email?
Maybe you're not on the mailing list
I'll forward it to you
 
Pedro is telling you to lay 5 intervals down on the real line, @Anthony, overlapping so that they cover $[0,1]$. Why must their total length be at least $1$?
 
Ram
@TedShifrin, can you help me to prove my question more formally instead of waving hands.
 
I wasn't paying attention to your question, @Ram, but in a few minutes I must go back to grading my students' homeworks. What is your question?
 
Ram
T1#T2/T1-B^0 how to prove that this is homeomorphic to T2?
some thing other than cut and paste technique
@TedShifrin
 
2:04 AM
Well, what other technique is there for such a thing?
 
Ram
T1 and T2 are connected and closed
 
These are surfaces, I take it?
 
Ram
yes
 
You're taking $(X\cup Y)/Y$, essentially, right?
 
Ram
if we collapse the boundary right?
 
2:05 AM
@TedShifrin That makes sense, but how do I saaaaay that.
 
@Anthony: You right down the inequality you get knowing that the intervals must overlap.
 
And also, then why is it that I can cover the rational numbers with zero length... it's so confusing.
 
@FernandoMartin Got it.
 
Yes, @Anthony. The hint for that is: geometric series.
you mean the boundary of the ball, @Ram?
 
It's so strange though.
 
Ram
2:06 AM
@TedShifrin, yes, boundary of the ball.
 
Yes. Because you're modding out by $T1 - B^o$, so the part that's left inside $T2$ is the boundary of the ball in $T2$.
 
@FernandoMartin The books looks cool, I can see myself reading it during winter.
 
@Anthony: There just aren't many rational numbers :P
I thought you're getting old during winter, @Pedro :D
 
@TedShifrin I am.
 
Getting old and reading math? Tut tut :D
 
Ram
2:09 AM
@TedShifrin, yes, I figured it out, but I want some one to confirm this. Thanks for your help.
 
Sure, @Ram.
 
Ram
Happy grading, I have to grade from next sem :-(, will be in grad school.
 
Yes, @Ram, I've been doing it for about 40 years now.
So I'm ready to let you young'uns do it :P
 
Ram
:)
 
Just take it seriously and read and give serious criticism/feedback. Don't just put checkmarks on the papers.
OK, back to work for me.
 
Ram
2:12 AM
@TedShifrin, thanks for suggestion. I never been to Math school before this sem.
Thats really helpful
 
Really? You're learning algebraic topology and you just started? I'm confused.
 
2:24 AM
Hi Ted
I decided against taking differential geometry next semester
I'm rather taking 2 intro analysis courses
 
2:53 AM
Hi Ted
I decided against taking 2 intro analysis courses next semester :)
 
lol Mike
 
"Although the function ... effectively yields the prime numbers at coefficients z1"
Heh.
 
@user127001 2? What do they cover?
Anyone reads Shakespeare here?
 
did macbeth, othello, midsummer, and much ado in high school. it's on my bucket list to read it all.
 
@seaturtles It's not Old English right? So I should be able to read it right?
 
It's closer to today's english than old english. One difference is people say trivial things in elaborate ways to show off how clever shakespeare is and to let the audience feel good about understanding it. (Or at least that was my impression.)
 
3:17 AM
@seaturtles I will recommend you one book with all his works then. Ready for it?
 
I already have his entire works.
In print form.
 
Oh, OK. I am getting one.
 
not counting any apocrypha
 
I have retired from Math and Eng SE. I will now focus on ELL SE, lol.
 
ping @pedro
 
3:20 AM
Then I will delete my account one day, lol.
@seaturtles Still waiting for you to tell your name, lol.
 
le sigh
 
@FernandoMartin ?
 
did you think about the problem?
 
@JasperLoy Dissing people whose first name is Sea?
 
@FernandoMartin Nope.
I am doing AM exercises though.
I should sleep.
 
3:22 AM
@KarlKronenfeld I am still confused between you and Arkamis!
 
Or... miss CA.
 
The teddie is back.
 
tb?
 
No, Ted Shifrin in chat.
 
Hi @Mike. You mean at grad school?
 
3:24 AM
oh, the later teddie
 
Hi @anon
 
Hi @ted, it's too late, lol.
 
You're always too late for something, Jasper.
 
I wonder what would happen if SE decided to remove all chat rooms.
 
Hi @user127001: You don't want to take my class? :) it's typically less abstract than analysis.
we'd get more work done, Jasper.
 
3:28 AM
:15150707
 
I am really shocked that his mum threw away all his math books.
 
@PedroTamaroff Nice number, there.
 
@KarlKronenfeld Yes. =)
 
Yes, Jasper, she sounds unfit to be a parent.
 
@KarlKronenfeld I forgot, are you a grad student?
 
3:30 AM
@JasperLoy nah
 
@KarlKronenfeld Are you gonna be one in future?
 
Jssper, don't you ask @Karl that every week?
 
@KarlKronenfeld Suppose $M\simeq N$, $A-$modules. $M'$ another $A$-module.
 
@JasperLoy I'd like to be one in the future.
 
Hehe, I have bad memory now.
@KarlKronenfeld OK, maybe we'll meet in future.
 
3:31 AM
Then $M'\otimes M\simeq M'\otimes N$, @KarlKronenfeld.
 
Anyone here good with elementary combinatorics? I can't figure out why I'm wrong here, and I'm headed AFK for the night... this is going to bug me; I can tell. math.stackexchange.com/questions/768266/…
(sorry for jumping into the middle of a conversation...)
 
Mine's definitely getting worse :) Seriously worried about Alzheimers ...
 
@JasperLoy Until then, I'll see you in my dreams, right?
 
@KarlKronenfeld Yes, you learn quickly.
 
Hi @anorton
 
3:32 AM
@TedShifrin I really hope there is a drug for it.
 
Hi
 
@PedroTamaroff Yuppers (tensoring with $M'$ is a functor after all).
 
Hi again @Karl
 
@KarlKronenfeld I was actually going to give a proof, but nevermind.
 
Hi, @Ted, not sure if I greeted you earlier.
 
3:33 AM
An old woman in the streets stopped there and started peeing. The urine flowed to my shoes.
 
No harm @Karl
 
I think maybe she has Alzheimers.
Anyway, Spivak has written a mechanics book I just realised.
Maybe he will write an entire series on physics for math people.
I lost interest in physics once I realised there was too much math even without it.
 
His goal 40 years ago was to understand through general relativity. That's what started him writing/learning differential geometry.
Now he's getting old ...
 
If I were interested in physics, I would buy the ten volumes by Landau and Lifshitz.
 
3:39 AM
I'm going to read Feynman when I retire.
 
It seems to me that Feynman isn't really good for learning physics. It's not a usual textbook.
 
He's masterful at giving the ideas. I don't want to learn pedantically, inch by inch.
 
I am waiting for someone to write say ten volumes on all of mathematics, similar to Bourbaki but covering almost all topics.
I guess that will never happen, unless I try to do it myself, lol.
 
Not likely. There is the Russian three-volume survey of math, written 40 years ago. But nothing like Bourbaki ....
 
What would the purpose of that be?
 
3:43 AM
@KarlKronenfeld For an interested student to learn almost all topics in a concise and consistent manner
It's really my dream to have such a series of books.
 
I don't think that is even remotely possible
 
I concur.
Have you tried learning from Bourbaki, Jasper?
 
By almost all topics I mean that many can be omitted. I think it is possible.
@TedShifrin The level of generality is too high for it to make sense to me. But the general topology was good.
 
@FernandoMartin I'm trying to find an injective $A$ module morphism $A^n\to A^m$ with $n>m$.
 
Well, there's the GTM series
@Pedro: That's the hard problem
 
3:46 AM
Precisely. You need to be very advanced to learn from it ... Basically a summary for experts.
 
@FernandoMartin Oh?
 
But hey Landau and Lifshitz could do the ten volumes for Theoretical Physics!
 
@PedroTamaroff You can pick your own $A$?
 
@KarlKronenfeld Yeah, I have to find a counterexample.
I was trying $\Bbb Z$. =P
 
I could give a hint @Pedro
 
3:47 AM
:P
 
Tensoring with $\Bbb Q$ makes $\Bbb Z$ unlikely.
 
2 mins ago, by Pedro Tamaroff
I was trying $\Bbb Z$. =P
try that some more
 
@seaturtles I think it should work, yes.
 
sips drink
 
sips saliva
 
3:50 AM
 
love that one
 
@PedroTamaroff That looks like you.
 
@JasperLoy Of course, I'm bold and noseless.
@seaturtles I think I got it.
 
I think Thai ghost movies are the scariest kind.
 
that's nice
 
3:52 AM
English ghost movies are the least scary.
 
@PedroTamaroff with $\Bbb Z$? popcorn popping
 
@seaturtles What is?
 
Show me, @Pedro. I'm befuddled.
 
I'm befuddled as well.
$\mathbb Q$ is flat, etc.
 
Anyone read Knapp's Basic Algebra, Advanced Algebra, Basic Real Analysis, Advanced Real Analysis? They cover so many topics...
I thought of buying them once, but did not for various reasons.
 
3:55 AM
Nope.
 
Hi @Ted. I had to make a decision between 'number systems and foundations of analysis' and differential geometry and I thought the number systems course would serve me better for my math background
 
@user127001 It's good to do analysis before differential geometry which uses analysis.
 
@KarlKronenfeld Maybe I should try a matrix ring?
 
@Pedro
 
I'd need to see syllabi and know the teachers to give good advice, @user127001, so I won't. :)
 
3:57 AM
please let me give you a hint
 
@Jasper the differential geometry professor told me I don't need any analysis
 
@FernandoMartin Did you read it somehwere?=
 
Our undergrad diff geo courses are not abstract, @Jasper
 
I ended up looking it up, yes
 
Is it cray cray?
 
3:58 AM
kind of
but I'd consider reading the question again
 
@FernandoMartin Why?
 
Unfair @Fernando
 
because you said you were asked to find a counterexample
 
No, I didn't.
I wanted to get one.
 
@Ted do you know either Nick Metas or Norman Weiss
 
3:59 AM
Example? Counter?
 

« first day (1360 days earlier)      last day (3666 days later) »