« first day (1264 days earlier)      last day (3771 days later) » 

5:00 PM
@Huy, the flaw, and the appeal probably, of that video is that it's very high-school-math-ish
high-school algebra and such
 
Huy
It saddened me that my friend who has a MSc in EE didn't notice the mistake ._.
 
high schools do not teach students that integers add up to fractions
 
my point was they learn the most basic rules of algebra
 
integer + integer = integer
is that basic?
 
um infinite sums?
 
5:06 PM
infinity is not an integer
 
@skull : As I have said numerously many times to numerous people, it depends on the sense you are talking about. If cauchy, then everyone has viral infection. If abel, ramanujan got serious mental disbalance problems, etc.
@skull Correct.
 
@skullpatrol, if you are curious there are a lot of fun threads to read about this
 
yes, they are fun :-)
 
@Peter Just to point out, in the Cauchy sense, see Riemann rearrangement theorem. A generalization to in turn divergent series are possible, though.
 
the point were it made click for me was this blog post terrytao.wordpress.com/2010/04/10/… but that is also linked in 9/10 of those threads
that was a few months back
to just refer to 'analytic continuation' at that point is a bit heavy handed - i like that he presents a real analysis alternative (that part clicked with me)
 
5:21 PM
1 + 2 + 3 + 4 + ... = -1/12 often crops up in Casimir effect... Just sayin'.
 
context, context, context please
Just sayin'
 
@BalarkaSen, yeah that's what Mr. Motl opens up with in his answer to the main math.se thread on the topic math.stackexchange.com/questions/39802/…
which is also liberally linked on the social media
@skullpatrol, you started with the non-sequiturs
 
@skullpatrol I am definitely not physicists, so I cannot provide further context, admittedly. I saw this one explained in Marty Green's blogpost once.
:P
 
@skullpatrol Hello skull, I'm fine. I see that you have come back to your original naming scheme?
 
@PeterSheldrick I know, I was being a numb skull
@JayeshBadwaik yep
@BalarkaSen we definitely need the Physics here :-)
 
5:27 PM
Right.
 
@skullpatrol How are you doing nowadays?
 
@JayeshBadwaik Fine thanks, how are your studies going?
 
Okay, here's where I have read it : marty-green.blogspot.in/2010/03/…
 
@skullpatrol My semester has just began, and I'm slacking off now, have to gear up soon though.
 
5:29 PM
ic ic
 
@Amr, hmm pretty advanced... complex analysis? (not all of it, but an appreciation at least)
 
Amr
@PeterSheldrick Rudin's real complex analysis suffices ?
 
sure
'the hottest topics of contemporary mathematical research'
is it really that hot?
it's been around...
 
Amr
@PeterSheldrick No idea, but I am intrested to learn geometric analysis
 
that part is definitely hotter than 'Riemannian Geometry'
and to be honest i saw some 'geometric algebra' but 'geometric analysis' is actually new to me
 
Amr
5:34 PM
@PeterSheldrick What other prerequisites do you think I might need ? I suppose differential geometry
 
Amr
Are functional analysis or PDE s needed
 
just reading the description... makes it sound pretty hot
 
Amr
:)
 
not sure about functional analysis
 
Amr
5:36 PM
hmmm ok
 
can't hurt...
 
Amr
Sure
but I have lots to self-study
 
physics from the sound of it
 
Amr
What do you mean ?
 
like, physics as a prerequisite
functional analysis is a bit technical and is used for example in numerics
but this sounds more like a theory extravaganza - not sure if they care much about actually solving stuff numerically
but i'm just guessing from the description
 
Amr
5:38 PM
@PeterSheldrick I don't understand what you meant by : "physics from the sound of it"
 
proper theoretical physics is probably a prerequisite from reading the description (from the sound of the description...)
 
Amr
ahh oh no
i thought this was a pure math book
 
so you have 'quantum field theory' in the description
that's not just quantum theory but also field theory (for example in electromechanics...)
 
Amr
i thought they would solve mathematically intresting problems that are also occur in QFT, but QFT wouldn't be needed as a prerequisite
this is just a guess
 
okay, find the book in a library and base any purchasing decision on that first
 
Amr
5:46 PM
@PeterSheldrick OK. Thanks a lot for your help. I will need to leve now to continue studying Rudin's book Bye
 
later
 
@Huy No - what if you pick $A$ to be the complement of $D$?
 
Huy
@Mike: Right.
@Mike: Could you by any chance help me here as well? chat.stackexchange.com/transcript/message/13276321#13276321
 
The actual method to proceed isn't immediately obvious to me. Did you solve that one yet?
 
Huy
5:58 PM
@Mike: It is a theorem in our lecture notes including a proof, if you want I can copy it, it's not too long.
 
I'd like that. What do you mean by $\text{Gl}(X)$ in the link you just sent?
 
Huy
Linear, continuous and invertible, I think.
(with linear, continuous inverse)
 
And $\sigma(A)$ is the spectrum of the operator?
 
Huy
Exactly.
 
Ah, well by definition of the spectrum, $(A-\lambda)$ fails to be invertible.
Hmm... that's not true
Give me a second
 
6:03 PM
Something to downvote: http://math.stackexchange.com/a/643970/23353
Copy/pasting answers isn't good. I almost wonder if this is a spambot that is trying to build credibility. (I already flagged for moderator attention.)

I only mention it here because, if it is a spambot, we could see more behavior like this.
 
@Huy Do you have the bounded inverse theorem? I.e. that a bijective bounded linear operator has a bounded inverse?
 
Huy
It doesn't sound familiar.
 
Hm.
Well, if you had that, it's clear why your function must be either not injective or not surjective: because if it were bijective, it would have a bounded inverse.
 
Huy
Is the proof to that theorem simple?
I know that if it's bounded it will be continuous.
Oh.
I had the theorem of open maps.
As it is linear and continuous and bijective, its inverse is linear and continuous as well.
Hence bounded.
 
No, the proof I know uses open mapping theorem, which requires Baire Category theorem
 
Huy
6:07 PM
Or am I making a mistake?
 
Oh, if you have that theorem, you're golden.
 
Huy
Exactly.
I have the open mapping theorem and Baire Category theorem as well.
I guess maybe somewhere hidden in the lecture notes that "theorem" was mentioned but I overlooked it. But it's clear now.
 
Yeah, it's an easy consequence of open mapping. But open mapping is a lot to assume.
 
Huy
I don't know about that. But now I see why it is in the chapter of "principles of functional analysis". :P
 
Well, I just mean that Baire is a hard theorem, so that I wouldn't assume every course covers it.
 
Huy
6:10 PM
Oh, okay. I don't know, I actually liked the Baire part more than the part about reflexive and seperable spaces.
 
I agree.
 
Huy
The proofs were much less abstract, at least to me.
 
It's an awesome theorem, I think.
 
Huy
I thought Riesz' representation theorem was mind-blowing :o
 
Did you do the Riesz representation theorem for measures, too?
I think I like that one even more.
 
Huy
6:13 PM
I don't know about that one. We just did it for Hilbert spaces.
To be honest, I kind of lack knowledge about measure theory because I didn't enjoy it very much.
 
measure theory is the pits ._.
 
I mostly agree with you, I found a lot of measure theory very tedious and with no really exciting results, but I think you'll like this one
Let's start with a locally compact hilbert space X, and consider the continuous, compactly supported, real-valued functions $C_c(X)$
 
I don't understand these things yet, but I am starting a Coursera course on Functional Analysis in a week. Maybe in a few months I'll have it down.
 
Now, given a functional $\phi: C_c(X) \rightarrow \Bbb R$, with $\phi(f) \geq 0$ if $f(x) \geq 0 \forall x$, then there exists a unique regular measure $\mu$
With $$\phi(f)=\int_X f d\mu$$
 
I find the frequent use of compactly supported functions in math to be a double-edged sword. For some applications, like statistics, its totally useful. For others, like physics, it makes things a real pain.
 
6:18 PM
Hello there! I am studying some basic Trigonometry and I'm stuck in a problem and hope you guys can point me out how to continue. It says:

"Determine the three cube roots of (2-j)/(2+j) giving the result in modulus/argument form. Express the principal root in the form a+bj, where j^2=-1"

I first calculated (2-j)/(2+j), which gives 0.6 - 0.8j (a+bj form). Then I calculated the modulus (r^2=a^2+b^2, it gives 1 in this case). But now, how do I continue? I do not get the first part of the exercise (I know what a cube root is of course, but...).
 
In other words, any such linear functional can be represented as an integral with respect to some unique measure. I believe this theorem can be restarted with a complex codomain as well, allowing the measure to be a complex measure.
 
Hullo @Mike.
 
Hi
 
Huy
@Mike: That does look pretty powerful.
Where did you learn about it?
 
It's used in the proof of the existence of a Haar measure on a compact group. I don't know the proof for a locally compact group, but I imagine it also uses it
 
Huy
6:21 PM
I have never heard of such measures.
 
@user2584228 That sounds like a question for Main, rather than chat.
 
Oh. A Haar measure on a group is a regular measure that's left-invariant; i.e. $$\int_G f(g) d\mu = \int_G f(hg) d\mu$$
or rather that $\mu(S) = \mu(hS)$
For compact groups it's also right-invariant, but that's not necessarily true for only locally compact.
 
Huy
I see.
 
@KevinDriscoll I did not want to open a question for this as I only need to have quick guidance about the text. I can do it nevertheless if you think it is better.
 
It's important (to me) because having a measure (or ability to integrate) on a group is rather powerful. It lets p-adic analysis really come into its own, for instance.
 
6:25 PM
Interesting : Find two complex numbers x and y such that 1) if one of them is transcendental then the other is not 2) both have transcendental status yet undecidable.
Ah, I found one with a respectively weaker condition that (1) applies.
$\log(\pi)$ and $\log(\log(\pi))$
If the first is not transcendental then the other is not.
Ughh, that's too trivial.
 
Any ideas?
 
Huy
@Mike: Let $(x_k)_k$ be dense in $M$, $a_0 \in A \subset M$. For $k,l \in \mathbb{N}$ with $A \cap B_{1/2l}(x_k) \neq \emptyset$ choose $a_{kl} \in A \cap B_{1/2l}(x_k), a_{kl} = a_0$ otherwise. Then, $A \cap B_{1/2l}(x_k) \subset B_{1/l}(a_{kl})$ and $$A = \bigcap_{l=1}^\infty \bigcup_{k=1}^\infty (A \cap B_{1/2l}(x_k)) \subset \bigcap_{l=1}^\infty \bigcup{k=1}^\infty B_{1/l}(a_{kl})$$ hence $(a_{kl})_{k,l}$ dense.
 
Ah, that makes sense. I like that argument.
 
Huy
@Mike: How long ago did you learn functional analysis?
 
Fairly recently. I first took a class in the summer.
And then I had a (rather unsatisfactory) semester-long class that I supplemented with Rudin somewhat.
The summer one was very short and we covered things too quickly to get a real grasp on anything.
Our teacher tried to cover the entirety of functional analysis in about six lectures and then moved on to the representation theory of Lie groups...
 
Huy
6:41 PM
A very interesting approach.
 
I got nothing out of it, personally, except for the first couple weeks of functional.
 
7:07 PM
Is it true that the number of Generators of Z_a X Z_b = # of Gen of Z_a * # Gen (Z_b)?
 
are those intended to be arbitrary groups
actually even if they are, no
 
they were, but after some thought i think a,b must be rel prime
 
i'm not sure what you mean then
$\Bbb Z_a$ is always going to have one generator
 
?
Z_5 has every element cept e a generator
 
oh, I misunderstood
I thought you meant the number of elements needed to generate the group
 
7:12 PM
no the actual count of hte generators
 
well, the number of generators of $\Bbb Z_a$ is $\phi(a)$
and by the chinese remainder theorem if $\gcd(a,b)=1$, then $\Bbb Z_a \times \Bbb Z_b = \Bbb Z_{ab}$
 
ok thats basically my question/ answer
 
yeah
 
guess Ill go read the chinese remainder theorem
 
you then have what you want by multiplicativity of $\phi$
 
7:14 PM
neat. thanks for the pointer
 
no worries
and yeah it's definitely false in general, $\Bbb Z_2$ has one generator, while $\Bbb Z_2 \times \Bbb Z_2$ has no elements that generate the whole group
 
yeah that was a duh moment for me
i learned these things once upon a time and its all slowly popping back into my brain
alright, back to reading - thanks again - toodles
 
hi @Mike
 
Hi @ted, it's too late to say hi.
 
it is? @Jasper
 
7:25 PM
I hate it when I try to gently nudge somebody to solve the problem themselves, and then some dork jumps in and posts a complete solution.
 
That's what I've been bitching about for months, @Daniel.
You've noticed I've basically quit answering.
I've gotten angry a few times and had the offender retract his solution.
 
@DanielFischer That sounds like me, lol. I like to post more or less complete solutions, but those are elementary ones.
 
still, @Jasper, "elementary" askers need to be made to think, too, not just fed the answers
 
@TedShifrin There are still a couple of questions left where a complete answer is adequate. And sometimes, the nudging works, that's always a great joy.
 
Yes, I've actually gotten some serious thanks from people whom I nudged and worked with ...
 
7:28 PM
@DanielFischer Wow, now you sound you are really committed to this site! I don't get a great joy just from some website on the internet.
 
I'm gonna have to bitch out my undergrad diff geo class ... about half of them didn't leave enough time or make enough of a serious effort on homework .... and there's no way they'll pass at this rate.
 
hi @Ted
 
@JasperLoy It's a great feeling when teaching works.
 
Hi @robjohn, do you know if an estimate of the type $|\pi(z)| \leq C (1+|z|)^N \exp(B|\Im z|)$ holds for all $z \in \mathbb C$ for some positive constants $C$, $N$, $B$, where $\pi(z) = \frac{1}{\Gamma(1+z)}$?
 
@Mike must actually be working today ... sooo quiet ;P
 
7:31 PM
@Ted I was also having that problem last semester with some of my students. They'd do half the problem set correctly and then just not do the second half.
 
@Ted Doing homework for the finance class. I should still drop it eventually.
 
although you might learn something interesting, I doubt the marginal learning is high enough :P
 
Solved two old problems today.
Oh, hullo @TedShifrin
 
@Ted I intend to just sit in on the lectures, but I haven't dropped yet
I think I'll see whether I can pass the first exam without studying for even a minute before officially dropping.
 
@Mike whats the title of the course?
 
7:32 PM
frustrating, @Kevin. I warned them specifically the first few days ... And some of them are my advisees. Even the 8 I've taught before didn't all exactly ace the assignment. I figure that if they have to do 5 problems, only 1 of which is specified, and they get to choose the level/content of the other 4, it shouldn't be hard to get close to 100%.
 
@Kevin "Mathematical Finance"
 
That'll make the professor feel good, @Mike.
Our course is called Mathematics of Options Pricing.
 
@Ted Well, don't tell him then!
 
@Ted @Mike Is it all Black-Scholes and such?
 
(He knows I intend to drop it and invites me to continue sitting in)
 
7:33 PM
Then no problems,@Mike. Frank F. is the only person on the faculty i actually know personally.
 
@Kevin Not even that yet. We've only just defined Brownian motion.
 
Our course started with a "review" of probability for the 95% of the class that hasn't yet taken the probability class. It's not a prereq.
 
@Ted That does sound totally reasonable. I have to think about what a behavioral economist would say though
 
say about what, @Kevin?
 
@Ted About allowing your students to choose a large subset of the problems they get to complete
 
7:37 PM
@Kevin: I started doing this about 10 years ago. This is the unique course I teach that way. I divide the problems into three levels (flower, pinecone, and pyramid) :P Generally, the most advanced students do more of the latter two, and the C student who needs to stick more to computations does mostly flowers. The one required problem on each problem set and stuff done in class forms the basis for exams.
 
@TedShifrin Flower, pinecone and pyramid? Very exotic choice!
 
I thought you'd appreciate the poetry of it, @Jasper :) I have little pictures on the left :)
 
@Ted Well, that sounds like some solid results and a good way of catering to disparate students
 
It's not fun keeping the grading scorecard with 34 students, @Kevin. That's about double the size I usually have, but I doubt I'll have 34 at midpoint.
 
@Ted Fall of 2012 I had 41 students and it almost killed me. I think I averaged 20 hours/week grading their homeworks.
So I feel your pain
 
7:43 PM
Did you abolish your account and start over, @Kevin?
 
@TedShifrin Abolish is another fanciful word, I just use delete, lol.
 
Well, you can delete me if you prefer, @Jasper.
 
@Ted Nope. I've never had much rep/activity on the site. I usually only ask questions if I'm really stumped and I mostly don't know the answers to the questions asked.
Why do you ask?
 
Oh, hello @Balarka
ah, cuz you'd been awol for a while and I guess I hadn't noticed your rep before.
 
Oh yeah, I've been around 400 for ages. Just been focusing on research.
 
7:48 PM
What properties does a function need for $f(0)=0=F(0)$
 
err... assuming $F(x) := \int_0^x f(t)dt$, then all you need is $f(0)=0$
 
hmm @Mike $F$ is a primitive of $f$
 
then you need it to have $F(0) = 0$ :P
there's nothing more you can say than that
 
Nothing such as being strictly increasing or decreasing?
 
nope
 
7:51 PM
Assuming $f$ is continuous, saying $F'(0)=0$ will give you $f(0)=0$. Otherwise, I have nada.
 
of course, the primitives of $f$ are all $\int_0^x f(t)dt$ up to adding a constant everywhere
so that if $F(0) = 0$, it must necessarily be $F(x) = \int_0^x f(t)dt$
but that's the best I can give you
 
I am specifically looking at $\int x^{1/x} \;\mathrm dx$ it is a crazy mf
 
I have opinions on that sort of integral that I think are controversial in these parts
 
@Alizter What is mf?
 
@Jasper m other f ucker
 
7:55 PM
A rude phrase, hence abreviated.
 
@Mike LOL
It could also be mezzoforte!
 
It could also mean mitten fest
 
Luckily for our colored tongues, it is none of those
 
Looks at tongue wait a minute
 
@Ted I'm already sick of this class and I've only just started the homework
I guess I'll just make the drop official tomorrow. :P
 
7:57 PM
@Mike What class?
 
howdy
 
Mathematical finance
 
@Mike Don't wait till it is too late.
 
@Jasper I have two more weeks
 
well, @Mike, you'd probably be sick of diff geo, too ... you'd like my pinecone and pyramid problems but would be bored by the routine ones.
Why is $x^{1/x}$ controversial, @Mike?
 
7:58 PM
Pinecone? Pyramid? I like those difficulty ratings, though I'm not sure how they correspond to medium/hard
 
May I ask a question : At the concatenation of two strings $x$ and $y$ , I guess the two strings belong to the same alphabet $\sum$ ?
 
pinecones are prickly but not deadly :) pyramids are difficult to climb. Flowers are child-friendly :P
 
@Mike I think it's pretty arbitrary choice of pictures!
 
@Ted The function's certainly not controversial, but my opinion on the beauty of evaluating integrals is, likely
 
I just explained, @Japser.
Oh, @Mike, I'm totally on your side.
 
7:59 PM
OIC.
 
I have nothing against those who like them, but I may never understand it myself.
 

« first day (1264 days earlier)      last day (3771 days later) »