« first day (2282 days earlier)      last day (3034 days later) » 
00:00 - 19:0019:00 - 00:00

00:00
whos good with Laplace Transforms, I have a question on how unit step functions work? Anyone Please? :)
@theartist as per the room's advice: Just ask; don't ask to ask.
@Obliv Cool :)
Ah. Yeah, seeing how someone else does something is nice, but it's not really "practive" if you don't figure it out on your own.
So it's useful to get an idea on how you're supposed to do something, but you can't just rely on reading other people's solutions.
That's why textbooks have exercises and not just examples.
@saturatedexpo
i heard you @AkivaWeinberger, without mentiining me;) hehe
What does "Saturated Expo" mean?
00:05
So in my notes it says that when you write $g(t)= (t-1)u(t-1)-(t-1)u(t-2)$, the notes say that this makes the function turn (t-1) ON at t=1 and turns (t-1) OFF at t=2. The problem I have with wrapping my head around this is that, why does it turn off at t=2, because at t=2 only the second term $(t-1)u(t-2)$ gets turned off, the first term remains, so how does the whole thing turn off?
it comes from starcraft 2, where you call an expansion (a new base with minerals to gather) which has been filled with "enough" workers "saturaded". @AkivaWeinberger
@TheArtist Try plugging in some $t$ bigger than $2$, say $t=10$.
If I'm not wrong, you get $9(1)-9(1)=0$.
It turns off because the first term cancels out the second.
@AkivaWeinberger $g(10)=(10-1)u(10-1)-(10-1)u(10-2)=9u(9)-9u(8)$
@AkivaWeinberger Doesn't cancel out ^^
i guess i'm just mad, that my High School degree doesn't give me a shit in studiing math. you can easily pass thourgh High School without any creativity.
@TheArtist Then I'm mistaken on what $u$ means.
What is $u(9)$?
00:11
@AkivaWeinberger step function
Then what is $u(9)$? Is it not $1$?
@AkivaWeinberger haha stupid me
@AkivaWeinberger thanks alot for that :)
You're welcome!
(I hadn't done this in a while so I was legitimately not sure what $u$ was)
(So thanks)
00:14
:) looks like you have a very good memory
00:24
@AkivaWeinberger but is an excercise of the same variety simply exchanging variables or slightly changing what's asked for? (i.e. speed instead of distance)
I'm not sure I understand
Another formulation: what do you think is a question of the same variety? Suppose someone asks: "i can bake 1 cake in 2 hours, how many cakes will i bake in 4 hours?"
I was the one that suggested the person you want to convince should do another exercise of the same variety, btw @akiva
oh SORRY ;)
variety meaning the same topic of interest that follow the same exercise format as before.
00:36
yo
and how do i give someone the tools to do it by himself without spoiling the result? in case of "i can bake 1 cake in 2 hours, how many cakes will i bake in 4 hours?" telling them about the rule of three?
00:56
hey @0celo7
hello
do you know Hahn Banach theorem for normed spaces ?
No
I don't know any functional analysis
oh ok
01:32
Suppose that X is normed space and $x,y \in X$ not equal to each other I am trying to think of closed disjoint subspaces which contains x and y.
I need it for something in functional analysis.
is there a trivial such thing ?
I guess it must have the point 0 in common
I just don't want them to share the point x or y respectively.
@Adeek what is wrong with taking closed balls?
subspaces being linear subspaces?
yeah
then no, they could be along the same line through the origin
if $x\notin\Bbb Ry$, then you can do it
I just want closed subspaces that seperate x and y.
Linear subspaces won't happen in general
How will you do it in $\Bbb R$, for instance?
01:39
what about a closed subspace that only contain x ?
and not y ?
...if $x\in\Bbb Ry$, then any subspace that contains $x$ will contain $y$, and vice-versa.
weird ..
Weird? They're parallel
How can you have one be in a subspace but not the other?
no weird because I want to prove something that needs this result.
Subspace or subset?
There's probably a condition that they're not parallel somewhere
01:41
well let us consider $R^2$ easier to visualize
If they're not parallel you can do it
Easily, in fact
Was a problem on my analysis midterm
I am trying to prove the following.
Let X be a normed space and $x,y \in X$ such that $x \neq y$ there exists functionals $f \in X^{*}$ such that $f(x) \neq f(y)$
I was thinking of using the following result
algebraic or topological dual?
$X^{*} = B(X,\mathbb{K})$ where $\mathbb{K} = \mathbb{C}$ or R.
I don't know what $B$ is.
01:43
bounded linear operator from X to K.
Ok, so topological dual.
I was thinking of using the following result.
Let X be a normed space, Y be a closed subspace of X and $z \notin Y$ . Then
there exists $F \in X^{*}$ such that $F|_Y = 0$ and $F(z) = 1$.
this result would get it right away
I don't think that result is true
However, I need to get a closed subsoace that contains x but not Y.
The kernel of a continuous functional is a closed hyperplane
you can't have it be a general closed subspace
01:47
?
The kernel of a linear functional is a hyperplane
Are you saying that $Y\subset\ker F$
oki
or that $Y=\ker F$
$Y \subset Ker(f)$
@Adeek you won't get that in general
01:49
no I mean you don't get it for any functional @0celo7
there exists one which gives you that property.
I can send you the proof of the theorem if you want.
I mean you can't find such a subspace in general.
I see
You need to treat two cases
The case when $x=ky$ and when $x\ne ky$.
If $x\ne ky$, then $\Bbb Rx$ is the subspace you're looking for
oh
You can prove it's closed because finite-dim normed spaces are Banach.
01:51
I guess we can consider the span{x} and span{y}
That's what I just said.
but this is true only when x and y aren't in the span of each other.
I said that 10 minutes ago.
I don't like saying things multiple times. Bye.
yes @0celo7 I was just trying to understand it in general normed spaces.
02:06
@0celo7 I solved it.
would you like to hear the solutions ?
solution.
 
2 hours later…
03:46
If $\Delta$ is the unit disk and $L$ is the trivial line bundle on $\Delta$, what does a metric on $L$ look like?
a metric on $L$ is just determining which vectors on $L$ have unit length, so it suffices to specify a positive function on $\Delta$
if the function is $f$, then you say $f(t) = \|(t, 1)\|$, where I mean the element $1$ in the fiber $L$
So I can say that, say, $f(z)=e^{-|z|^2}$ is a valid metric on $\Delta$. My book says that a metric is some inner product, which confused me.
It is an inner product on each fiber. But your positive function gives us such an inner product.
Note that when the fibers are 1-dimensional, an inner product is really easy to define!
04:04
I see. Thank you.
hey @MikeMiller
just proved really cool thing
Suppose that Y is subspace of normed space X, then if we have $T \in B(Y,l_{\infty})$ then there exists $S \in B(X,l_{\infty})$ such that $S|_Y = T$ and $||S|| = ||T||$
we just consider the sequence $T_i : Y \rightarrow K$ and use hahn banch theorem seperately on each i.
cool stuff happen in infinite dimension it is actually much more rich than finite dimension stuff.
Yeah, I like it a lot. But it's also very wild.
yeah
 
2 hours later…
06:34
Hello?
It's me.
I was wondering if after all these years you'd like to meet.
Ummm... OK.
I have a sequence...
Which needs a formula.
This is violating the chat rule, but can I ask for this sequence here?
 
2 hours later…
08:24
anyone see this
0
Q: Solve Recurrence Relation for maximum and minimum in an array.

souravI know that recursive equation of this algo is $T\left ( n \right )=2T\left ( \frac{n}{2} \right )+2 $ Given that $T\left ( 1 \right )=0,T\left ( 2 \right )=1 $ and its solution is also given here,ijust want to clear my doubt where i am stuck at. I solved this eqtn as--: $\Rightarrow T\left...

 
1 hour later…
user228700
09:32
Hey everyone :-) I've a quick question. Are the parametric coordinates for any parabola given by $(at^2, 2at)$ ? ($a$ is one quarter of the length of the latus rectum and $t$ is the parameter)
https://www.kickstarter.com/projects/2086970389/whisk-wiper-whisking-made-easy
It feels like there's a certain maths principle being exploited here, however i am not sure what it is
0
Q: Solve Recurrence Relation for maximum and minimum in an array.

souravI know that recursive equation of this algo is $T\left ( n \right )=2T\left ( \frac{n}{2} \right )+2 $ Given that $T\left ( 1 \right )=0,T\left ( 2 \right )=1 $ and its solution is also given here,ijust want to clear my doubt where i am stuck at. I solved this eqtn as--: $\Rightarrow T\lef...

@Secret I don't see why there would be. It is just one of those simple ideas that are obvious in hindsight but requires someone to be the first to come up with them
I don't know, I might be distracted by that nice flower pattern formed when the whisk slide out of it and somehow it exploit the geometry of something of the whisk
10:01
Hello!!
Does someone of you have an idea for my question:
0
Q: Calculate the velocity of the particle

Mary StarWe have that a particle is moving with the effect of the gravity $F=-mgj$ and it has the initial position $r(t=0)=10j$ and initial velocity $r'(t=0)=10i$. We assume that $m=g=1$. How can we calculate the velocity when it falls on the ground ($y=0$) ? Do we use the formula $F=ma=m\frac{du}{d...

?
10:36
When we have that the motion of a particle in polar coordinates is givwn by the realtions $r=\cosh (t) \ \ \theta=t$, do we have that the acceleration is equal to $r''=\cosh (t)$ ?
11:05
Ah no, it is $a_r=0$ and $a_{\theta}=2\sinh t$ and so $a=2\sinh t$, right?
11:48
0
Q: Solve Recurrence Relation for Maximum and Minimum in an Array.

souravI know that recursive equation of this algo is $T\left ( n \right )=2T\left ( \frac{n}{2} \right )+2 $ Given that $T\left ( 1 \right )=0,T\left ( 2 \right )=1 $ and its solution is also given here,ijust want to clear my doubt where i am stuck at. I solved this eqtn as--: $\Rightarrow T\lef...

12:14
@TedShifrin Bonjour Ted, j'ai regardé livre de Stephen Smale, il est bien fait. Cependant, je n'ai pas trouvé d'exercices où Cauchy-Lipschitz; théorème d'explosion en temps fini; Inversion locale soient mélangés. Aurais-tu, par hasard, des noms de Théorème (ou exercices) utilisant ce mélange ? Par exemple, il y a le théorème d'Hadamard qui stipule que si f est de classe C^2 et que l'image réciproque de tout compact est compact alors f est un difféomorphisme. Merci d'avance
12:41
0
Q: Area of triangle formed from equation of of its sides.

RamanujanThe area of triangle formed by the lines $$a_1x + b_1y + c_1=0$$ $$a_2x + b_2y + c_2=0$$ $$a_3x + b_3y + c_3=0$$ Is $$\frac {\Delta^2}{2\lambda_1 \lambda_2 \lambda_3}$$ Where $$\Delta = \begin{vmatrix} a_1 & b_1 & c_1 \\ a_2 & a_2 & c_2 \\ a_3 & b_3 & c_3 \\...

user116211
13:14
@balarka, completed the level 5 of that Halloween game.
user116211
is $$\sum_{h=1}^n h^k$$ always $O(n^{k+1})$?
it is if n and k are naturals
13:30
@Sophie How would you even define that if $n$ isn't an integer?
I want to generalize it for all real k
Okay.
It's not going to be true for $k<-1$, then, since each term is bounded above by $1$ and therefore the sum is bounded by $n$ not $n^{k+1}$
Indeed, if $k<-1$ then this converges to the Riemann zeta function $\zeta(-k)$ as $n\to\infty$
it might still be true for $0\geq k >-1$, but I dunno
@Semiclassical Well, being bounded by a constant implies being bounded by anything increasing
...yeah, point
It might not be a useful bound, but it is a bound
13:47
@Semiclassical It's true for $k = 0$ trivially.
Being equal to $n$ it is clearly $O(n^{0+1})$
Caption: Abstract algebra experiment: Tracking the dependence of the entries in a Cayley table given axioms
Note if we impose the axiom aa=a and aq=e, then ae=e can be derived. After that, when all possible words were written out, it is clear (assuming the algebra is non abelian but associative) that defining the word qa will give most of the entries in the table
Can I simplify the sum $\sum_{n=0}^\infty \frac{(-x^2)^n}{n!}$ to $\sum_{n=0}^\infty \frac{(-1)^n*x^{2n}}{n!}$ ?
$-x^2=(-1)x^2$
13:59
yes
just wanted to verify
Huy
Huy
@BalarkaSen: wanna help me with some basic geometry?
@huy What counts as 'basic'?
Caption: Therefore one example is setting the axiom qa and then ee, and you lock all the entries in the cayley table. There are other choices, but note the correlations introduced when a axiom is set
it's probably phD level
@Huy What do you need?
14:14
@MAFIA36790 I did it days ago
@Huy Sure, if I can
Oh wait
Huy
Huy
I remember from high school that one can construct the image of the inversion at a circle by drawing chord and tangents (or the other way around, depending on whether we start inside or outside of the circle). I want to prove that this construction indeed yields inversion at a circle, and I see how I can do this with this theorem (a^2 = pc) of which I don't know the English name.
Is it gonna be Euclidean geometry
Huy
Huy
Is there a way to prove it using similarity/intercept theorem?
What even is inversion at a circle?
14:17
@Huy Prove what, exactly?
Huy
Huy
$z \mapsto 1/\overline{z}$ if you work in $\mathbb{C}$
Oh, $a^2 = pc$. OK.
@Huy that's hardly Euclidean geometry
Huy
Huy
you usually define the inversion by saying that $MP' = r^2/MP$
mmm
Huy
Huy
14:18
$P'$ is the image of $P$, and those are distances
so obviously I can rewrite $MP'/r = r/MP$ which is why I'm thinking there might be a proof using ratios/similarity
@Huy the wiki article mentions this fact
Huy
Huy
@0celo7: the wiki article mentions it can be proven using those concepts?
(I only checked german wiki and there they referred to a^2 = pc)
no, it mentions that OP is to r as r is to OP'.
Huy
Huy
ah ok
@Huy I am not sure how to describe your picture in terms of circle inversion.
Huy
Huy
14:23
@BalarkaSen: do you mean this?
(not sure what you mean otherwise)
I misunderstood the question; I thought you wanted to prove the thing in the Pythagoras' theorem picture you linked
Huy
Huy
no, I want to prove MP' = r^2 / MP. I know it follows directly from a^2 = pc (given in the first picture), but I want to avoid anything related to Pythagoras
why do you hate ol' Pythag
Huy
Huy
cuz I'm not teaching Pythagoras yet but circle inversion would be a nice end to the chapter about maps of the Euclidean plane
@Huy Ok, so trying to understand your theorem statement. What is the given relation between P and P'?
Huy
Huy
14:28
@BalarkaSen: take P inside the circle. construct the chord perpendicular to MP, intersect with the circle. construct tangents. their intersection is P'. theorem: MP' / r = r / MP
(and the same should work if P is outside the circle)
fair enough
Huy
Huy
that (to me) would prove that "this given construction is indeed circle inversion"
Huy
Huy
(need to go afk for 15, but will be back if you figure it out or have an idea)
let me draw something
@Huy The quadrilateral you have in your picture is symmetric about MP', not? Because the two sides are radius, and the other two are tangents from the same point, which have the same lengths.
So the central angle at M is divided into two equal angles by MP'. That should imply similarity of the appropriate triangles.
Well, the important part is that the angle at P is divided into two equal parts by MP', but w/e.
Huy
Huy
15:08
@BalarkaSen: right, those two triangles are similar, thanks
 
1 hour later…
16:31
@Huy sure, no problem. what're you upto these days?
17:04
[Abstract algebra]
Everything is easier with matrices:

Suppose one is being asked to compute the associative laws of $(\mathbb{Z}/3,+)$ by hand. First rewrote the cayley table of $(\mathbb{Z}/3,+)$

\begin{array}{|c|c|c|} \hline + &0&1&2\\ 0& \hline 0&1&2\\ \hline 1&1&2&0\\ \hline 2&2&0&1 \\ \hline\end{array}

as follows

$$\begin{pmatrix}0&1&2 \\ 1&2&0 \\ 2&0&1\end{pmatrix}=\begin{pmatrix}0&0&0 \\ 1&1&1 \\ 2&2&2\end{pmatrix}+'\begin{pmatrix}0&1&2 \\ 0&1&2 \\ 0&1&2\end{pmatrix}$$

where $+'$ is operating $+$ entrywise, Then associative law (where $a\in \mathbb{Z/3}$) can be condensed into
Theoretically , one can automate the process in Mathematica by defining $+'$ and $+$
17:18
I'm reading about disc/vector bundles. If $p \colon E \to B$ is a disc bundle with $B \subseteq E$ and $p$ is the identity on $B$, then why is the inclusion of $B$ into $E$ the zero section of $p$?
@abenthy Eh, the zero section is just the canonical way the base space is included in the total space of the disk/vector bundle. There may be other inclusions if eg $E$ is the trivial bundle.
If $\varphi \colon U \times D^n \to p^{-1}(U)$ is a local trivialization, $U \subseteq E$, then why is it forced to sent $(u,0)$ to $u$ for $u \in B$?
because this would make $B \hookrightarrow E$ the zero section, right?
@BalarkaSen Sure, my question is that if $B \subseteq E$ and $p$ is the identity on $B$, then why is this inclusion the zero section?
Ah.
Well, write the inclusion down in local coordinates.
Ah, and then use that the zero section is unique?
I've gotten to the point in my self-study where it's starting to look like I'm on the path to learning cohomology.
Just from peaking ahead into theorems and other things.
If I remember correctly, @MikeMiller does a lot of work with that, right?
17:48
Sorta.
In some sense most every topologist (and algebraic geometer, and half of number theorists) do a lot of work with cohomology.
Is there any reading you recommend I do before I get there?
Nah.
What are you reading?
Differential Geometry and Topology side-by-side
Boothby and Munkres specifically.
I've read that the exterior algebra of differential forms has something to do with the de Rham cohomology elsewhere in my studies
More specifically, I ended up digging into geometric algebras.
Rather than algebraic geometry.
18:22
Schoolwork is tiresome.
@BalarkaSen That's why I left. Footnote: worst mistake I've made in my life.
I think we have different things in mind when we say "school" :)
Elementary schoolwork was unrealistic.
@BalarkaSen I'd imagine--what do you have in mind?
How is a child supposed to make a scale replica of Egypt one week and then write a book report on Martin Luther King the next and STILL have time to make a working baking soda volcano?
18:25
high school
You're in secondary school?
@BalarkaSen College felt largely the same to me, to be honest.
More proofs, though that's because my area's HS curriculum was poor.
yeah, @Axoren
I don't believe it.
@Fargle: I can believe that, but makes me feel scared.
18:27
@BalarkaSen Don't be--just find that which intellectually stimulates you.
Unless you meant w.r.t. proofs, in which case I'll say it's never TOO demanding in undergrad.
Especially for someone at your level coming in, as you're by far better than I am and I'd like to call myself a Pretty Smart Dude(tm)
I meant the bit about college feeling the same to you as high school.
I dunno, I guess I'm also just easily exhausted?
Though, like I said, if you stay stimulated you'll be fine. That's the main difference, and it's a big one.
High school gives a lot of busy work. College gives a lot of work that keeps you busy, but is mainly intended to actually teach you.
I'd think so. That's reassuring in any case.
Definitely made me feel better my first couple of years, haha.
Hi @TedShifrin
18:32
Speaking of college, hi @Ted!
good evening @Ted
Hi @Balarka, @Fargle — I'm currently 3 blocks from Stanford.
Nice! I trust you got there with no hiccups?
cool. hope you're enjoying the trip.
Hi @Alessandro — going to lunch in an hour with a former UGA student who's Italian :)
No hiccups ... Just 8 1/2 hours of driving.
18:34
what's UGA? Eating with Italians is usually a good idea :P
@Balarka: My students never complained about busy work. They had a different sort of complaints!
@TedShifrin Ew.
@Alessandro: UGA is where I used to teach.
is it fair to notate $f\circ f$ as $f^2$?
@TedShifrin: Busy work is fine, as long as it's thinking about cool math, watching movies and reading!
@saturatedexpo I think standard notation is $f^{\circ 2}$
18:37
thanks
That doesn't sound like busy work ....
I've seen $f^n$ for iterated functions, but it's usually stated what it means the first time the symbol appears since it could be interpreted in different ways
@BalarkaSen i've never seen that notation.
Right. If we're studying iterations of maps, then $f^n$ is used.
so if i have a recursive function with a shown limit i could notate that as $f^{\circ\infty}$?
ugh, I wouldn't do that, @saturate
18:38
@saturatedexpo Balarka's given notation is a good choice as it's not ambiguous, but $f^2$ is fine as long as $f$ isn't something like $\sin$ where the notation $\sin^2 x$ means $(\sin x)^2$.
Then there's always $\sin^{-1}$.
while looking it up I found out that there's a university in France which is also called UGA @Ted, now I only need to find out where Georgia is because my knowledge of the US is pretty terrible
@Alessandro It's right above Florida (a.k.a. America's, erm, leg)
Well, I'm not there anymore, @Alessandro :)
I've also seen $f^n$ for the $n$-th derivative, but that's usually written $f^{(n)}$, just use the symbol you prefer but write explicitely what it means the first time you do
18:41
Context is important in math notation!
Well I don't know where Stanford is either
Yeah, for derivative it should have parens.
Northern CA, @Alessandro.
I don't know where most universities in Italy are, to be fair.
California turns out not to be on the coast where I expected it!
Oops.
Biggggg country ..
why do people use "CA" for "California"?
18:45
Two-letter abbreviations for all states
Any good math today, Balarka?
@TedShifrin I proved the Gauss-Bonnet yesterday.
Not how you proved it, I guess, though.
the coast to coast distance in the US is more than $3$ times the length of Italy, that's huge
You only have epsilon of the theorem.
The boundary term is very important.
18:48
Ok, yeah, I did it for closed surfaces.
The holonomy connection is pivotal.
Again, differential forms makes a winning proof (generalized by Chern to high dimensioms).
Plus leads to Chern-Simons invariants, etc.
Did you prove it using diff. forms. in 3.3?
Interesting, nice.
Apparently my advisor is a fellow of the AMS now.
18:55
whoa, great
@MikeMiller Welcome to the club :)
He would probably tell me it would be better for me if I was a fellow of the AMS.
Yup. I saw that. I know (of) about 12 on the list.
@TedShifrin One of them is from UGA, right?
Hmm. I just wrote an answer to a question in Rudin but I'm really really unsure if I've done it right.
18:57
Danny Krashen (or how you spell it?)
Danny, yes.
That's what he said when I told him it would be helpful if he won the fields medal.
LOL .... you ass.
What question, @Fargle?
18:59
It's the one in chapter 1 walking you through defining powers for all real numbers.
Yeah, me too.
@MikeMiller Who is your advisor again?
Try to find out from the list ;D
I think he's the only one cited for Floer homology
Also hi everybody
Hi bye @Danu
00:00 - 19:0019:00 - 00:00

« first day (2282 days earlier)      last day (3034 days later) »