« first day (3650 days earlier)      last day (1667 days later) » 

04:59
@ArtificialStupidity Hello! Too fat man no neck!
@robjohn Sir, how can I prove that these two inequalities contradict each other
$$ 0 \lt |1-l| \lt \epsilon \\ 0\lt |l| \lt \epsilon$$
Given that $\epsilon \gt 0$ is constant, and $l$ is any real number.
I don’t know why I’m unable to prove it, I think it should be easy.
05:26
@Knight they are only contradictory if $\epsilon\le\frac12$. Then the triangle inequality says $1\le|1-l|+|l|\lt2\epsilon\le1$.
@robjohn How you figured out that $\epsilon$ must be equal to or less than 1/2?
@Knight look at what I just wrote and try to answer your own question.
If $\epsilon=\frac34$, then $l=\frac12$ satisfies both inequalities (no contradiction).
@robjohn How is $2\epsilon $ is less than 1?
If $\epsilon\le\frac12$, then $2\epsilon\le1$.
@robjohn Yes sir. But how would I know from next time where I have to take $\epsilon \lt 1/2$?
05:37
I'm not sure I understand your question. Next time, if it's the same question, you know. If it is a different question, it may not be $\epsilon\le\frac12$.
Lol :)
I agree sir, I meant how you came up with the condition $\epsilon \lt 1/2$. If in future I get a different question, how should make my that clever choice?
Think of what the inequalities are saying: $|l|\lt\epsilon$ says that $l$ is closer to $0$ than $\epsilon$. $|1-l|\lt\epsilon$ says that $l$ is closer to $1$ than $\epsilon$. When can those be at odds?
What is the closest that something can be to both $0$ and $1$?
1/2.
Mid way.
So, have you answered your question?
If I want to prove that these two inequalities contradict each other $$ |x-l|\lt \epsilon \\ |l|\lt \epsilon$$
then, I think the good choice should be $\epsilon \lt x/2$, right?
05:42
How about these: $|x-l|\lt\epsilon$ and $|y-l|\lt\epsilon$
@Knight $|x|/2$
Then $\epsilon \lt \frac{|x-y|}{2}$
right?
@Knight yes
Sir, you have taught me something which I shall always be grateful to you for.
You’re really awesome !
@Knight because $|x-y|\le|x-l|+|y-l|\lt2\epsilon\le|x-y|$
draw a picture
05:46
@robjohn yes
@LeakyNun Yes, got that. You mean number line, ha?
Mark x and y and see the magic :)
YES! @Leaky is now giving my perennial "draw a picture" advice!!!
LOL
OK, so you stole it a while ago.
05:50
lol
06:04
@Knight where could $l$ be?
hi everyone. can someone tell me where can i get the AIEEE papers with solutions on fficial website ?
it sounds like you're screaming... AIEEE!!
2
@robjohn yes . I am looking for the past year paper with solution of AIEEE
@robjohn do you know the official link to these papers ?
I didn't see a quick answer on this page
@ronakjain I have never heard of this examination before.
@robjohn ok. No problem !
06:34
@Knight Hi. I am here. :-)
@ArtificialStupidity How do we know that you are really here? You look somewhat unoriginal.
@robjohn Sir, it’s been an hour and a new doubt has crept inside my mind.
@Knight flush it out
We wanted epsilon to be a constant but when we write $\epsilon \lt \frac{|x|}{2}$ isn’t $x$ a variable?
We are saying that $\epsilon$ needs to be less than a variable?
and variable can change?
Not if you declare it not to change. Sure the notation $x$ represents a variable, but a variable can contain a constant. However, it can hold one constant at one time and another constant at another time.
Sure, it is not like $\pi$, which is given a particular value most of the time, but it could be given another value at some point in a paper that doesn't care about the ratio of lengths in geometric shapes.
@WeavingBird1917 math has been stepped on by a giant.
If we consider this function $$f(x)= \begin{cases} x & if~x~is~rational \\ 0& otherwise \end{cases}$$ and try to prove that it’s limit is zero, when x is near to zero.
Lol, I am learning some graph stuff and decided to visualize MathSE tags by similarity.
We would do something like this $$ \text{consider an epsilon greater than zero} \\ 0\lt |f(x)| \lt \epsilon \\ 0\lt |x| \lt \epsilon ~~~(for ~rational ~ x)$$
Now, if I take $\epsilon \lt \frac{|x|}{2}$ then of course, my inequality will be (in @LeakyNun word) S N I P E D.
@robjohn What should prohibit me from doing that?
where did $\epsilon\lt\frac{|x|}2$ come from and why should it be relevant to the inequalities that you've mentioned?
The inequalities you've stated are not even close to the inequalities we were talking about before and $\epsilon\lt\frac{|x|}2$ is not related to them at all.
07:00
Yes, we were discussing $|1-l| \lt \epsilon$ and $|l| \lt \epsilon$.
@robjohn Maybe I need a certificate issued by an official Certificate Authorization agent to prove my identity. :-)
@ArtificialStupidity Hola
@ArtificialStupidity RealStupidity seems more original
never mind. It was a bad joke
@robjohn No problem. That is fine! :-)
07:12
@robjohn I kindly request you to please give me an explanatory proof that there is no limit of $f(x)$ at any point $a$. $f(x)$ is defined above (that rational and irrational one)
I request you to please do that.
@Knight $\lim\limits_{x\to0}f(x)=0$
so your request is impossible
Hi everyone. Can someone tell me how to type using mathjax. I can read using mathjax but not able to write
@ronakjain have you installed the ChatJax bookmark? that helps to read it
07:27
@robjohn I have made a bookmark as hrishab had explained
@ronakjain okay, so you see $\int\frac1x\,\mathrm{d}x$ rendered as an integral?
Then read this tutorial about using MathJax
@robjohn yes I can.
@robjohn but I did not use that link
@robjohn is had used this one
4
A: How to use MathJax on the mobile web chat?

Avyansh KatiyarWe'll be using robjohn's bookmarklet to make mathjax work on your mobile device. As a matter of fact he's pretty much written a step by step guide for doing so. Copy the text from the box below. Create a bookmark to this (or any) page. Replace the contents of the bookmark URL with ...

@ronakjain that is how to install ChatJax. That is also described in the link in the upper right sidebar on this page, where it says "$\LaTeX$ in chat"
The link I give above is for how to write MathJax
07:48
@robjohn I'm sorry, I missed that $a \neq 0$.
Prove that $$
f(x) = \begin{cases} x & if~x~is~rational \\ 0 & otherwise \end{cases}$$ have no limit $l$ for any point $a\neq 0$.
08:06
@robjohn will I need to memorize the commands if LaTeX. They seem quite difficult for memorizing to me
08:34
@Knight Since you won't give up on this, suppose it did. Then there would exist $\delta>0$ such that if $|x-a|<\delta$ then $|f(x) - \ell|<|a|/2$. Since the set of such points $x$ is an interval around $a$, there is a rational such $x$, say $x_0$, and an irrational one, say $x_1$, such that $x_1$ is further from $0$ than $a$, i.e. $|x_1|>|a|$. We get the inequalities $|f(x_0) - \ell | = |\ell | < |a|/2$ and $| f(x_1) - \ell | = |x_1 - \ell | < |a|/2$. Then
$$ |a|<|x_1| \le |\ell| + |x_1 - \ell | < |a|$$
Does a 'nice squiggly line' in $\mathbb{R}^2$ always have a bijection to some closed interval $[a,b]$ in $\mathbb{R}$? Intuitively I think so because a closed interval in $\mathbb{R}$ is a 'nice squiggly line' in $\mathbb{R^2}$. By 'nice squiggly line' I think I mean a continuous path.
@ronakjain as you use them, they will become easier to remember.
Suppose that $\lim\limits_{x\to a}f(x)=L$. Then the definition of a limit says
$$
\forall\epsilon\gt0,\exists\delta\gt0:\forall x,|x-a|\le\delta\implies|f(x)-L|\le\epsilon\tag1
$$
Choose $\epsilon=a/4$. For the $\delta\gt0$ guaranteed above, we have that
$$
|x-a|\le\delta\implies|f(x)-L|\le a/4\tag2
$$
we can pick $x_1\in\mathbb{Q}$ so that $|x_1-a|\le\min(\delta,|a|/4)$ and $x_2\not\in\mathbb{Q}$ so that $|x_2-a|\le\min(\delta,|a|/4)$. Note that since $|a|\le|x_1|+|x_1-a|$ and $|x_1-a|\le|a|/4$, we have
@CalvinKhor i mixed up rational and irrational here but thats not so important
@Threnody is $f:[0,1]\to \mathbb R^2, \ f(x) = 0$ a "path"? It's continuous
@CalvinKhor What definition for 'path' do we usually use?
I only have the definition for a 'continuous path'
you may want to avoid self-intersections, is my point @Threnody
08:43
@CalvinKhor What if we allow intersections?
Then two parameters in the domain map to the same point
violating injectivity?
Actually reading your question, if you only want a bijection of sets of points, the answer is yes, the moment your continuous path has more than one point (because the cardinality will match), but it might not be very nice
08:55
Thank you :)
sorry for reading something you didnt write the first time round lol @Threnody
 
1 hour later…
10:19
@Hippalectryon HOLA
Will someone please try this question I am going to Post.....
You have 5 simultaneous equations.
(cunningham's law is in full effect here) or, you can notice that $f(x) - (6-x)$ has at least 4 distinct roots: 1, 3, 4, 5 (I can't read what f(2) is)
presumably f(2) = 4
in which case $f(x) - (6-x)$ has roots 1, 2, 3, 4, 5, so $f(x) - (6-x) = A(x-1)(x-2)(x-3)(x-4)(x-5)$
and since $f$ is monic, we conclude $A = 1$
Monic?
monic = leading coefficient is 1
10:30
@LeakyNun what is Cunningham's law ?
@LeakyNun but what is its relation with my question ?
never mind
@LeakyNun will you please explain that one
tell me what you dont undersatnd
5 mins ago, by Leaky Nun
in which case $f(x) - (6-x)$ has roots 1, 2, 3, 4, 5, so $f(x) - (6-x) = A(x-1)(x-2)(x-3)(x-4)(x-5)$
5 mins ago, by Leaky Nun
and since $f$ is monic, we conclude $A = 1$
10:33
I don't have understood anything from starting of your answer
what is f(1) - (6-1)?
It's 4
what is f(1)?
It's 5
@LeakyNun what next ?
what's 6-1?
10:37
It's 5
so what's f(1) - (6-1)?
It's 4
what's 5-5?
I am given the value of these points in the function. So why are you asking such questions ?
It's 0
10:37
so what's f(1) - (6-1)?
It's 4
given that f(1)=5 and that 6-1=5 and that 5-5=0, what's f(1)-(6-1)?
@LeakyNun what do you want to tell ? Why are you asking silly questions ?
@LeakyNun please explain it rather than asking such questions ?
@EdwardEvans linguam latinam in schola didiscisti?
@ronakjain because you keep saying that it's 4 for some reason
10:41
@Leaky if I learned Latin in school?
Nah I didn't
@LeakyNun oh...I just took 6-1 as f(6-1)
did they offer it?
@ronakjain ok
no lol, I only had Spanish and French in school, I taught myself German
10:42
so can you see that f(x) - (6-x) has roots 1,2,3,4,5?
15 mins ago, by Leaky Nun
in which case $f(x) - (6-x)$ has roots 1, 2, 3, 4, 5, so $f(x) - (6-x) = A(x-1)(x-2)(x-3)(x-4)(x-5)$
15 mins ago, by Leaky Nun
and since $f$ is monic, we conclude $A = 1$
great, so that's the first sentence
so now do you understand the second sentence?
@Leaky I got some cool seminar talks to give this semester
so the third?
@EdwardEvans sopra che?
10:43
What is monic ?
14 mins ago, by Leaky Nun
monic = leading coefficient is 1
@Leaky one on elliptic curves over global fields (I'll be proving weak Mordell-Weil and introducing the Kummer pairing via Galois cohomology) and then one on local Langlands
but I'm not sure what talk I'll have for the Langlands seminra
seminar*
@ronakjain so can you find f(6) now?
10:45
great
But how did you come to the first point
pattern recognition
f(1,2,3,4,5) = 5,4,3,2,1
screams out "I'm linear" to me
@EdwardEvans great
11:17
@LeakyNun is there any particular method of integration which may be applied to all types of problem ?
 
4 hours later…
15:09
@CalvinKhor Cal, can you please tell me how did you make that businessman choice (lol) of taking $\epsilon \lt \frac{|a|}|{2}$?
@robjohn Sir, can you please explain me one more time why you took $\epsilon = |a|/4$? And can give me some advice of how to make such a clever choice? (I know I have already asked it, but please explain it one more time).
15:30
@LeakyNun hi :-)
15:43
@robjohn: You are good at writing a LaTeX renderer on web browsers. There is a challenge that might be interesting to you. The first person to port MathJaX to Blazor WASM will be hailed as the messiah.
3
16:19
@EdwardEvans just complete the Langlands program and talk about your results
16:30
@ronakjain yes but it's meant for computers
(Risch algorithm)
@Alessandro that is the goal of my talk
16:48
Howdy, @Edward, @Leaky, demonic @Alessandro
takes a nap
Hiya @Ted
Suppose u got 10 000 coins and you bet 5 coins on 50/50 odds. If you lose, you bet double than you bet beforehand, if you win you bet 5 coins again. Does this strategy work in theory ?
Are your chances of winning high if you bet enough times ?
@TedShifrin hey ted!
16:53
Hi Stan
I've found the center of mass of the trapezoidal prism
Yippee.
So my next step I think is identifying the bounds of integration
with the trapezoidal prism centered on the origin
is that right?
I'm confused.
Oh I'm finding the moment of inertia tensor
for the trapezoidal prism
16:55
And you found the center of mass without doing any integrals.
I guess you could similarly do the MI (you're not doing the whole tensor, just one entry of it) in pieces.
I found the geometric centroid of a trapezoid
So my first step was
which gave me $\bar{x},\bar{y}$
and then the $z$ is just half the length
Oh, so you didn't do the weighted average thing I suggested.
so $\frac{\ell}{\2}$
no i wasn't super creative
i found the formula and asked someone on math SE to prove it
and someone did
and i was satisfied with the proof
Still, for the moment of inertia, you can if you want break the triple integral into easier-to-solve pieces.
16:58
@TedShifrin so i was planning to break up the integral into 3 parts, the left triangle, the box, and the right triangle
There's demonic!
and i can write y in terms of x i think
The left and right triangles give the same contribution by symmetry.
At the moment, I'm trying to do things in the simplest math way possible because I seem to understand the physics poorly. I tried something simple like taking the inertia tensor for a cube. Then cutting the cube in half and trying to compute the overall tensor for the cube from each half and got it wrong.

I'm gonna go write up my approach and then share it. be back in a bit
You keep saying tensor. Are you computing all 9 entries? Or just one?
We had agreed you wanted the moment of inertia about the $z$-axis, when you set up your coordinates appropriately.
17:40
@TedShifrin I have decide to compute all 9
because I am going to, for my project, use only the Z axis it is true
however for my research overall, it will be highly useful to have this
I am planning to also do a 3D simulation at some point
and, in general, I think i need practice with the intuition
so I can't see a reason not to do all 9
it doesn't take that long to do it when you use a symbolic calculator for half of it :')
I posted my approach here. If you get a moment to read my set up that would be awesome
from a math perspective, i'm most concerned the bounds of the integrals are wrong
from a physics perspective, you can see the concern i mention in the question
18:01
@TedShifrin I'm on holiday in Tuscany, the place is beautiful, but the internet is awful, so I'm only here occasionally!
18:25
How much time it took you to get comfortable with Real Analysis?
It’s been 3 months now, and I still feel a little uneasiness in simple proofs.
 
2 hours later…
20:26
Is there a philosophy why an angle-preserving Euclidean map corresponds to a distance-preserving hyperbolic map?
Why it "makes sense"?
20:40
@Emolga What's the exact statement
I figured it's involved enough to warrant a post, so I work on phrasing it now
But what I wrote is rather exact...
I mean, there's an iff
I don't know what a hyperbolic map is
a function in the disk is conformal iff it is a hyperbolic isometry
Oh sure, that's easy.
The Poincare metric is conformal to the Euclidean metric
20:52
Is that "expected"?
Maybe it is, if we start from saying hyperbolic space is where distance is a conformal invariant. So hyperbolic space is where maps preserve angles iff they preserve distance.
Ok so there are two components to this: (1) Angles work exactly the same as in $\Bbb R^n$ as in $\Bbb H^n$ (2) Conformal auts of $\Bbb H^n$ are isometries of $\Bbb H^n$
(1) is easy; just write down the Poincare metric
(2) can be proved as follows
imagine if less than a dozen people have the ability to turn the internet off in an instant.,
Extend a conformal automorphism of $\Bbb H^n$ to the ideal boundary $\partial \Bbb H^n = \Bbb R^{n-1} \cup \{\infty\}$. If it leaves $\infty$ fixed, it's just a Euclidean isometry. If it sends $\infty$ to $0$, compose with an inversion to make it fix $\infty$ and back to first case. If it sends $\infty$ to some other point on the boundary, translate and make it fix $\infty$.
oh wait don't worry about it that's literally true\
boop
And why are there no more hyperbolic isometries except those?
21:12
One way to write down all hyperbolic isometries is to use the hyperboloid model; $\Bbb H^n$ sits inside the Minkowksi space $\Bbb R^{n, 1}$ as the unit sphere, and the isometries are exactly those which preserve the Minkowski metric. So that would be $SO^+(n, 1)$
$\text{Isom}^+(\Bbb H^n) \cong \text{SO}^+(n, 1)$
21:25
@Emolga Ah maybe here's one way to intuit this. Let $\phi : \Bbb H^2 \to \Bbb H^2$ be a conformal automorphism. Recall that for any triangle $\Delta \subset \Bbb H^2$ with vertex angles $\alpha, \beta, \gamma$, $\text{Area}(\Delta) = \pi - (\alpha + \beta + \gamma)$. Thus, $\phi(\Delta)$ and $\Delta$ have the same area!
Of course, area preserving and conformal forces isometry
Hm I am a little surprised at how simple that was. Doesn't the usual proof use that weird complex analysis lemma with Blaschke factors?
I guess you need to know conformal map takes geodesics to geodesics, but that should be simple, right?
Not to me, a @Balarka.
Yeah. I was trying some boundary argument, but maybe that's not clear.
("area preserving and conformal forces isometry"?)
How do we go from areas to lengths?
21:40
@Emolga Remember that if $g$ is the Riemannian metric, the Riemannian volume is $\sqrt{\det(g)}$, so the scale factor pops out if you scale $g$
Using this, you can prove area preserving conformal forces the scale factor to be $1$
But anyway this doesn't help because apriori it's unclear $\phi(\Delta)$ is a geodesic triangle
You should post this as a question; it's a good one. Maybe its not as easy as I thought
I've never heard the fact that Emolga quoted before.
It follows from Schwarz lemma, right?
If you have a biholomorphic aut of $\Bbb D$, multiply with a Blaschke factor and a rotation to make it fix two points.
Then its identity
How did we end up in the holomorphic category? I thought we were in arbitrary dimension and just hyperbolic.
Let's do 2 dimensions; conformal is the same thing as holomorphic there
So given a conformal hence biholomorphic automorphism of $\Bbb D$ you want to show that it's a hyperbolic (orientation preserving) isometry
So the hyperbolic version of the Schwarz lemma is that holomorphic mappings are distance decreasing (in the disk model).
Well, it's easy computation that holomorphic automorphisms are isometries. That's even in my baby notes.
But it follows from what I just quoted for the Schwarz lemma. Apply to $f$ and $f^{-1}$.
21:54
Right
That's what I was saying
I mean the interpretation in terms of hyperbolic metric, not classical Schwarz.
It's often got Ahlfors attached.
But do you know a complex analysis slash computation free proof that conformal => isometry in $\Bbb H^2$?
No, that's the thing I said I'd not heard before.
The proof I was thinking for higher dimensions cannot be specialized to dimension $2$, because conformal automorphisms of $\Bbb R^n$, $n \geq 3$ are very restrictive: they are compositions of isometries, stretching and scaling
* and inversion
22:22
Are the hermitian (self-adjoint) operators on a Hilbert space spanned by the positive operators?
@BalarkaSen did you know that every cofibrant category is a poset?
22:43
whats a cofibrant category
it's a category such that the unique map from the empty category is a cofibration
 
1 hour later…
23:46
@Knight you write the whole thing without choosing epsilon and then at the end make it work. Practice. And draw pictures
@Knight maybe because you did Riemann integrals before you knew how to define a limit

« first day (3650 days earlier)      last day (1667 days later) »