« first day (1138 days earlier)      last day (3887 days later) » 

12:00 AM
@AlexYoucis Just "OK"? Grrrr...
 
Grounds for divorce is a great song
 
@AlexYoucis Listen to "the Fix", "One Day Like This", "Starlings", "Jesus is a Rochdale Girl" and so on.
MUSE I love too.
 
@PeterTamaroff Know them, not super big fans. Especially The Fix is eh.
I don't really like Muse haha
 
@AlexYoucis Try this
 
> We need definitions!
 
12:03 AM
@PeterTamaroff I will. But, while I do, you try this: one of my favorite songs. youtube.com/watch?v=RCitWSMLf90
 
@DanielFischer Feed me definitions! I will... produce... theorems!
 
I thought that was coffee?
 
@DanielFischer LOL, both.
@DanielFischer Point risen
@AlexYoucis I like "We're Away" by Elbow.
 
@PeterTamaroff One problem is that a lot of definitions are standard, and if all those would be included, we couldn't find the questions anymore.
And I bet the ones with unusual definitions would be the ones who don't include them.
 
@DanielFischer Yes, that is why I said "relevant to". Of course, if someone cannot see the importance of including the def, it would be strange if they can realize which definitions are relevant. Maybe people can get educated, who knows.
But say in Topology definitions vary wildly!
 
12:13 AM
Well, yes, there are people calling $T_4$ spaces normal and vice versa.
 
@DanielFischer Heh, I call a space normal if it is $T_1$ and closed disjoint sets can be separated by disjoint open sets.
 
@PeterTamaroff As it should be, normal = $T_1 + T_4$.
 
@DanielFischer ;)
 
@DanielFischer Wait a minute.. You don't like having a chain of implications $T_n\implies T_m$ when $m\le n$?
 
@KarlKronenfeld I do. But I also like every $T_i$ being one single separation axiom.
Also, compact = quasicompact + Hausdorff.
Seriously, though, I'm fine with both conventions, as long as people make it clear which one they use.
 
12:26 AM
@DanielFischer I think the compactness thing is usual in "lower" mathematics.
But then we start to stick to Bourbaki's standards, right?
 
Uh, I actually don't remember which definition Bourbaki used. Methinks it was quasicompact + Hausdorff, but I'm not sure.
 
@DanielFischer Yes, it is that.
ODE's make things run so smoothly.
 
Certainly less painful than pure power series manipulation. But I have the impression the OP wants the latter.
 
@DanielFischer Seems the OP is a masochist.
 
Heh, aren't we all sometimes?
 
12:33 AM
@DanielFischer Guess so!
@DanielFischer Have you ever checked Kolmogorov's Introduction to Real Analysis? It is an amazing book!
Kolmogorov and Fomin, to be fair.
 
No, never looked at that.
 
@DanielFischer Guess it is under your toes, but it is quite awesome.
 
Kolmogorov has a good reputation, so I can imagine it's a good one.
 
I deem an author great when it makes seemingly complex results appear either natural or easy. I guess this also comes with experience, but in general poor authors are those who entangle themselves with poor definitions or convoluted proofs.
I'm not saying a hard proof is bad, rather, that a proof is bad when it is made harder than it should be. Or when the ideas are mixed up, or unclear.
 
Generally, but some proofs are convoluted, no way around it.
 
12:38 AM
@DanielFischer I guess so.
 
Can I have a second opinion on math.stackexchange.com/questions/493876 am I being excessively unclear or is the OP, well, how can I say it?
 
@DanielFischer I have no idea what he means by the "nullspace" and the "image" of a vector space...
 
Well, yes. I went ahead and assumed he meant the null space and image of a linear map.
 
@PeterTamaroff There's implicitly a linear map $V\to V$.
 
Needn't necessarily be an endo, the codomain could be a different space.
 
12:44 AM
@DanielFischer I usually try to stay away from ill posed or unclear questions.
 
Wise man, @Peter.
 
@DanielFischer I wonder what you work in. You know your math! Did I ask this before? (Maybe I did, I'm a little dory-ish)
 
@DanielFischer Oops, right, I didn't think it through (or at all :)).
@DanielFischer I don't think there is a way to be clearer without giving a full proof. The OP needs to read your answer more carefully.
 
@PeterTamaroff Nothing regular. Sometimes I teach a bit, sometimes I program a bit.
 
@DanielFischer Ah, how nice. What do you teach?
 
1:00 AM
Rehi @Peter ... Just checking. Did you get my email?
 
@TedShifrin I couldn't take the time to think of your lines problem. I'll try to do that tomorrow. I don't think I can just answer it off the top of my head. I did get your email, thanks!
 
Oh, you don't need to answer it so fast ... and it won't be off the top of your head.
 
Mostly Maths and English, for those pupils who didn't learn it well enough in school.
 
@DanielFischer Oh? You teach in college?
 
No. Pupils as in 13-18 years old.
Problems in school? Engage a helper.
 
1:03 AM
@DanielFischer Since you said "didn't learn it well enough in school." Guess I missed high school?
 
Could have become a Taxi driver like so many others.
But I don't like driving that fast ;)
 
Food is here. Yummies. =D
@TedShifrin I am pretty sure it won't, heh. Wanna give me a deadline?
@TedShifrin Do you have any advice?
 
@Peter: No deadlines. It's up to you and whether you're interested!
Advice?
 
@TedShifrin A pointer.
 
1:18 AM
You folks sure eat late!
 
It's past ten here.
 
Precisely.
 
What is a decent dining time? 8, 9? Guess so.
 
In Europe I eat that late, but I prefer to eat 4-5 hours before bedtime. :)
 
@TedShifrin Well, then I'm safe!
 
1:20 AM
LOL, yeh, you don't get up at 7 am :)
 
@TedShifrin Nay. My productive levels are off the roofs at night.
 
If you didn't spend so much time here, imagine how productive you'd be :D
 
@TedShifrin But... OK. You've got a point.
Yet this helps me quite a bit!
 
That happens from time to time.
 
For example, explaining stuff to others is, in my experience, a very healthy thing for one's knowledge.
 
1:25 AM
Yes, it's good. It's one reason I did so well on GRE years ago ... But still ...
 
@TedShifrin So, shall I say farewell and get back to Landau? =O
We usually have good math talks here!
 
Sure :) You don't need me :)
 
@TedShifrin Well, not with number theory, I guess!
@TedShifrin Here, have fun!
 
Ha ha. In general, I prefer not to try to deliver a course lecture here :)
 
@TedShifrin I was being slightly sarcastic! Of course, that question is waaaaaaaaay too broad!
 
1:58 AM
@anon
 
y?
 
> if n has order 2^(m-1) mod 2^m then the units mod 2^m are cyclic generated by n, so there is only one subgroup of order two, so only phi(two)=one element of order two (namely -1). but this is false: 1+2^(m-1) has order two as well.
When you say "so there is only one sgp of order two, you mean subgroup of $\Bbb Z_{2^{\ell}}^\times$?
 
yes
 
Ah, OK.
I fail to see why. Let me think a second.
Oh, OK.
Because if $a^2=b^2=1$ then $n^k=a$ and $n^j=b$ gives $a=b$.
@anon Yes?
 
2:35 AM
huh? why which part of what I said exactly?
 
@anon I mean, there is only one element of order $2$ because it is a cyclic group. Pardon my derpness.
 
yes, an element of order 2^(m-1) would mean the units are a cyclic group (which is false)
 
@anon OK.
@anon I have a question.
@AlexanderGruber Hey there.
 
@PeterTamaroff y?
 
@anon Sorry.
I was meaning to ask.
Suppose I have an abelian group $G$.
Let $G_0$ denote the trivial subgroup.
Then there is a canonical way to produce an ascending chain of subgroups $$G_0\subsetneq \cdots\subsetneq G_m=G$$
Which goes as follows.
If $G'$ is a subgroup, define the index of $a\in G$ in $G'$ to be the least positive integer for which $a^k\in G'$.
Denote this by $|a|_{G'}$.
If $H$ is any subgroup, define for $a\notin H$ the (group) set $H'=(H,a)=\{ha^m:h\in H\;,0\leqslant m\leqslant |a|_{H}-1\}$
@anon What is this construction? Does it have a name? Does it have more general analogues?
It is used to prove an abelian group of order $n$ has exactly $n$ distinct characters, say.
 
2:53 AM
So what are the $G_i$s you were talking about?
 
@anon $G_1=(G_0;a)$, $G_2=(G_1;a')$, $\ldots$
The claim is that $|(H,a)|=|a|_H\cdot H$.
 
What is $a'$?
 
@anon Elements one picks that are not in $G_k$ (until one exhausts all of $G$)
 
my idea of "canonical" probs differs from yours...
 
This happens since $a\notin H\implies |a|_H>1$.
@anon That was a bad choice of words, sorry.
 
2:56 AM
I'd call it a cyclic filtration
cuz the factors (i.e. $G_i/G_{i-1}$s) are cyclic groups
 
@anon Awesome name. You'd call it that?
(Or did you pick the name somewhere?)
 
Well, filtration is a standard term.
although we'd talk about "series" in group theory
 
@anon "Filtration series"?
 
the idea is more useful in field theory I'd say
it's relatively obvious any finite algebraic extension can be obtained by finitely many simple extensions
this allows induction to prove e.g. primitive element theorem
 
3:11 AM
@anon Any finite multiplicative subgroup of $\Bbb C^\ast $ must be of roots of unity, right?
 
mmhmm
the elts of the sbgrp have finite order...
 
@anon Aha.
Dunno, sometimes I think there are hidden monstrosities out there.
 
there are
one of them is called the Monster
 
@anon ORLY?
 
3:14 AM
@anon When I was about 6 y.o. I thought 1,000,000 was so large one would die before getting to it.
That is, counting.
 
3:26 AM
You're not there yet :)
 
@TedShifrin Ah?
 
Well, are you? :D
 
@TedShifrin I never tried to do so! =D
 
Clearly @PeterTamaroff is not a scientist :-P
 
I have 28880 seconds in 8 hours, so I could spend about 35 days to count to 1M.
And assuming I count 1 per second, which is quite slow. Then again, I would take quite some time to say "novecientos noventa y nueve mil novencientos noventa y uno, novecientos noventa y nueve mil novencientos noventa y dos, novecientos noventa y nueve mil novencientos noventa y tres, ..."
 
3:35 AM
just give an existence proof by induction and leave the counting to the engineers
 
engineers prefer to count money
 
LOL
 
@AnthonyCarapetis Let's get all depressed by finding the greatest integer one can possibly count to before dying!
 
@PeterTamaroff: that set is not bounded above, do you mean the least? ;)
 
3:37 AM
@AnthonyCarapetis I was about to add that detail.
 
Most of us mathematicians can't do arithmetic well at all, and I'm far closer to dying than you all!
 
@TedShifrin That is very questionable.
Biologically, yes.
But the human society has much more variables.
I take the highway four days in the week. Then again, I like to think I am a cautious driver.
 
Well, true, every time I return from the 150 mile round trip to ATL, I'm amazed I made it alive.
 
@TedShifrin (The avid mathematician would commit suicide to disprove you, I am not such a mathematician.)
Or am I...?
 
I hope not!
 
3:41 AM
@TedShifrin I was quoting Cosmo.
 
On that note, tennis comes early. G'night. (Who's Cosmo?)
 
@TedShifrin From Fairly Oddparents. Quite a nice show for kids.
 
later
 
@TedShifrin Bye byes.
Back to Characters.
Mathematical characters, not cartoons, I mean.
 
Dirichlet characters?
 
3:45 AM
Don't misrepresent yourself! Hee hee.
 
@TedShifrin Didn't get it!
 
Characters are associated to representations.
 
@TedShifrin Heh, don't know about that... yet.
I remember one time I confused representation with presentation @anon! =)
 
4:14 AM
indeed
 
4:48 AM
Guys, I have a problem understand trigonometry.
In a triangle ABC,
AB = AC * tan(x)
That's so confusing.
tan(x) is supposed to be AB/AC
 
@Nick Yes, that is precisely what you have there.
 
so the entire thing will just prove AB = AB
 
$\overline{AB}=\overline{AC}\cdot \tan x$ gives $$\frac{\overline{AB}}{\overline{AC}}=\tan x$$
 
which is either 1=1 or 0=0
 
@Nick What do you need?
 
4:51 AM
O_O oh, I just realized. This equation is to help me find either AB or AC
=_= sorry for the question
I was sleepy and confused
 
@Nick ;)
 
Rarely if ever expressible as a ratio of integers.
 
@anon
 
yo
 
@anon I am wondering about the following.
In his little section on orders, Landau proves that:
$$|5|_{2^\ell}=2^{\ell-2}$$
And then that for any $\ell >0$; any odd number $a$ satisfies $$a\equiv (-1)^{\frac{a-1}2}5^b\mod 2^\ell$$
precisely for those $b$ in a particular class mod $2^{\ell-2}$.
$b\geqslant 0$.
He uses this in his chapter on characters for the following:
If $d>0$, $(d,k)=1$ and $d\not\equiv 1\mod k$, then there is a character such that $\chi(d)\neq 1$.
@anon Could we have done this with a prime $\neq 5$, or in another way?
It seems that he must rule two cases: if $d\not\equiv 1\mod k$ one either has $d\not\equiv 1\mod p^\ell$ some odd prime in $k$, or $d\not\equiv 1\mod 2^\ell$, and gives different proofs for this.
For the first, he uses this
For the other case, what I just quoted above.
@anon Thought you may know some other, less "elementary" method.
 
5:14 AM
I'll have to get back to you on that. I'm through half a handle.
 
@anon "I'm through half a handle."?
 
@PeterTamaroff He's drunk!
 
@anon Oh, beer.
 
haha water
 
"Handle" usually refers to liquor
 
5:16 AM
@anon I leave Landau's approach for you to see later.
@KevinDriscoll I thought of the big handle in beer mugs.
 
@PeterTamaroff Ah, good guess but no. Its the handle on a liquor bottle.
@PeterTamaroff Why is all this 'mod' stuff important? I always thought of the 'mod' function as just an odd curiosity that was useless
 
@KevinDriscoll My liquor bottles don't have handles.
@KevinDriscoll Do you cherish your soul Kevin?
 
I do!
 
@KevinDriscoll Being serious, as you can see, modular arithmetic is a very extensive and interesting subject.
@KevinDriscoll (I wonder who introduced you to modular arithmetic to make you believe that.)
 
hmm, I confused .75 with 1.75mL. disregard however badass I may have sounded.
 
5:24 AM
@anon What are you drinking, dude?
 
some weak sauce vodka. girly stuff is easier to down.
 
@anon HAHAHA, OK. I will join you with some Honey Moonshine and Rhum/caramel filled chocolate treats. Cheers!
FUUUUUUUUUUU
It has 36% vol/vol.
Thought it would be sweeter.
 
5:39 AM
@anon Aw, yiss.
I found the zip is more enjoyable if one lets it sink in the tongue a bit. Else it burns too much.
(Sort: active)
 
@PeterTamaroff: It makes sense if you're thinking of $f$ along a curve $x \mapsto (x, z(x))$
 
@AnthonyCarapetis True, but still nope.
 
?
it's physicist's notation
but that's the correct formula for $df(x, z(x))/dx$
 
 
3 hours later…
8:32 AM
Greetings high-minded people!
 
Greetings highest-minded person!!!
 
@cyberskull hehe :-)))))))))))))))) Hi! How are you doing?
 
@Chris'ssis Fine thanks, how are you my friend :D
 
@cyberskull happily, very productive!:-) Last night I created 47 new questions. Let me show you the last one.
 
user87637
@Chris'ssis How did you come up with so many?
 
8:37 AM
$$\lim_{N\to\infty} N^2*\left(\sum_{n=1}^N \arctan\left(\frac{1}{n^2-n+1}\right) - \arctan(N-1)\right)$$
@JasperLoy I think I've just reached the creativity peak in this life. I'm full of ideas.:-)
 
That is a beauty.
 
user87637
@Chris'ssis It's best to never peak. That means you will always be better!
 
@cyberskull yeah, it's amazing.
@JasperLoy hehe, I'm just a human being with ups and downs. It's OK for me. :-)
 
user87637
I think I am going to not log in to SE for a while, it's getting boring. No lhf for me!
 
@JasperLoy enjoy the beauty of things around and you'll never ever be bored again. :-)
 
8:43 AM
True dat^
 
 
1 hour later…
9:45 AM
I better let that limit for now.
 
@Chris'ssis I need to look at that...
 
@robjohn I just created that last night before going to sleep. It's very cute!
 
10:08 AM
@Chris'ssis $\lim\limits_{n\to\infty}\frac{N^2}{N^2-N+1}$
 
@robjohn right! I'm just preparing now another version of it.
@robjohn true
@robjohn can you see the problem from the comment I deleted? (That improper integral)
 
Yes
 
@robjohn :-) It's very beautiful. (I created it this morning)
 
10:24 AM
@Chris'ssis $0$
 
@robjohn I think the answer is different from 0.
As regards the limit above, I plan to create a form involving the double series. (It'll be rather tricky)
 
@Chris'ssis This is the one with the $\arctan^3(x)$?
 
@robjohn the improper integral evaluates $\pi^3/8 \log(\pi)$. Yes, it's $\arctan^3(\pi x)-\arctan^3(x)$ in numerator.
 
@Chris'ssis Yes, I am just now looking at the tail
 
@robjohn OK
 
10:33 AM
$$
\begin{align}
\int_0^\infty\frac{\arctan^3(\pi x)-\arctan^3(x)}{x}\,\mathrm{d}x
&=\int_0^\infty\frac{\arctan^3(\pi x)}{x}\,\mathrm{d}x
-\int_0^\infty\frac{\arctan^3(x)}{x}\,\mathrm{d}x\\
&=\lim_{N\to\infty}\int_N^{\pi N}\frac{\frac{\pi^3}{8}}{x}\,\mathrm{d}x\\
&=\frac{\pi^3}{8}\log(\pi)
\end{align}
$$
 
@Chris'ssis You have been changing avatars frequently. :)
 
@KarlKronenfeld hehe, yes. It changes automatically. :-)
@robjohn amazing solution.
@robjohn shot=solution
 
@Chris'ssis shot?
@Chris'ssis what were you thinking of for a solution?
 
@robjohn I was planning to use the differentiation under the integral sign.
 
@Chris'ssis Ah. well arctan limits to $\pi/2$, so it seemed the way to go.
 
10:47 AM
Yeap.
@robjohn when both limits of the integral tend to infinity, then we should always have in mind the squeeze theorem ...
Actually I said that to myself in the past, but it seems I neglect it sometimes.
 
@Chris'ssis It's an even function, so if we made the bottom limit $-\infty$, the integral would be $\frac{\pi^3}{4}\log(\pi)$
 
(surely, first of all we need to bring the integral into a convenient form)
@robjohn right
 
 
1 hour later…
12:02 PM
Is there something wrong with this question:
http://math.stackexchange.com/questions/494333/adding-zeta-function-limits-is-this-an-identity-for-any-complex-number-s-and-an
or is it not even wrong?
 
12:18 PM
@MatsGranvik What do you mean by is it wrong? It seems like a pretty well formulated question.
 
ok, just checking.
 
@MatsGranvik If you're looking to improve it (and perhaps give it a better chance of being answered), you might like to add some context about how you encountered the problem and what you've attempted to try and solve it yourself.
 
@DanielRust I will do that.
 
@MatsGranvik: for clarity, you might want to define the behavior of the If function.
 
12:48 PM
@DanielR ok.
 
 
1 hour later…
1:49 PM
How can someone even accidentally format a question like this? math.stackexchange.com/questions/494398
 
@DanielRust What's the problem?
 
It just looks a mess. One full stop with no space before or after it, one full stop with both a space before AND after it, a random line break in the middle of a sentence, no capitalisation. Even if English isn't your native language you should know basic writing skills.
 
2:16 PM
Let's do a quick survey here: what's the most beautiful math formula for you?
 
@cyberskull You must like triangles
 
$\int K = 2 \pi \chi$
 
$$\int_M K\: dA+\int_{\partial M} k_g\: ds=2\pi\chi(M)$$
 
Don't know too much calculus
What are these?
 
2:27 PM
@saadtaame It's the Gauss-Bonnet Theorem.
 
@DanielRust Curvature thing?
 
@saadtaame yep. It roughly says that the total curvature of a surface is only dependent on the topology of the surface.
 
@DanielRust How is the curvature of a surface defined?
 
@saadtaame In the Gauss Bonnet theorem it's specifically the Gaussian curvature (there are other types of curvature). The definition is here: en.wikipedia.org/wiki/Gaussian_curvature
 
2:40 PM
Do you know of a resource on the internet that sums up "everything" about integrals? I'd like to learn about them
 
@saadtaame That's a pretty tall order...
 
"integrals" is pretty broad :p
 
You're best off picking up some calculus textbooks
 
hmmm things that would serve me well in my self-study of topology and geometry?
 
I'd say a good working knowledge of calculus is a definite advantage if you want to start learning geometry (and less so topology)
 
2:44 PM
Hi @TedShifrin how are you?
 
I'm sure @TedShifrin can offer some better advice than me :)
 
@TedShifrin Please
 
Unless in topology you include my favorite, differential topology. Transversality theory is essential to a working topologist.
I butted in, @Daniel :)
Hi, @cyber.
 
@TedShifrin as you should :D
 
differential topology is a station in the way of course
 
2:46 PM
Ooh, I missed Gauss Bonnet?
 
I'm thinking about taking an online course this semester on Morse Theory actually
 
morse theory is cool!
 
@TedShifrin What do you recommend?
 
Gauss Bonnet is the overlap of differential geometry and topology — the baby case of characteristic classes.
 
I followed along with the lecture notes last year but I didn't really put any effort in so didn't feel like I really followed it that well
 
2:47 PM
Morse Theory is beautiful. Sigh @ online classes. :)
 
Yeah they're not great :-/. I have to take a certain number of them to qualify for my funding.
 
For what @saadtaame?
 
@TedShifrin Learn integrals for geometry and topology
 
The students sitting in my grad diff geo course who aren't working on homework exercises aren't getting much out of the course. That's just a fact of educational life.
 
Resources?
 
2:50 PM
You don't need all the tricky methods of integration, necessarily, but the concept of integral and integrating differential forms over manifolds is crucial.
What level are you at, @saad?
 
Beginner
 
That doesn't help me.
Be specific.
 
um, learned basic integrals at high school. Never looked at them ever since
 
What math have you studied and learned since?
 
not much, first chapters of a topology book. first chapters of Artin's algebra, some graph theory in my algorithms class..
 
2:53 PM
Weird disjointed ... You're in CS?
 
yes
math things i study by myself
 
You need multivariable calculus, linear algebra, and then real analysis and multivariable analysis.
 
alright
thanks
 
With an obvious bias, I recommend my multivariable math book :) Hubbard & Hubbard is popular.
 
do i have to do these in order?
 
2:56 PM
@saadtaame You can probably find some decent online resources too, lecture notes and the like.
 
But diff geo is quite multivariable-math intensive. Point set topology or algebraic topology, not.
 
If you're learning from Artin successfully, you're smart, so my two recommendations above woukd teach you theory and computation together, which is more interesting.
Yup. As I said, it's a biased recommendation :) But good enough for Stanford, Yale, Vanderbilt, among others:)
 

« first day (1138 days earlier)      last day (3887 days later) »