« first day (1797 days earlier)      last day (3430 days later) » 

00:01
Seems simple enough, although part of the end of the method is a bit fuzzy.
user54412
"seems simple enough" -- I've never heard disturbing functions described as such
Well, relative to some of the other options out there.
Just like 1 million kilometers can be considered small, in some cases.
user54412
definitely the most symbolic prestidigitation I've had in an astro course -- which is why we left it to a Russian to teach
@HDE226868 your mom didn't think my 100 kilometers were small :^)
. . . That's a new one.
@ChrisWhite ?
user54412
00:09
So why hasn't anyone answered with making the star larger? Equilibrium planet temperature scales as $T_* R_*^{1/2} d^{-1/2}$, so bigger, hotter stars will have a larger absolute range of "habitable" $d$.
user54412
I suppose absolute differences in semi-major axis correspond to smaller energy differences further out, so there's more chance for planet-planet interaction. But the increased mass of the star could partly offset this.
@ChrisWhite That's the best question I've heard asked all day. You could argue that there's a narrow range of masses in which stability will increase while the timespan for complex life to evolve is above some minimum amount, but I doubt anyone's taken that into account.
People are fans of the G-star-ish system.
Even a small change in mass could help, I suppose.
user54412
Also, I'm pretty sure Niflheim isn't in the habitable zone.
00:14
@DanielSank I know about that
user54412
Ah someone already pointed that out.
I don't recall dong any coupled spring problem on this chat
@Secret listing the math you know
O, that
yeah, it helps you guy to explain things in case I have a question (because I tend to have tons of questions when I start self studying)
by letting you know my full academic background
00:29
()
I tend to use the chunk idea to help me grasp the concept and build illustrations. The neutrino oscillation question qualitative description asked yesterday is an example.

However, because my physics intuition still sucks, I often need to ask others in the field to check whether my illustrations are accurate enough. This is the reason I asked @ACuriousMind and @dmckee about it

This is because (particularly in the physics community, where they seemed to be extremely allergic to any potential cranks), I don't want to be accused crank because I got a wrong illustration or interpretation
One can say asking someone to help me check my illustration/interpretations is a step to help me to build my intuition, besides doing more exercise in textbooks
So basically, when I am still learning, I am trying to deepen my understanding by converting the maths into the physics via pictures, illustrations and descriptions, until I am trained enough to be able to do the other way without much error, when it becomes necessary to start building new models to explain and predict my experiments
Besides textbooks, forums and experts in the field are the best place to check whether I have made a mistake in the interpretation

Cause you know, there are some guys that made interpretations that is considered nonsensical by the physics community. I want to avoid that for my case
Interpretation questions, are however rarely answered in any place other than face to face, which frustrates me
I remember there's one scenario in the past in a science forum where the frustration is so great that I resort to an extreme approach: To act like a crank by posting my attempt in answering the question as if it is a fact

Then the science community's panoria and attention of cranks and trolls was being exploited as it means they often cannot resist the urge to correct me, thus unknowingly answered my question
why do people often associate me with JD?
I have heard many times that acuriousmind and others said I am like JD
in the chat
and I have no idea what I am similar to him
Do I need to prove I am not by posting my academic statement in UNSW???
00:46
how would that prove anything
It proves I have a physics and chemistry major, and not a computer major
it proves I really know how to do maths (as demonstrated earlier in my attempt on answering my own electromagnetism question)
@Secret his bio could be a lie
maybe he's the construct
and you're the real deal
how can you prove this I wonder
I don't know what I can say to prove that I am not
If I said we have different IP, you will said JD's IP is a proxy
If I post my UNSW qualification, you will said JD's is a construct

Basically, everything that I can think of that can be used as a proof, will be dismissed as can be easily made up because this is the internet
so you admit it
I honestly cannot think of anything that is not fabricable by the internet
I said not, but you would not believe
obe
obe
00:52
@0celo7 This all over...
what all over
obe
obe
Icosahedron.
oh yeah
what happened to him
obe
obe
He died.
he was this smart Canadian kid
::shrug::
obe
obe
00:58
lol
who are you, anyway
you kinda just showed up
@Secret I'm keeping my eye on you...
obe
obe
I'm really JD.
xD
blocked
4
Q: Application of Cauchy-Schwarz with Sobolev norms

0celo7I'm working through the problems in the initial value formulation chapter in Wald's General Relativity. A short summary of the problem. I have to show that $$\sup_{x\in A}|f(x)|\le C||f||_{A,k}$$ where $A\subset \mathbb{R}^n$ satisfies the uniform interior cone condition, $C$ is a constant, $k>...

anyone welcome to solve it
obe
obe
really old.
I still don't know how to solve it
obe
obe
01:07
fml.
what
@bolbteppa buddy
pls
@Secret if you solve it
JD could not
honestly, that problem ruined Wald for me a bit
obe
obe
runs away from wald
lol you thought Wald was all fun and games?
Wald is srs bsns
obe
obe
rly
read it at your own peril
note that I won't be able to help with everything
obe
obe
01:17
dude you're supposed to be my guardian angel in physics.
er what
obe
obe
lol
like I said
good luck with that first problem on the second set
I will not help
too difficult tedious
ask the dude who finished how he did it
if he said it was not tedious, he's a lair and a cheat
obe
obe
he already took qft before.
supposedly.
well ACM looked at it
and I looked at it
FWIW, we had the same solution
and it's pretty rough
obe
obe
01:21
then I forfeit.
bye.
@ChrisWhite Now you'll flip out:
The author of that website got up to 60 in a single system actually. — PyRulez 32 mins ago
oh my gooooood
can you put that in non-astro nerd language?
The writer of a blog supposedly found a way to make 60 Earth-like planets orbit a star and not collide.
Again, I'm not sure I buy it.
I do
just find the stable solution :^)
obe
obe
oh I'm an expert in stable orbits.
01:34
is there some theorem that says this is not possible?
unless there's a theorem, there's no reason to think it's not possible, right?
@HDE226868
01:47
@Slereah you around
@0celo7 what happens if $A$ contains only one point?
erm
what
ah
pick $C$ to be very large
or
wait
wtf is the Sobolev norm again
maybe just assume $A$ is not a point
@0celo7 Can $f = 0$?
not sure, why?
Because if $f = 0$ the inequality becomes an equality trivially
01:58
ok
but it has to work for every f
Thus the inequality is just accounting for the fact that $f$ can be zero as well, in which case the equality is trivially satisfied, after that maybe it becomes a strict inequality
well I don't know how to prove inquality
The point is, if $f = 0$ you don't have to do anything, it's obviously true trivially, so what you wrote up in your answer shows the inequality to hold.
I'm not convinced that I proved inequalty
Do you think I did?
But does the thing you wrote in the last line not show up as merely one of the many terms in the Sobolev norm? Am I reading it wrong or?
02:02
no
because $\psi$ is in there
and depending on $\psi$, the derivative of $\psi f$ could look a lot different than that of $f$
user54412
@HDE226868 Forget about point mass stability for a moment. I strongly suspect that nonideal mass distributions become more important in these ridiculous systems.
>ridiculous
you're not gonna get a 60 planet system with that attitude
what's the largest known solar system
(The stakes are high... given I have not learnt sobolev norm before. I hope the cauchy swartz stuff learnt in my linear algebra, the euclidian norm and the standard inner product might be useful, let's see...)
btw there's a theorem that makes this true
but I want to do it without the theorem
because that theorem is stuck 300 pages in the average PDE book
without which theorem?
02:11
Sobolev embedding
that's how Hawking-Ellis proves it
but they just cite some PDE analysis book
user54412
In fact, here's my (poorly worded) conjecture: Given any algorithm for procedurally adding planets to a system that converges to a point in parameter space that is provably stable, even if a neighborhood of this point is also stable, its measure goes to 0 for any fixed deviation from ideal point masses given to the planets. That is, for real systems with higher multipole moments, you cannot make arbitrarily complicated systems even if guided by a perfect setup with point masses.
I ain't about that life
I am pretty sure if any answer I wrote will not include this, because I never heard about that therom before
@ChrisWhite proof?
just solve the n-body equations :)
not a big deal...
@bolbteppa do you think I have proved inequality?
@0celo7 I thought you were happy with your intuition and just wanted to know about the equality, but I'm not sure about the intuition... About the inequality, yeah so $\psi$ being $\mathcal{C}^{\infty}$ between $r = H/3$ & $R = 2H/3$ means you have to justify the last step right?
02:28
no
my intuition is shit
don't trust that mofo
@bolbteppa yup
It might come down to that but you just need to show it, probably will tbh
well how?
I don't see any way to show that
if $\psi$ is really crazy...
Oh, since $\psi$ is $\mathcal{C}^{\infty}$ the function & all the derivatives are bounded so you can just bound $\psi$ by some constant on that section of the domain right?
yes
YES
write up an answer
I was getting stuck thinking of it being $\sqrt{x}$ and it not working at $x = 0$ but then I realized that's not possible!
(shifted appropriately)
Lets just be sure about that
02:38
that function is not $C^\infty$
Yeah
Thanks for this chapter, really liking the motivation for the sobolev norm
hmm?
are you typing up an answer
I like your idea but I need to see it fully fleshed out
Can I just copy-paste that?
no, you need to prove that this boundedness works
I think I know that it does
but it would be nice to maybe have an example?
It's just $||\partial_r^k(\psi(r) f)||_{L^2} \leq ||\partial_r^k(\psi(r)_{max} f)||_{L^2} \leq C||\partial_r^k f||_{L^2} \leq C||f||_{Q,k}$ (right? Just written in terms of an integral as you did in your post, I have no idea how to define a function on that crazy cone to match the bounds haha)
02:45
mmmm
mmmmmmmmmmmmmmm
maybe
02:56
@0celo7 Well, no, but there are a lot of potential issues the more planets you add.
@0celo7 Pending confirmation of exoplanets in certain other systems, it's the Solar System. After that, it's HD 10180 (7, possible 2 more, which would give it 9 over the Solar System's 8) and then Kepler-90.
@0celo7 I'm okay for writing it up as long as the answer makes sense, any problems let me know
I made the gradient, laplacian, divergence and curl all look the same in arbitrary coordinates yet still link to easily plugging in numbers, Zee(wa)hoo!
03:11
I give up, I have tried to self learn, but I still don't understand Sobolev spaces nor its norm (especially the multidimensional case)
@Secret that shizz is crazy
Sobolev spaces are a way of not only measuring a function but also it's derivatives on an interval basically
That's what I have read but when the whole thing is expressed in a precise definition, I lost track on what's going on
I might try again to understand later with this source instead of wikipedia
https://www.math.psu.edu/bressan/PSPDF/sobolev-notes.pdf
wikipedia is really not helpful in understanding new topics
Okay, well the vague jist is as follows, when you look at a vector like $\vec{v} = (1,2,3)$ you might think of an arrow, but you might also think of that as representing the polynomial $f(x) = 3x^2 + 2x + 1$. If I keep going I can represent a polynomial as an infinite dimensional vector, and clearly you can see Taylor/Fourier expansions of arbitrary functions implying all sorts of fun... Well, the general theory for thinking of functions as vectors is that of Hilbert spaces.
The norm of a vector measures it's length, similarly the norm of a function measures the area under the curve or something like that (depending on how you measure)... Anyway, I just turned functions into vectors, what happens if I try to turn linear (ordinary/partial) differential equations into some vector analogue? I'll get sobolev spaces.
In other words they are Hilbert spaces where you allow derivatives into the mix
Ok, this sounds quite intuitive, let's see if I can churn out something from $$||\partial^k_r(\psi f)||_{L^2}\le C_4 ||f||_{Q,k}$$ using your description and check the result with you
This math.harvard.edu/~canzani/math253/Lecture11.pdf has a nice motivation and an actually really easy proof of Sobolev embedding which I can't believe is that easy haha
for the norm
Short video on Sobolev spaces youtube.com/watch?v=_pCbaAX7UM0 and an example youtube.com/watch?v=aU3h3gsdA-c
03:31
$\psi(r)=1$ for $r<H/3$ and $\psi(r)=0$ for $r >2H/3$.
What's the value of $\psi(r)$ for $\frac{H}{3}\leq r\leq \frac{2H}{3}$?
All you know is that $\psi \in \mathcal{C}^{\infty}$
I really could have used Sobolev embedding 5 months ago :\
Do I expand this correctly?
$$\phi(r)\in C^{\infty}$$
$$k > \frac{n}{2}$$
$$||\partial^k_r(\psi f)||_{L^2}\le C_4 ||f||_{Q,k}$$
$$\sqrt{\int_{\mathbb{R^n}}(\partial_r^k(\psi f))^2d\mu}\leq C_4 \left(\sum_{|\alpha|\leq |Q|}\int_{\mathbb{R}^n} |D^{\alpha}f|^k d\mu\right)^{\frac{1}{k}}$$
03:47
I think so except the $1/k$ should be $1/2$?
If 1/k is 1/2, then shouldn't the sobolev norm shown on the right with be $||f||_{Q,2}$ instead?
Mixing up notation a bit, the $k$ in your last line should be $2$ and the $|Q|$ should be $k$ to match up to what we've done
Good luck
Ok I see, but where did the Q from the $||f||_{Q,2}$ gone in the expansion?
$$\phi(r)\in C^{\infty}$$
$$k > \frac{n}{2}$$
$$||\partial^k_r(\psi f)||_{L^2}\le C_4 ||f||_{Q,2}$$
$$\sqrt{\int_{\mathbb{R^n}}(\partial_r^k(\psi f))^2d\mu}\leq C_4\sqrt{\sum_{|\alpha|\leq k}\int_{\mathbb{R}^n} |D^{\alpha}f|^2 d\mu}$$
04:48
I don't think the following makes sense...
Given
$$D^{\alpha}\Lambda_f=\int f D^{\alpha} \phi d\mu=(-1)^{|\alpha|}\int g\phi d\mu=\Lambda_g$$
$$\phi(r)\in C^{\infty}$$
$$k > \frac{n}{2}$$
$$||\partial^k_r(\psi f)||_{L^2}\le C_4 ||f||_{Q,2}$$
$$\sqrt{\int_{\mathbb{R^n}}(\partial_r^k(\psi f))^2d\mu}\leq C_4 \sqrt{\sum_{|\alpha|\leq k}\int_{\mathbb{R}^n} |D^{\alpha}f|^2 d\mu}$$
If we assume equality
$$\sqrt{\int_{\mathbb{R^n}}(\partial_r^k(\psi f))^2d\mu}= C_4 \sqrt{\sum_{|\alpha|\leq k}\int_{\mathbb{R}^n} |D^{\alpha}f|^2 d\mu}$$
$$\int_{\mathbb{R^n}}(\partial_r^k(\psi f))^2d\mu= (C_4)^2 \sum_{|\alpha|\leq k}\int_{\mathbb{R}^n} |D^{\al
Slight correction...
Integrate wrt r k times
$$\psi f=\int\cdots \int\sqrt{k}C_4|D^{\alpha}f| d^kr+C(\Omega)\frac{r^{k-1}}{(k-1)!}$$
Now $\sqrt{k}C_4 > 0$
$$\psi f=\int\cdots \int|\sqrt{k}C_4||D^{\alpha}f| d^kr+C(\Omega)\frac{r^{k-1}}{(k-1)!}$$
$$\psi f=\int\cdots \int|\sqrt{k}C_4D^{\alpha}f| d^kr+C(\Omega)\frac{r^{k-1}}{(k-1)!}$$
 
3 hours later…
07:38
@0celo7 I am here
 
1 hour later…
08:49
"Massless particles aren't characterized by spin because square of Pauli-Lubanski operator (which is Casimir operator of Poincare group) for them is equal to zero. They are characterized by helicity."
This sounds wrong
Am I insane
09:03
@0celo7 Aaaah....good ol' Sobolev embedding theorem.
@Secret What do you want to prove? I could not follow...
you have a function $\psi$ and a function $f$. What is the regularity of $\psi$; and what is the regularity of $f$?
It is not necessarily Sobolev that you want to use there (even if it may be ok).
My hint: $$\bigl\lVert \lvert\nabla\rvert^r \psi f\bigr\rVert_2\simeq\bigl\lVert \lvert k\rvert^r \hat{\psi}*\hat{f}\bigr\rVert_2$$
apart from some $(2\pi)^{d/2}$ factor
09:37
Hi @yuggib
et al.
:-)
hey
et al. is the most prolific author in the field of science really
@skillpatrol Hi there
;-)
09:56
there's a lack of GR questions lately
I am sad
 
2 hours later…
11:29
Anyone who says "good ol Sobolev" is insane
@0celo7 I said good ol' Sobolev embedding theorem
the poor Sobolev is dead
@yuggib do you mind writing an answer for my math.SE question
@0celo7 If I could...can you show me your math.SE question? I got lost in the discussion above
On mobile, you can look at my profile
4
Q: Application of Cauchy-Schwarz with Sobolev norms

0celo7I'm working through the problems in the initial value formulation chapter in Wald's General Relativity. A short summary of the problem. I have to show that $$\sup_{x\in A}|f(x)|\le C||f||_{A,k}$$ where $A\subset \mathbb{R}^n$ satisfies the uniform interior cone condition, $C$ is a constant, $k>...

Or I can post it
11:39
ok, so a clarification: how do you define the Sobolev norm?
what $A$ and $k$ stand for?
is that the Sobolev norm you think: $$\lVert f\rVert_{A,k}^2=\int_{A} (1+\lvert \xi\rvert^2)^k \lvert\hat{f}(\xi)\rvert^2d\xi $$?
I'll answer that when I can get out my laptop
11:54
Ok, let's look
fucking butterfingers
$$||\phi||^2_{S_1,k}=\int_{S_1}\left\{|\phi|^2+\cdots+\sum_i|\partial^{k}_i\phi|‌​^2\right\} \,\mathrm{d}x$$
ok, so that's the same thing that I wrote above
@yuggib
really
doesn't look like it
yep
how much do you know about Fourier transforms?
ah
that explains it
(I know analysis is not your strong suit)
11:59
I'm sure if I work that out and use that one theorem (Parseval?)
at least I think that's what's going on
the Sobolev spaces with base space $L^2$ are best understood in Fourier transform
@yuggib I don't even have a strong suit
what baffles me is that I was so much better at math than everyone in my high school
because Fourier transform maps $L^2$ into itself (Parseval)
and derivatives are multiplication by the variable in Fourier transform
knew it!
ok, well if you can use that, swell
actually, it should be Plancherel and not Parseval...
12:01
I just want to be done with that exercise after 9 months
@0celo7 you are a small fish in a big pond now pal :-)
@yuggib same thing?
@0celo7 but to solve the exercise, you are basically proving Sobolev embedding in Hölder spaces...and that is not straightforward
my physicist's proof is just $\langle g|f\rangle=\langle g|f\rangle$
@yuggib see, this exercise is in a book that does not talk about any of that
@0celo7 yes probably
12:03
maybe the author had a simple but flawed proof in mind?
@0celo7 I know, but that's the thing
maybe he thought about a simple justification
but I do not see it being a physics exercise
the proof is necessary to do the Cauchy problem in GR
Hawking-Ellis, the canonical reference on the Cauchy problem, just cites the embedding theorem
Wald never calls it that, he just leaves it as an exercise
and cites some unpublished dissertation for the proof he has in mind
maybe I should look up the dude on linkedin
@0celo7 I think you should learn something about Sobolev spaces and embedding, then everything should be clearer
well sure
but that's cheating
if else, you simply believe the theorem to be true
12:13
I can't wait until I learn some analysis
It's not possible to self study that
yes, probably you just need to wait
Build on the part of math you know best for now, like algebra.
I don't know shit about algebra
Get learning.
I was thinking about getting Arnold so I can practice geometry
But I have too much to read as it is
12:18
What part of math were you best at?
I got perfect grades in high school
in math and science
I know some geometry
But that's about it, really
Try your best, that's all you can do :-)
However, let's try to get $\lVert \partial_r^k (\psi f)\rVert_2\leq C\lVert f\rVert_{Q,k}$ step by step
that is not difficult
first: use the chain rule of differentiation
you get inside the first norm, term of the like $\partial_r^{l}(\psi)\partial_r^{k-l}(f)$, for $0\leq l\leq k$
now separate each term from the others, using triangle inequality
you get $$\lVert \partial_r^k (\psi f)\rVert_2\leq \sum_{l=0}^k\lVert \partial_r^l (\psi) \partial_r^{k-l}(f)\rVert_2$$
now consider a single term
$\partial_r^l (\psi)$ is bounded on $Q$, for the function is $C^\infty$ and the domain is compact (it is a closed cone)
let's call $C_l(\psi)=\mathrm{max}_Q \partial_r^l\psi$; then we have by definition $\lVert \partial_r^l\psi\rVert_\infty=C_l(\psi) $
therefore, by Hölder inequality, we have $$\lVert \partial_r^k(\psi f)\rVert_2\leq \sum_{l=0}^k C_l(\psi)\lVert \partial_{r}^{k-l}f\rVert_2\leq C \lVert f\rVert_{Q,k}$$ by definition of the Sobolev norm
Now a possible issue is that the constant $C$ depends, through the $C_l(\psi)$, on $H$.
And it may explode when $H$ becomes very big. But probably the cone condition (that I do not know very well) should avoid such explosion of the constant
Observe that your $C_2$ depends on $H$ as well
Anyways, that is surely not the best way of proving the result @0celo7.
@Danu Have they updated this to reflect that the higgs boson has been discovered?
@yuggib What I am trying is what you and 0celo7 just went through, except as revealed above, I got the sobolev norm completely wrong because I don't know about it and self study for it does not help much

For the reason on why i am trying to solve this problem, see the entry on the star board
12:59
ok
specifically "how can you prove this I wonder"
@Secret Play and find out ;)
@danu hopefully that does not take too long (of the idle games I have seen out there, this is the only one that does not suffer form diminishing returns as you go to higher units)

Unless they have updated the game mechanics, the increase is a constant 10 times form level to level
@Secret I've been playing for about an hour and I'm getting decently close, I think
@Secret They have updated.
Oh I got the Higgs just now
The next step I believe shoudl be the pentaquark
or the final neutrino mixing angle
Let's see...
O btw, if you are interested, 0celo7 have some very disturbing conspiracy about me.
We don't know of any ways to prove it without being accused of fabriacation however
13:09
@Secret ?
check star board:
"how can you prove this I wonder"
You've provided no evidence to the contrary
Our latest question (above sobolev stuff) was supposed to be a prove that I can do maths, while at the same time solving the problem in Wald.

But this question contains knoewledge not listed in my academic statement (as I have done 2 months ago), thus obviously I am unable to solve it because I don't even understand the problem despite I tried
We currently don't have another question that can be used as a proof, but perhaps they will soon came up one

And we have argued that the electromagnetism stuff I post yesterday is not strong enough as a prove

So yeah, this is a pretty annoying conspiracy to be resolved
Proving you can do math means nothing
JD could be the non math construction
JD can't do math though
13:26
To you, anything is a construction, I don't know if anything can convince you that I am not JD

since there is no pathway that can be mathematically rule out that it is not a construct, because it is possible for a conspirer to came up with a completely separate identity, to the point of making different IPs, fake bios, acting dumb, etc.

I don't see any advantage of me wasting so much resources just to fabricate a separate identity, though
sock puppets always have common talking styles, but this arguement is not strong, because a conspirer can just made the mask moreperfect so that the styles don't match

But once again, what I can gain from this process?
A side note, actually,@Slereah, have I demonstrated in the past starting from when I first joined this chat, that I can do maths?
@Slereah exactly
so someone who knows math could have constructed JD
in fact
there are some people who I know are not JD
these are the people who were in chat when he was banned
so @ACuriousMind, @Slereah and myself
@yuggib ok
so everyone else is suspect
Was JD banned?
chat banned
Woo
Forever?
13:40
oh no
30 minutes
that's happened to me too
Can't we make it forever
and also banning him from SE
no
you're a barbarian
you can't just ban people because you disagree with them
How about let's turn the question around:
List everything that makes YOU think I am JD?
that's not how this works
13 hours ago, by Secret
I have heard many times that acuriousmind and others said I am like JD
You guy do said that, especially whenever I post weird pictures
>0celo7: @ACuriousMind HE'S JD
>Acuriousmind: Ah, that's the secret.
>0celo7: although I think JD just printscreens
http://chat.stackexchange.com/transcript/message/24518527#24518527

« first day (1797 days earlier)      last day (3430 days later) »