« first day (4450 days earlier)      last day (867 days later) » 

01:12
@robjohn So that guy who deleted his post (on Chern-Gauss-Bomnet) only to have you undelete, has deleted his identity and now has an unreachable user…… ID. What garbage.
 
2 hours later…
03:18
@TedShifrin I would assume that the question is a homework or exam problem and the user deleted their account to prevent anyone finding out they got help here.
@leslietownes One way I found is that: Using Schwartz reflection, extend the given map to $\Bbb D\to\Bbb D$ then use Schwartz lemma.
03:31
sad to see that kind of goofy stuff for high level questions. you almost expect it with 'contest math' and things where users might be more casual about wasting other people's time.
@onepotatotwopotato that's cool. not the approach i know but it seems like there's often a way of reducing complex analysis to the schwarz lemma.
@leslietownes Yeah, I’m skeptical, yet ….
03:56
Seeing as how I think I may have broke wolfram alpha, I will ask the pros, what do: $3\cos(3.7k) = 7\cos(3.5k)$
I do not remember trig at all can I just cancel the cosine functions to solve for k?
no.
for your application, are you actually required to solve this equation? or is it simply an expression that comes up in an analysis that might not need the value of k to x digits.
Shirley there must be a way
Well I have two sets of x,t values for a wave $s(x,t) = s_{max}\cos(kx-\omega t + \phi)$
3.5,0 and 3.7,0 and I'm asked for f which is $v=\lambda f $ and $k = \frac{2\pi}{\lambda}$
so I thought to solve for k
wait why did i set them equal to each other..
those are big parameters for a trig function!
oops, k=kilo in an engineer's muddle
04:12
yeah k is the wave number and omega is the angular frequency
it's a sound wave problem, giving me the position of two air molecules at time t = 0 with different displacements s_max
and asked me to find frequency, i'm also given the speed of sound as 340 m/s
i don't like partial problems where important parts are revealed bit by bit.
OH
that was not a comment for you, just a general observation
Yeah no worries, I just had an epiphany
just ignore the red laser flitting across your forehead
04:16
my epiphany is crumbling
i find it amusing that people just turf stuff like this out math.stackexchange.com/q/4549327/27978 without any preamble
epiphanies are made to crumble. don't give up.
i found some of the best came after some alcohol, but something evaporated between 0.08 and sobriety
I could go for a nice cheap rosee
i will sip a Mezzacorona Pinot Grigio shortly...
There's got to be a way to solve for k in this situation..
i dont know enough of your situation to hazard a guess. did you ask me some 'physics' question about a wave packet moving recently?
04:22
Yes I don't remember the problem exactly but I tried to word it in a more mathematical way
The issue is the h-bar has been very dead lately
at the times I use it anyway
i am having another epiphany
Nevermind I will ask my professor tomorrow :D
"Don't call me Shirley." One of the best lines from Airplane.
Yep :)
well you know a that a phase shift of $0.2k$ results in a from $s_\max$ to ${3 \over 7} s_\max$.
this lets you compute $k$.
linguistic comedy seems to be an older thing. I recall monty python being in the same genre of comedy sort of
It's on the tip of my frontal lobe..
indeed it lets you compute the wavelength $\lambda$ and from there $f = {v \over \lambda}$.
the $k$ is irrelevant really.
04:30
OH
you really need to draw a little picture for this sort of thing to see what is going on.
So you're saying set $\phi = .2m$
the point of the $kx-\omega t$ part is just the combined phase contributed by distance and time.
It seems odd to me that two molecules .2m apart can vibrate at the same time t=0
if you can figure the drop caused by a phase change of $0.2k$ then you can figure the entire wavelength (assuming there is no aliasing here).
04:32
like a wave can't travel instantaneously
have you even been in a swimming pool?
no one is saying it is traveling instantaneously
draw a picture of a snapshot of the wave at $t=0$ and note that a peak of $7nm$ is at $3.5m$ then (presumably) the first place it drops to $3nm$ it is at $3.7m$. No time or instantaneity involved.
from this you can figure out $\lambda$.
imagine taking a snapshot of the wave against the side of a swimming pool and how you would estimate the wavelenght.
@copper.hat as you might have suspected, that question is evolving. potentially relevant parts are being revealed bit by bit
aummmmmmmmmmm
I guess because this is a displacement function not the equation of the wave.. so yeah that makes more sense.
I was confusing the two, assuming the time tracks the evolution of the wave, but it's the displacement of the molecules
so as t increases the molecules experience different levels of displacement
i have no idea where you are pulling molecules from
04:40
the air molecules displaced by the sound wave
these problems are not experiments in creative writing
forget about that and just think of the function from which you wish to extract $\lambda$.
they are basically asking if you can compute $\arccos$.
I was trying that at the very beginning but I was told I cannot solve for k that way
not that it is relevant here, sound is a pressure wave (that is, longitudinal)
the molecule thing is someone's fancy writing.
Hmm yeah I will definitely give this another try tomorrow, it is past my bedtime. Thank you for your support mr. hat
or asshat as my friends call me
jk
04:48
lol
good luck & gn
tyty
Friends? Again you refer to … friends.
i had one once
decided to keep it all age appropriate here, don't want to risk a suspension like one of the cured folks
Should I even wonder?
04:55
the edit history on that post is funny.
it's been updated again with more hypotheses.
ted: copper being copper.
Which post?
i don't mind a question so much (in chat anyway) where details trickle out as the asker genuinely comes to a realization that other things might matter. but this post is clearly about some specific situation where they stripped all of the detail out in the first instance. they know that stuff might matter.
Oh, the silly L^2 post.
just save everybody's time and say hey, in paper X they claim Y but don't explain it. is it a special case of some general thing? i know it is not a special case of general things A, B, C that i thought about for a while, for simple reasons.
If each singularity of $f:\Bbb C\to\Bbb C$ is at most a pole, then we can consider $f:\Bbb C\to\Bbb C_{\infty}$ as an analytic function.
We can't allow essential singulairty.
The statements 'The meromorphic functions in the extended complex plane are the rational functions' and 'The analytic maps $\Bbb C_{\infty}\to\Bbb C_{\infty}$ which are not identically equal to $\infty$ are the rational functions' are equivalent statements?
Or the latter statement is stronger?
05:49
they might not be tautologically equivalent depending on the definitions of 'meromorphic,' 'analytic map' on those domains, etc., but they strike me as capturing the same statement.
In the usual sense: $f$ is meromorphic on $D$ means $f$ is holomorphic on $D$ except a discrete set of poles. And for $f:\Bbb C_{\infty}\to C_{\infty}$, $f$ is analytic at $\infty$ means $f(1/z)$ is analytic near $0$ and $f(z_0) = \infty$ is analytic at $z_0$ means $1/(f(z))$ is analytic near $z_0$.
But anyway it seems the statements are equivalent. The point is about the essential singularity but the continuity of $f$ ensures it cannot have an essential singularity because any deleted nbd of essential singularity has dense image.
 
3 hours later…
08:50
\begin{align*}
\int_0^{2\pi} {\cos\theta\over 5+4\cos\theta}\ d\theta & = \int_0^{2\pi}{1\over 4}{\cos\theta\over (5/4)+\cos\theta}\ d\theta\\
& = {1\over 4}\int_0^{2\pi} 1-{5/4\over 5/4+\cos\theta}\ d\theta\\
& = {\pi\over 2} -{5\over 16}\int_0^{2\pi}{1\over 5/4+\cos\theta}\ d\theta.\\
\int_0^{2\pi}{d\theta\over 5/4+\cos\theta} & = \int_{|z| = 1}{1\over 5/4+(z+z^{-1})/2}{1\over i z}\ dz\\
& = {4\over i}\int_{|z|=1}{1\over 2z^2+5z+2}\ dz\\
& = {4\over i} 2\pi i {2\over 3}\\
& = {16\pi\over 3}.\\
Something is wrong here...
09:16
Poll time guys
"A YouGov survey of 1,737 British adults, carried out on Thursday and Friday for The Times newspaper, showed support for Labour at 52%, down two points from a poll by the same organisation in late September, while the Conservatives were up just one point at 22%."
Question: Do you as a mathematician, accept the YouGov polls' accuracy when you consider the fact that there are about 46 million UK voters and only 1713 voters were interviewed by YG
 
1 hour later…
10:34
Yes and no. Really depends on what kind of "sample" they used. It is possible to skew the results based off of using the wrong sample and sample size. However, a bigger sample size doesn't always lead to more accuracy. At some point, a bigger audience just brings diminishing returns. Its hard to tell where that point is for this particular surgery, which is why I said, yes and no.
Survey*
 
2 hours later…
12:40
@TedShifrin Why does it matter if it's $+b$ or $-b$?
13:12
@leslietownes thanks Leslie :-).
@Koro
Hello
Hi @Shinrin-Yoku !
I want to ask one question now: Fatou's lemma remains valid if the hypothesis that $f_n$ is non negative measurable is replaced by the hypothesis that $f_n$ is measurable and $f_n \ge -g$ where g is non negative integrable. This is exercise no. 2.19 in Folland's book. I know how to do this.
@Koro Off the top of my head, I don't know, but an "obvious" thing to try is to consider the new sequence of functions $F_n := f_n + g$.
Or something like that.
13:18
Revised exercise: Suppose that $f_n$'s are integrable, g is integrable and $f_n\ge g$. Then Fatou's lemma's conclusion holds for $f_n$, i.e. $\int \liminf f_n\le \liminf \int f_n$
@XanderHenderson: yes, I know solving that exercise.
For this revised exercise: I have written a solution and I was told that the solution is not correct. But I believe that the solution is indeed correct. So can anyone please review my solution and let me know in case there is some flaw in it. I'm sharing the solution now:
Who told you that your solution is incorrect, and why did they make that claim?
Is the proof that the algebraic numbers form a field without any determinants given in any text?
@XanderHenderson one of my teachers did without exactly mentioning which part in the solution, but since they will be away from college for some days, I can't discuss with him personally.
@copper.hat @Thorgott tagging you as we discussed a similar problem very recently.
@Shinrin-Yoku what is 'without any determinants'?
That has no dets in the proof @Koro
13:25
@Koro I'll try to have a look in a bit. That being said, my first complaint is that, by the time you take measure theory, you ought to be writing in complete sentences.
@Shinrin-Yoku: you may look up Gallian's book.
@XanderHenderson: did I not write a complete sentence in the proof shared in the images? I'm afraid, I don't understand.
Thanks a lot for taking a look at that. :-)
Oh I think you mean the last line in the first image where I wrote " " "
yeah, that sentence itself is wrong. One says $h$ is integrable but not "$\int h$ is integrable.".
@Koro Part of writing in complete sentences is using only as much notation as you actually need in order to convey ideas. Symbols like $\therefore$ are, in my opinion, not helpful.
Ah I see. Thanks a lot for informing me that. I write it interchangeably with "hence". Come to think of it, I saw a post on mse, that criticised using $\therefore$ but I didn't actually see why.
13:31
And your very first statement isn't really a sentence. What does "$f_n \ge g = g^+ - g^-$" mean? Presumably, you mean something like "Suppose that $g$ is integrable, $g^+$ and $g^-$ are the positive and negative parts of $g$, and that $\{f_n\}$ is a sequence of measurable functions such that $f_n \ge g = g^+ - g^-$."
I mean, I'm being pedantic, but it helps to orient the reader.
14 mins ago, by Koro
Revised exercise: Suppose that $f_n$'s are integrable, g is integrable and $f_n\ge g$. Then Fatou's lemma's conclusion holds for $f_n$, i.e. $\int \liminf f_n\le \liminf \int f_n$
@Koro Sure, but that isn't in the actual answer that I am reading. I shouldn't have to look at two documents to understand what you are saying.
@XanderHenderson Oh I didn't incorporate it in the images. yes, $g^+, g^-$ are as you say. So is true for $f_n$'s.
I am making a general statement about mathematical writing---documents you produce should be as self-contained as possible.
Xander: I understand. The image I shared should be self contained.
13:34
And, honestly, I am saying this as someone who grades a lot of problems, but often forgets what the actual statements of those problems are.
It is really helpful to have it all in one place. It saves time, makes me happier, and makes it far less likely that I will grumpily take off points for annoying me.
Infact, 3 images are also annoying for a reader. (after seeing the content on 1st image, while reading the second image, what if I need to go back to the first image.) :(
I thought about that after emailing that to the teacher also. But the email was already sent.
:(
Are you using a dedicated scanner, or a cleverphone?
In either case, scan to .pdf. It makes everyone happier.
(or just learn TeX---you're taking measure theory; it might be time to start learning TeX)
I wrote this solution in ipad and then shared the pages to my gmail.
sending like this happens in images not in pdf in the writing app that I use.
Using what app? Notability, for example, will happily export a .pdf.
Notability
Oh yes!! There is PDF option as well.
I saw only two JPEG and PNG.
13:41
Yeah, I use Notability every day. It definitely has an "export to .pdf" option (at least, the desktop version does; I have no idea about the iPad version).
Okay, so I don't like the step you highlighted.
I need to think a bit more about why that is bothering me...
Let me explain that: So $F_n+g\ge 0$. Therefore, $\int F_n+g=\int (F_n+g)1_E+\int (F_n+g)1_{E^c} $ for any measurable set E.
$1_E (x):= 1$ if x is in E and is defined as 0 if x is not in E.
No, I know what the notation means. It is just that you have written that step, and then the second integral doesn't appear in the next line. It is hard to follow the reasoning. Again, a bit of language helps to explain where you are going.
the next line simplifies 1st term on rhs of highlighted part.
I agree I should have written that before sharing the images here.
Should I write that in mathjax and then share?
Ah... in that case, "The first integral simplifies to..." would help.
I really am critiquing your writing more than the math. Let me look at the math.
@XanderHenderson thanks a lot :-).
13:54
@leslietownes What do you think?
btw I modified the definition of annulus because of the continuity issue.
@koro I don't see anything obviously wrong with the mathematics of your solution, but it does seem like overkill to me. You claim that you already know that if $f_n \ge -g$, then you get the appropriate Fatou analog. So why not split your original integral into the region where $g$ is positive, and $g$ is negative, and apply Fatou on the region where $g$ is positive, and exercise 2.19 on the other region?
That's what made the solution 3 pages long. Because the Folland's exercise is not a theorem that was covered in my class so I can't write "by exercise 2.19 in Folland, we have...".
Also, my edition of Folland has a completely different exercise at 2.19 (it is on complex integration). What book are you working out of?
second edition
It's 2.18.
I made a mistake above. It's 2.18.
Ah. There it is.
14:03
Do you know of any website that offers exercises (also harder ones) to Calculus, for example? Or would you use a book for problems?
Honestly, I would give a proof of that exercise as a lemma, then prove the thing you are asked to prove.
@XanderHenderson I wonder what subtle errors the teacher was talking about then. :(
In any event, I have not thought deeply about elementary measure theory in at least a decade, so I don't know how I would prove either result off the top of my head, and don't see any obvious errors with your approach.
The basic structure of your argument looks fine, but that doesn't mean that I'm not missing something small (but important).
Thanks a lot for reviewing my solution even despite the write up not being self contained. :-)
I think (but not sure yet), that he referred to the following piece:
40 mins ago, by Koro
Oh I think you mean the last line in the first image where I wrote " " "
39 mins ago, by Koro
yeah, that sentence itself is wrong. One says $h$ is integrable but not "$\int h$ is integrable.".
Khan Academy for example seems to just offer easy problems that are auto-generated
 
1 hour later…
15:28
@Shinrin-Yoku not sure what you're thinking of, the standard proof dosen't have any determinants
Hi @Thorgott
Could you give a skecth
The only proof I am aware of is this :maths.tcd.ie/pub/Maths/Courseware/NumberTheory/ch12.pdf
@Thorgott
For a given function $Y(t) = (x(t),y(t))$, determine if $Y(t)$ is a solution system....Question: What exactly is a solution system? I tried searching it in the textbook I am working through and it only appears once without explanation.
So, for example, $(x(t),y(t)) = (2e^t,-e^t)$ and we want to determine if $Y(t)$ is a solution system, but I am unfamiliar with this term.
16:23
@TedShifrin Are you aware of any such proof
?
16:39
all you need is that an element $a$ in a field extension over $\mathbb{Q}$ is algebraic iff $\mathbb{Q}(a)$ is finite-dimensional
But how do you prove that ?@Thorgott
any refs?
Oh...wait...I wasn't given the entire problem. We're supposed to consider the system $dx/dt = 2x + 2y$ and $dy/dt = x + 3y$.
So, it's just a simple plug and chug.
16:52
Let $\mu_1 >= \mu_2$ be two natural numbers. Let $S = \mathbb{C}^4$ and let $A = \mathfrak{sp}(4,\mathbb{C})$ the complex symplectic Lie algebra of $Sp(4, \mathbb{C})$. Let $V$ be an irreducible representation of $A$ of highest weight (1,1). I need to prove $S^{\otimes(\mu_1 - \mu_2)}\otimes V^{\otimes \mu_2}$ has a irreducible sub-representation of weight (mu_1, mu_2).
I found a vector of weight $(\mu_1, \mu_2)$ but am not sure how to proceed. I know every representation of $\mathfrak{sp}_4(\mathbb{C})$ is totally reducible.
@Koro you would need to give me some clue what you are doing before i could comment...
Fatou's lemma remains valid if the hypothesis that $f_n$ is non negative measurable is replaced by the hypothesis that $f_n$ is measurable and $f_n \ge -g$ where g is non negative integrable. This is exercise no. 2.19 in Folland's book. I know how to do this.
Revised exercise: Suppose that $f_n$'s are integrable, g is integrable and $f_n\ge g$. Then Fatou's lemma's conclusion holds for $f_n$, i.e. $\int \liminf f_n\le \liminf \int f_n$. This is what I tried to solve in the images shared above.
@copper.hat
17:09
@Shinrin-Yoku if $a$ satisfies a polynomial of degree $n$, then $1,a,\dotsc,a^{n-1}$ generate $\mathbb{Q}(a)$ by a little bit of identity juggling. if $\mathbb{Q}(a)$ is finite-dimensional, the set $1,a,\dotsc,a^n$ will be linearly dependent for $n\gg0$, which yields a polynomial relation for $a$.
Step 1: Define $F_n:= f_n - g^+$ so that $F_n+g^-\ge 0$.
Step 2: Define $E=$ the set where lininf $F_n$ is non negative.
Step 3: Apply Fatou's lemma on non negative measurable function $F_n+g^-$. LHS thus obtained is $\int \liminf (F_n+g^-)=\int (\liminf (F_n) +g^-)=\color{blue}{\int_E (\liminf (F_n) +g^-)}+\color{green}{\int_{E^c} (\liminf (F_n) +g^-)}$, where the second equality holds because of non negativity and measurability of $F_n+g^-$
Step 4: Showing that the highlighted integrals can be written as sum of integrals.
This summary looks like often seen in Robjohn's answers :-).
what could $g^-, g^+$ possibly be if $g \ge 0$?
copper: in the revised exercise g is not given to be non negative.
this looks like another 'got the answer, delete & by bye': math.stackexchange.com/questions/4549334/…
ah, i see.
if $g\ge 0$ then $g^-=0, g^+=g$
Definition: $g^+= \max (g,0), g^-=\max(-g,0)$
17:15
why not just replace $g$ by $-|g|$?
and fall back to the first answer?
Where?
@copper.hat This epidemic is getting worser.
you are given that $f_n \ge -g$ hence $f_n \ge -|g|$.
ohh, that's very nice. :-)
some forms of lazy are ok :-)
@TedShifrin i'm not sure how those deletions should be handled.
17:21
I've been complaining in here to robjohn and Xander. They repost and scold the OP.
maybe i should do same.
Got it @Thorgott
but nonetheless, the solution that I shared is not wrong. Right ?
how to show Q(a+b) is finite dim a,b algebraic ?
@robjohn @XanderHenderson sorry to bug you as mods, another ask, get answer, delete & run : math.stackexchange.com/questions/4549334/…
17:22
@Thorgott
@Shinrin-Yoku $\Bbb Q(a+b)\subset \Bbb Q(a,b)$
@Koro i don't really have the energy to drag through the whole proof, it is not clear why you got the -1 that was highlighted, but in fairness you pulled it out of thin air and you could have just written $f_n-g$ instead of $F_n+g^-$ which is the same thing. Don't make your reader do extra work.
what Ted said
An what would be the basis @TedShifrin? (I got it’s finite dimensional but would $\{a^jb^k: O\leq j < m, 0\leq k< n\}$. Be the basis?) where m,n are the smallest naturals for which rational polynomials of degree m,n exist such that the polynomials satisfies $a,b$?
@Thorgott?
no, but that is a generating set
but there can be relations between $a$ and $b$ that make this set no linearly independent
17:31
Why not?
for the most degenerate case, consider $a=b$
@copper.hat Ohh that was because there was a problem in separating the integrals that way and $F_n +g^-$ seemed to simplify that.
you have to put yourself in your reader's mind, not your own.
makes Sense @Thorgott thanks a lot.
@TedShifrin also.
I guess the only Ada antge of the dets proof is that it actually gives us a polynomial @Thorgott
*advantage
yeah, I have therefore written all the steps that I used in the solution.
17:41
@Koro well, it reads more like a magic trick than a set of directions
yes
the determinant is a very natural perspective
because the minimal polynomial of an algebraic element is the characteristic polynomial of left multiplication with that element on the field extension it generates
@copper.hat In my opinion, the answerer probably shouldn't have answered the question in the first place (it is not of very high quality, and is likely a duplicate), but I have undeleted it. Thanks for the heads up.
@XanderHenderson thanks!
Hrm... I cannot find a good dupe target. That surprises me. Isn't that, like, a basic result in operator theory?
@Koro i don't follow the reasoning at the second highlighted point (-2).
18:01
Ugh... the railroad crossing in town is going to be closed for three days next weekend. Getting from the north side of the crossing to the south side is going to require a 30 mile detour.
And the best part is that I am supposed to report to the county complex for jury duty while this is going on.
So what is usually a five minute drive is going to take me almost an hour.
@copper.hat I don't have a bike in serviceable condition at the moment. And getting it into such condition would require a day-trip to Flagstaff (90 miles each way).
copper: $F_n=f_n-g^+$ has been substituted for $F_n$ in the inequality $\int \liminf F_n\le \liminf\int F_n$. Then, $\int F_n=\int f_n-g^+=\int f_n -\int g^+$ because $f_n$'s and $g$ are given to be integrable.
18:04
Also, I don't really want to show up to jury duty all sweaty and gross.
(It is about 5 miles each way, with significant hills on both sides). Honestly, it would probably take me 45 minutes to bike it (given how out of shape I am), so I wouldn't even be saving any real time. :/
18:27
@XanderHenderson Do you prefer a custom flag or a chat message (either in here or Cured) for suspected "homework cheat" deletions? On SO & Physics.SE the mods prefer flags for stuff like that.
It's a common problem on several sites, but it seems to be most common here on Math.SE.
18:41
@PM2Ring A flag makes it easier to keep track.
And it's a Good Idea to keep track, if you want to catch repeat offenders.
@PM2Ring That was my implication, yes. :D
0
Q: Infinite system of partial differential equations

geocalc33If $D=\{d_n\}$ is the space of coordinate systems on $\Bbb R^n$ then consider the equation: $$ \Delta_{d_1}f=\Delta_{d_2}f= \cdot\cdot\cdot$$ Where we have homogenized the coordinate variables (i.e. used the same variables for each coordinate system) and $\Delta_{d_1}$ represents the Laplacian op...

Xander that is an example of what i was talking about with an infinite system of equations
18:55
I have no idea what you mean by "the space of coordinate systems", but it seems to me that the only possible solution is the trivial one.
Me too
@XanderHenderson yeah I think "the space of coordinate systems" might need to be defined better
It needs to be defined at all.
You haven't defined it in your question.
I thought it didn't need definition
Everything needs a definition.
I already said that I have no idea what the space of coordinate systems is.
You haven't attempted to clarify...
do you know the cartesian coordinate system?
19:00
I know what a coordinate system is.
I do not know what the "space of coordinate systems" is.
Though it would be useful if you were to define a "coordinate system" so that you can reasonably define what a "space of coordinate systems" is.
the collection of all coordinate systems you can put on the manifold i suppose
With what structure?
@Koro: It seems that you are still working on some version of Fatou's lemma where $f_n+g\geq0$ ($g\in L_1$ but not necessarily nonnegative). At issue it seems that you are having troubles with cancelling $\int g$ from the inequality $\int(\liminf_n f+g)$. Set $f:=\liminf_nf_n$. Since $f+g\geq0$ then $I=\int (f+g)$ exists as an element in $[0,\infty]$. If the integral $I<\infty$, then $f$ is integrable and that's that.
You call it a space, so what structure does it have which makes it a space?
@Koro: Otherwise $I=\infty$. Let $E_+=\{f\geq0\}$ and decompose $f+g=(f_++g)\mathbb{1}_E-(f_--g)\mathbb{1}_{E^c}$. It is easy to check that $f_-\in L_1$. Thus the only option is $\int f_++g\mathbb{1}_E =\infty$ This implies that $\int f=\int f_+=\infty$ and done, the cancelation is valid also in this case.
19:03
If I have random variables X_1, ... , X_n all sampled independently and uniformly from [0,1] and let X = min_i X_i, am I right that E(1/X) doesn't exist?
@XanderHenderson I think diffeomorphisms from $\Bbb R^n$ to $\Bbb R^n$ would qualify
@OliverDíaz I understood how to solve that. The problem was that I shared my solution with one of my teachers and he said that there were some errors in it.
So that's what I was discussing.
6 hours ago, by Koro
user image
The teacher didn't pinpoint the errors, just said that there were some.
I'm convinced that there are no conceptual errors in it.
@Koro It suffices to decompose $f:=\liminf_nf_n$ into $f_+$ and $f_-$as $g$ cause no problem.
I did give an example of two coordinate systems in the post, plus a corresponding solvable linear pde. So for those specific coordinate systems in $\Bbb R^2$ there's a general solution @XanderHenderson
To address that, E was defined.
19:18
@Koro I see that there is a little error in the yellow part. The integral on positive measurable functions is subadditve in general. First you need to justify that $\int f_-+g\mathbb{E^c}$ causes no problem, that is $f_-\in L_1$, and thus that $\int f_+ - g=\infty. I wrote the details here
@geocalc33 You still haven't answered my question: what is the space of coordinate systems (or, if you equate this with diffeomorphisms, the space of diffeomorphisms). What structure does this space have? Or are you simply referring to the set of these objects?
If so, why do you keep calling it a space?
In any event, just consider the diffeomorphisms from $\mathbb{R}^n$ to $\mathbb{R}^n$. There are uncountably many. What gives you any reason at all to believe that you will get a non-trivial solution?
Also, why? Why does anyone care if there is any function which is invariant with respect to the Laplace operator expressed in every possible coordinate system?
What is the goal?
@OliverDíaz isn't it finitely additive in general? Theorem 2.15 in Folland says additive.
(finitely)
@XanderHenderson I don't have an immediate goal in mind
@geocalc33 Then what motivates the question? Why are you asking it?
I was planning on asking first, finding motivation later
19:37
That is, quite literally, not how it works. The motivation behind a question is the reason that a person might ask the question. Why are you asking this question? Why should anyone care what the answer is? What motivated you to post it to Math SE?
@XanderHenderson 30 mile detour to go 6 feet? That's insane.
@TedShifrin Welcome to the rural west.
And gas is only $6+ a gallon.
@TedShifrin About \$4.15 here.
At least, that is what I paid on Friday.
Wow, over $7 some places in SD.
@geocalc The Laplacian actually depends not on coordinate systems (although I know you're thinking of sums of second derivatives) but on a metric. Just look on $\Bbb R^2$ with different metrics (as many as you like) and think about the different equations you get.
19:47
@Koro Notice that in your solution you wrote $\liminf(f_n+g_-)$ in the second integral in the yellow. it should have been $f_n-g_-$ which is not necessarily $\geq0$. Again, $g$ is not the problem, it is $f:=\liminf_nf_n$. That is why I suggest you work with $f_+-g - f_- -g$ . Think about it for a little while. You can also take a look at the edit the my solution to your posting from yesterday.
20:17
I am solving the following exercise and I managed to to 1-3. Could you give me a hint for exercise 4? I can't do it somehow.
So I first want to compute the density but I don't get how to do it because I wanted to compute the distribution function and derive it but I don't think that I have enough information to compute the distribution function
The distribution function is give by $\sum_n P[X = n, Y \le t]$ and this is straightforward to compute. The form of the integral (do not integrate) shows the distribution function. Or you can differentiate if that is what floats your boat.
@copper.hat sorry so you mean $P(Y\leq x)=\sum_n P(X=n,Y\leq x)$?
Sorry I had a tipo
i presume it is the marginal distribution of $Y$ that is to be computed
yes to the correction
20:30
Ah okay, and then I don't need to compute the integral explicit but I need to derive all at the end?
do what you think you need to do, that was a hint.
okay thanks!
however you compute the density, usually differentiation.
Sorry for the stupid question but I thought that the first step $P(Y\leq x)=\sum_n P(X=n, Y\leq x)$ makes sense but thinking about it I t don't make sense. Could you maybe explain this to me why the two are equal?
 
1 hour later…
21:43
Minakshisundaram–Pleijel zeta function - is there another name for this? maybe 'geometric zeta function?'
centuries of ancestors of Minakshisundaram and Pleijel worked very hard at giving them those names, and you will use them.
@leslie Can you explain to me why this isn't nonsense?
i don't get it either.
He wants a real analytic function on $(-1,1)$ that is differentiable at $x=1$ and has all derivatives = 0 there.
So it seems like $e^{-1/(x-1)^2}$ is the only sort of thing to do, but he rejects it.
@robjohn Any thoughts?
22:01
yeah, i don't understand the basis for rejecting. there must be more about the goal that isn't stated (or maybe wasn't thought about). i do respect bountying an unanswered post from four years ago.
Wants an explicit formula for the coefficients.
It's far from obvious to me that the power series we get for that function will even be convergent/continuous/differentiable at $x=1$.
22:55
Good news. I got the highest score in an inter-college mathematics Olympiad
The biggest difference maker was the euclidean geometry worth 7 points, which I specialize in. That question gave me a huge lead
ted: this is my reaction. seems not unlikely that it would be at the edge of an interval of convergence, where generally who knows what will happen without some deeper thoughts.
23:15
$\not$
goku: congratulations.
would two expressions of the form Alog(B)=4Alog(C) mean B = 4*C?
I don't know how logs work admittedly
23:31
Well, isn’t it time you learned?

« first day (4450 days earlier)      last day (867 days later) »