« first day (686 days earlier)      last day (4634 days later) » 

04:00
He called my $\LaTeX$ lame or something of the sort =(, =P
heh. Frank does submit a fair number of suggested edits IIRC.
@anon i think iyengar asked this question on the main site again.
@FrankScience I'm asking why do you include the parenthesis. It is so superfluous.
That Wikipedia uses them means nothing to me.
04:04
@Peter It's convention.
@FrankScience Useless one, then.
@Peter $a\bmod b$ means a binary operator $\bmod$.
@Eugene As it turns out, a finite inverse system with unique maximal object is isomorphic to that object. (In reference to the problem I posted yesterday.)
@FrankScience No it doesn't, unless you define it as such.
@anon wow. that is weird.
04:06
(I say "as it turns out" ironically, because it should have been blazingly obvious)
wait.
i guess not.
hahahah
when i first read what you wrote i thought it was weird. then i saw unique there and i realized "duh!"
The Who is such an awesome band.
@PeterTamaroff that is true.
He's a Pinball Wizard ♫♪
But if instead we say a group $G$ is infinite and consider only subgroups of finite index, we will no longer have the trivial subgroup, and thus no longer the quotient $G/\{e\}$ in the system, and things become nontrivial (I think). It's essentially the profinite completion but without the stipulation that groups be normal, and we're working in the category of G-sets rather than groups.
04:09
@Peter OK, that's nothing important.
I wonder whether Dirichlet generating function is useful while solving the recurrence?
@anon i'm trying to think about this in terms of tate modules.
you mean dirichlet series?
@FrankScience As in $\sum \chi(n)/n^s$?
generating functions are useful for solving recurrences, or proving solutions
Yes, $\tilde A(z)=\sum_{n\ge0}a_nn^{-z}$
04:11
@FrankScience Well, a simple example would be another kind of GFs
Sorry for my mistakes, I'll input it again.
$A(x) = \sum a_n x^n$
You can prove the Binet Formula with that, for example.
@FrankScience If it's a recurrence that works with divisors of the index, the dirichlet series might help.
@Eugene I'm reading WP on it now..
@anon I've see one example in cs.SE, $f(n)=f(\lfloor n/2\rfloor)+f(\lceil n/2\rceil)+e_n$
and a dirichlet series is supposed to help with that? was it an asymptotics problem?
04:14
$\tilde A(z)=\sum_{n>0}a_nn^{-z}$
@anon He gets an asymptotics result.
What is $a_n$?
@anon But it's all greek to me.
@PeterTamaroff an arbitrary sequence
@FrankScience πραγματικά?
@anon btw, i'm a trainwreck all the time!
04:16
@Eugene Oh, don't be mean with yourself! Be mean to others!
@Peter It's all Greek to me means that I have no idea on it.
@FrankScience I know!
@anon Maybe this one is more interesting.
I need to prove that $$1 > \frac{n}{ \sigma (n)} > \prod_{i=1} \left(1-\frac 1 {p_i}\right)$$ where $p_j$ are the prime factors of $n$. (Without Euler's totient, for sure).
@anon I guess I have used a wrong tool in my problem
And then $$ \frac{\sigma(n!)}{n!}\geq H_n $$
Curioser and curioser.
Well, since $n$ is in the sum defining $\sigma$, of course $\sigma(n)>n$. (except when $n=1$). For the right inequality you'll first need that $\sigma$ is multiplicative...
04:23
@anon I have proven that already.
@PeterTamaroff That one's obvious. $n!H_n$ is a subsum of $\sum_{d|n!}d$.
@Peter $$\sigma(n)=\prod_{p\mid n}\left(1+p+\cdots+p^{v_p(n)}\right)$$
What is the $v$ up there? I write that differently.
@anon I think I got it.
$v_p(n)=\sum_{k>0}[p^k\mid n]$
@FrankScience You're rigorous. I like it! =)
@anon Yeah, grabbed it by the horn.
Thus $$\frac n{\sigma(n)}=\prod_{p\mid n}\frac{p^{v_p(n)}}{1+p+\cdots+p^{v_p(n)}}$$
@FrankScience You're spoiling it for me.
@FrankScience You're lucky I'm listening to Stevie. Else I'd get mad.
$$\frac{1+p+\cdots+p^{v_p(n)}}{p^{v_p(n)}}=1+p^{-1}+\cdots+p^{-v_p(n)}>1+p^{-1}+\cdots=\left(1-p^{-1}\right)^{-1}$$
@FrankScience Are you serious?
04:32
Chang $\rt$ to $lt$.
$\rt$ to $\lt$.
No
I'm wrong, sorry
Change $\gt$ to $\lt$
where $1+p^{-1}+\cdots$ means $\sum_{k\ge0}p^{-k}$ (infinity sum)
@Eugene are tate modules relevant to nonabelian theory?
Incidentally, can we give an asymptotics for $\sigma(n!)$?
plug $n!$ into asymptotics for $\sigma$? :P
@anon LOL Maybe use Stirling then
or involve $v_p(n!)$ and make things messy...
04:40
It is too difficult to manipulate the primes.
@FrankScience What asymptotics do you ahve for $\sigma(n)$?
In addition to probability generating function $G(z)=E(z^X)$, is there any generating functions works well to determine the expected number $EX$ and the variance $VX$?
@anon Can't we use Lagrange's Identity somewhere?
@PeterTamaroff Impossible. For example, when $n$ is prime, we have $\sigma(n)=n+1\sim n$; when $n=2^k$, we have $\sigma(n)=2^{k+1}-1\sim 2n$
Can't you just say $O(n)$?
04:43
what do you mean by Lagrange's identity?
@PeterTamaroff lol, for $\sigma(n!)$?
@anon No, for $\sigma(n)$.
That means that $\sigma(n) \sim C n$ right?
no, it means $\le Cn$ for all $n\ge N$ for some $N$ for some $C$...
@anon Oh, right.
I think $\prod_p\left(1+p^{-1}\right)$ diverges.
I'm not sure.
they're related to elliptic curves.
04:47
@FrankScience It does.
@anon Oh, yes, because $\sum_p p^{-1}$ diverges.
You don't even need that.
@anon Hehe yes, overkill.
Frank got it.
:5030511 One thing.
derp
$$\frac{\sigma(n!)}{n!} =\sum_{d \mid n!}\frac 1 d $$
04:51
Sorry, I have no idea on calculus.
I think equality is only true for prime $n$.
@FrankScience What do you do?
@Peter What?
@FrankScience What are your interests?
@PeterTamaroff I view that as a corollary to $\sigma_\alpha(n)=n^\alpha\sigma_{-\alpha}(n)$.
@Peter I'm now very eager to sombody solving my problem.
04:53
@FrankScience Which one?
@anon Well, it is the $d \mapsto n/d$ "trick"
yes
This one. I have to redescribe it because it's possible that generating function is not a good idea.
are the moderators still looking at flags? i've had some under review for awhile now
@anon $\sigma_\alpha(n)=\sum_{d\mid n}d^\alpha$?
@FrankScience Yeah.
04:57
@Eugene What do you mean by?
The divisor functions.
2 hours ago, by anon
I prefer divisor-sigma, $\sigma_s(n):=\sum_{d|n}d^s$.
@Eugene Why is it not a question??
@PeterTamaroff just read the comments. it can't be answered.
04:59
So, basically the OP has not wrapped their head around the definition..
yes
because whether or not this is a question is debatable.
@anon It's a very difficult one.
@anon There's a problem I've mentioned when I took part in math competitions, It's very old-story.
Eh?
@anon Prove that there are infinity many $n$ such that $n\mid \sigma_k(n)$ where $k>1$ is a positive integer.
@anon I was looking for the notebook and luckily found out this problem.
When $k=1$, it might be an open problem, so $k>1$ is necessary.
05:14
What about proving there are infinitely many integers such that $\sigma(m^2)=\sigma(n^2)$?
infinitely many distinct pairs
Burton says $(k,10)=1$ and $m=5k$, $n=4k$.
@anon Potato, patata.
patata is incorrect.
@anon Patata is correct. I can count to it.
Just as easily as I can count to potato.
Actually $\rm potato \equiv patata \mod \text{translation}$
I'm off. Be well, people.
05:24
@anon ?
Indeed.
@anon i flagged that too
@anon Pay attention to security. You should avoid private messages when snapshotting.
private messages?
The real security issue is I didn't wash any of the metadata.
For example, the browser you're using, the operating system, etc.
05:28
well i now know you live in a central timezone
I could care less if people know I use chrome and OS for my normal surfing.
And I've already given away my location a few times.
@anon Incidentally, how to show \hbox{!`} in MathJaX?
Does \hbox{!`} not work? What's it supposed to be?
$\hbox{!`}$
It's only !`, not the reversed !
05:34
no idea
it would be a pretty dumb troll for me to ask "what is an elliptic curve" on the main site right?
yes
fair enough.
unless it were April 1st
two months too late for that
05:37
For april 1st I changed my username, gravatar and profile to match that of robjohn.
hopefully MSE will be around next year
@anon that sound pretty cool.
can two users have the same username though?
yes
i should think of something really good for next year.
Eh? Can I ask for a proof for Goldbach's conjecture in April Fool?
In fact there are four "anon", four "anonymous", and eight "Anonymous"
@FrankScience pretty sure that was done, like three times just this AF
05:40
@FrankScience it's only funny if you ask for it in various restatements
Judging this answer, is it useful?
isn't it only useful if you find it useful?
I don't really think it's useful, but somebody has upvoted.
nobody has
05:44
i mean the answer has no upvotes.
the answer has one upvote and one downvote
I've downvoted because I want to set a bounty and I don't like to bounty him.
If there's no solution and the deadline comes.
he gets it?
No, the deadline will come.
so what happens to your bounty?
05:51
I've mentioned a rule in FAQ that if I haven't decided to give the bounty when the deadline comes, the bounty will be given to the solution which gets the most upvotes (if $\ge2$), and if somebody else upvotes him, the bounty will eventually be given to this April-Fool (at least, I think it is) solution.
yah that is irritating sometimes. upvotes here are incredibly random. as long as it looks somewhat credible it gets something
And I've mentioned that when I post a problem in fashion (for example, asymptotics, calculus, and so on), the upvotes are usually high, although I think it might not be a very good problem, but when I post a problem out of fashion (usually discrete and elementary), the upvotes are usually low.
Incidentally, is formal-power-series taught in most colleges?
yes
In which branch?
difficult questions on this site are without glory
@FrankScience take your pick
05:59
@Eugene what?
@FrankScience it's basically taught in every branch of math from analysis to number theory.
@Eugene Thanks.
But the problems are too elementary to put into mathoverflow.
@FrankScience they are very important after all
@Eugene Formal-power-series?
yes
06:02
But I cannot tell the formal one from the other.
the other?
in which the convergence is an important problem.
ah yes. absolute convergence in a lot more useful IMO
sadly you don't have that a lot
But if it converges at $z=z_0$, the power series converges absolutely in $|z|<|z_0|$, but there're power series which diverge at any point.
For example, $G(z)=\sum_{n\ge0}n!z^n$.
that's why you need them as formal objects no?
06:07
Yes, and the differential equation is also formal.
ugh. i hate odes
By the way, where should I post the problems which seem difficult but elementary?
@Eugene Shakespeare liked them :-)
@robjohn he did indeed. i knew i'd pay for not capitalizing
@FrankScience need some elaboration
@Eugene You find odes odious?
06:11
@robjohn yes. both kinds.
For example, I'm not sure there's one solution for the latest problem I've told you
@FrankScience you can ask on meta if this seems suitable for migration to MO
@Eugene Is there any previous example?
@FrankScience of migration to MO?
just look for some on meta
@robjohn do you have any idea about 4(a)?
@Eugene Just like this?
06:15
i'm trying to get an intuition as to why that small area between the curve and the trapezoid is bounded above by the slopes of the tangent lines.
@FrankScience yes
@Eugene but it is said that More migration requests, flagging for moderator attention is enough. – Willie Wong♦ Jul 9 '11 at 9:57
@FrankScience then you can do that then.
i would suggest posting it on meta.MO first to see if it's appropriate for MO
@Eugene Thanks
and remember to link to your MSE question
en
I remember that I've posted a problem which is really elementary but no body answered until one pointed out that it was solved in 1963 by Don Knuth but with a bit tricky, clever but elementary technique.
06:19
i see
@Eugene That might be a sign typo...
so it is the sum
If it's $f'(k)+f'(k+1)$ over $4$ then I can explain it easily with three triangles.
@anon yeah
oh man. i meant meta.MO
06:25
Oh, sorry
What should I do next?
they have a "is this question acceptable" category
Oh, wait
i hope you have thick skin though because MO users aren't as welcoming as MSE
I remember that I should register for meta.MO?
@FrankScience oh? then just leave it as it is on meta.MSE then
06:28
@Eugene Did you register?
@Eugene I've registered on MO but not meta.MO
@FrankScience yes i did. it's not an issue though. qiaochu uses MO so he'll know if it's appropriate
do you need the answer in a hurry?
Actually, no. But without answer I feel painful. Can you feel this if you're waiting for an answer?
nope.
i keep myself pretty busy in the meantime
i'll give your meta thread it's first upvote so it get attention
Thanks. Incidentally, how to register in meta.MO?
the not signed in link
06:35
The righter is Sign in?
The Not signed is not clickable.
yes
Apply for membership?
yes
Well, I first know that apply has such meaning.
It's not appropriate to post in meta.MO now, but the next time?
@FrankScience well you're probably better off in meta.MSE. MO people are less welcoming sometimes.
06:39
Better off?
yes.
Can you explain in simple English?
as in it's probably safer to ask that here.
So vote to delete in meta.MSE and post in meta.MO?
just leave it in meta.MSE is what i'm saying
see asaf's comment?
also who knows additional attention might just get your question answered.
07:04
Yes
07:45
@Eugene are you still having trouble with 4(a)?
@BenjaminLim: hey there :-)
hey
just saw the ping from jordan saying he works harder but has no results to show for
@Jordan You know what, just shut it, I use to have pity for you but now, I am sick.
2
@robjohn next semester I am doing a several variables reading course with my supervisor
it's gonna be full on
@robjohn
@BenjaminLim that sounds intense
@robjohn 240 pages worth of
08:15
@Eugene: Since $f$ is concave, for $x\in[k,k+1]$, we have
$$
f(x)\ge f(k)+(x-k)(f(k+1)-f(k))\tag{1}
$$
Integrating yields
$$
\int_k^{k+1}f(x)\,\mathrm{d}x\ge\frac{f(k)+f(k+1)}{2}\tag{2}
$$
Integrating
$$
f(x)\le f(k)+(x-k)f'(k)
$$
yields
$$
\int_k^{k+1}f(x)\,\mathrm{d}x\le f(k)+\frac12f'(k)\tag{3}
$$
Integrating
$$
f(x)\le f(k+1)+(x-k-1)f'(k+1)
$$
yields
$$
\int_k^{k+1}f(x)\,\mathrm{d}x\le f(k+1)-\frac12f'(k+1)\tag{4}
$$
Averaging $(3)$ and $(4)$ gives
$$
\int_k^{k+1}f(x)\,\mathrm{d}x\le\frac{f(k)+f(k+1)}{2}+\frac{f'(k)-f'(k+1)}{4}\tag{5}
hello. I'm trying to understand wikipedia's explanation of positive-definite non-Hermitian matrices. I'm a bit stuck because I don't know what it means for a vector to be 0. What does it mean for a vector to be 0?
@MattЭллен all components are 0
@robjohn oh! ok, thanks :)
08:56
Hi ... does function has be be Vector function to satisfy Green's theorem??
09:14
0
Q: Green's theorem

experimentXLet the partial derivatives of Q be defined in the enclosed Region R.How to picture this geometrically? $$ \iint_S \frac{\partial Q_y(x,y)}{\partial x} dx dy = \int \left ( \int\frac{\partial Q_y(x,y)}{\partial x} dx \right ) dy = \oint Q_y dy$$ I don't understand how it changes into closed loop?...

 
2 hours later…
11:17
Holy cow I slept right through two appointments! 8-).
Which day is it today...?
Hmm, already Wednesday 8-(.
11:54
@BenjaminLim It is because you have things easy and you don't know what it is like when it is hard.
@Jordan Whatever, let the haters hate.
I am starting to get the feeling that no one here really had to struggle a lot in school
Your feeling is probably correct in some cases.
especially people like Peter or Ban that are 19 doing third or fourth year college level math
And they only reinforce my idea when they attempt to say that they are doing first year math, like they are completely oblivious to the world outside of their private school education
What do you mean?
12:10
People like them talk down to me a lot whether or not they realize it, saying things are so easy that I cant get at all
Well I find it weird that you find it suprising that people in a math forum excels at math It would be like going to a car showoff, and be surprised that people are good with cars.
And I guess most people struggle with various things in their lives, some with math some with other problems. But the difference is that they keep their wining to themselves, and just deal with their own problems =)
Well it is really difficult to keep a good attitude when I work so hard at something and not even once it can pay off
@Jordan What do you consider to be a "pay off" in math?
Proving the goldbach conjucture ^^
12:28
I just want to get a decent grade, it would be amazing if I could get an A on a test
but really the best I have done on a test is a B on the easiest test in a class I already took one time with a reall easy teacher who gives you half points just for showing work
so really I got like a D on the test
@Jordan You are not the only one getting mediocre grades.
@N3buchadnezzar But how many people put in as much work as I do and have to take each class twice just to get a C or B?
12:57
Hello freinds.
Anyone knows the "Newton's 350 year old puzzle"?

« first day (686 days earlier)      last day (4634 days later) »