« first day (4669 days earlier)      last day (647 days later) » 

00:00
your right ted, i should be more attuned to the expression
@D.C.theIII Even better, an isosceles right triangle.
i saw $\phi: A \to B: a \mapsto \phi(a)$ once, and now i use it everywhere
I'm going to turn into @冥王Hades and make finding regions of integration my obsession instead of angles in geometry
@shin That's pretty standard. Mike Artin used to use $a\rightsquigarrow \phi(a)$, and so I copied him most of my life.
yeah artin uses funky arrows
i didn't know it was standard
cool
00:04
What's abhorrent is the people who use $\mapsto$ instead of $\to$.
!!! but how do you distinguish a map between sets from a map between elements
what if I make a huge mistake and actually interpret $a$ maps to $B$ the entire image
I see people writing $f\colon X\mapsto Y$ a lot on MSE. They're wrong, of course.
this will create disasters
ohh, my god that is disgusting
but artin's arrows are just too long
jesus christ artin we get it please shorten your arrows
@D.C.theIII in the Old World we have textbooks to prepare kids to take final exam in HS and usually these books throw very sophisticated equations and inequalities at the reader.
00:31
@noballpointpen A bit vague.
@TedShifrin excuse me?
“The Old World” is a huge place.
I am letting this to the reader's curiosity.
“Leaving”?
Sorry, not native English speaker.
00:42
not "a" native English speaker???
hahaha just kidding
Yes, I often forget to put articles or put them in the wrong places.
peak slav energy
01:02
ted: its size is all a matter of perspective, man. maybe it is we who are small.
01:23
I just have no idea to what it refers, whether I'm small or not.
what if it doesn't need to refer to anything, what if its meaning is really the friends we made along the way
What are friends?
proofs you can understand are friends. maybe the converse is true, so that friends are maths you can understand
Gives a whole new meaning to antisocial.
this is why a proof you cannot understand is an enemy
ted: of interest wrt the "compactness" of varieties en.wikipedia.org/wiki/Complete_variety
01:46
That ain’t happening unless you’re in projective space.
Let $M>0$ be given. Define $\gamma = x/y^2$. We know that $|\gamma + 1/\gamma| > 1$, regardless of the values of $x$ and $y$. If $x \to 0$: $\gamma \to 0$ and $1/\gamma \to \infty$. Fix $x$ to satisfy $|\gamma| < 1/M$ and choose $y$ small enough so that $|\gamma + y^2/\gamma| < 1/M$.
Right?
Not that I can see. Changing $y$ changes $\gamma$. You have too many letters in this alphabet soup.
I don’t see where you ever used your first thing. Yes, $x+1/x\ge 2$ for $x>0$. So?
02:01
It was crucial in the first half of the problem and I kept it near my hypotheses. Going to work out something from it...
Sorry for that gamma, how I write it on my paper sheet is much more distinguishable from y.
That’s not my issue.
What are you actually trying to do?
Show that $g(x,y) = \frac {xy^2} {x^2+y^6}$ has no bounds near $(0,0)$.
Nevermind, I am going to leave right now. Maybe will hit the chat with it tomorrow. Goodbye.
Have you seen other functions lime this?
Bye.
02:50
splitting field is such a metal name
fuaaa now that is metal
and nobody calls them killing "vector" fields, either, just killing fields. wikipedia must have chosen that as the name of the entry just to declutter the wiki namespace a little bit
the slayer song is actually about the mathematical object
on the subject of striking mathematical names, i've recently encountered the unfortunate Tits buildings
which wikipedia was quick to name "Buildings (mathematics)"
there's also the Damm algorithm
and en.wikipedia.org/wiki/Cuntz_algebra (sometimes a problem with text checkers, pronunciationwise it is not a problem, "koonts")
03:03
hahaha that's great
let's not forget the Watda-Fark Theorem
 
4 hours later…
07:21
0
Q: How to prove that outer measure $|A|=\lim_{t\to \infty}|A\cap (-t,t)|$?

KoroLet $A\subset R$. Then, $|A|=\lim_{t\to \infty}|A\cap (-t,t)|\tag 1$ $|A|:=\inf\{\sum_j l(I_j): A\supset \cup I_j\},$ where $I_j$'s are open intervals in $\mathbb R$ and $l(I_j)$ is the length of the interval $I_j$, and infimum is taken over all sequences of open intervals whose union contains $A...

have you looked at answers to related questions on MSE?
6
Q: Continuity of outer measure induced by measure from below

steveThis question comes from Bass, Ex 4.15. Given a finite measure space $(X, \mathcal{A}, \mu)$, we can define an outer measure $\mu^*$ as $$ \mu^*(A) := \inf\{ \mu(B) : A \subset B , B \in \mathcal{A} \} \:. $$ (One can check that this is an outer measure and that $\mu^*$ agrees with $\mu$ on $\m...

11
Q: Continuity from below for Lebesgue outer measure

AceLet $\{E_n\}$ be an increasing sequence of subsets of $\mathbb{R}^n$, measurable or not. Then $$m^* \bigg( \bigcup_{n=1}^{\infty}E_n \bigg)=\lim_{n\rightarrow\infty}m^*E_n$$

terence tao has a related solution in math.ucla.edu/~tao/245a.1.04f/midterm.pdf
Ahh, I didn't see this one. I know this continuity for measures but didn't know that it holds for outer measures too.
And considering that it is in chapter 2A of Axler's, it didn't occur to me that it was expected to know this result.
I did see other related questions such as |A|= $|A\cap(-t,t)|+ |A\cap (-t,t)^c|$
But still the same problem (described in the post): limit of the rightmost expression is to be shown 0 in case |A|< infty.
07:49
Isn't that one of the convergence theorme?
08:34
Let V=M_n(\Bbb{R}) and A\in\V . Define T_A(X) =A^TXA . I want to know the structure of the operator T.
Rank, nullity, eigenvalues, diagonalizibility etc
I have found an amazing result trace(T_A) =trace(A) ^2 and det(T_A) =(det A)^{2n}
09:23
If the determinant of a matrix is the area scale factor, is the square root of the determinant of a matrix the linear scale factor? I thought so but it doesn't seem to work with all matrices? Why not?
Done! Please ignore my previous comment.
10:14
"harmful" in what way?
11
Q: Is it harmful to use the word "Cancel"?

JTP - Apologise to MonicaElsewhere, among a group of high school math teachers, I encountered a discussion of the term 'cancel'. Most (>20) people in the discussion had very strong feelings about why the term should be eliminated. I searched among posts on this stack to find the word common used and its meaning seems wel...

Don't use that word you may harm yourself!
10:27
@Xnero it depends on the dimension. In a $3\times3$ matrix, the determinant represents a volume. In a $2\times2$ matrix, an area. In a $1\times1$ matrix, a length.
 
2 hours later…
12:12
Is there a function in PARI/GP that computes the n-th fibonacci-number modulo some integer p>1 , where n is very large ? Or must I program such a routine ?
12:52
By definition, $$\limsup\limits_{n\rightarrow\infty}x_n=\lim_{n\to\infty}y_n,$$ where $y_n=\sup\{x_n,x_{n+1},\ldots\}$.
In problem 27a) in chapter 22 of Spivak's Calculus, we are asked to show that $y_n$ converges if $x_n$ is bounded. Spivak argues that $y_n$ is decreasing and bounded below by a lower bound of $x_n$. That is all clear.
However, consider the sequence $-1,0,-2,0,-3,0,\ldots$. It is not bounded below, yet the limit superior exists and is finite, namely $0$. This example confuses me; by Spivak's argument, it seems like $y_n$ needs to be bounded below in order for it to converge, but the example shows it need not be. Any help is appreciated.
EDIT: "...it seems like $x_n$ needs to be bounded below in order for $y_n$ to converge..."
13:37
To summarize my problem; what conditions does $x_n$ need to satisfy in order for $\limsup\limits_{n\rightarrow\infty}x_n$ to be finite?
It seems like we have the following three cases:
Case 1; if $x_n$ is only bounded below, the limit superior is not finite.
Case 2; if $x_n$ is only bounded above, the limit superior may or may not be finite (cf. $-1,-2,-3,-4,\ldots$ and $-1,0,-2,0,\ldots$).
Case 3; if $x_n$ is bounded above and below, the limit superior is finite.
 
3 hours later…
16:21
@schn If the sequence is bounded above and there is a subsequence not converging to $-\infty$, then the limsup will be finite. :)
$R'/R$ algebraic extension of rings, $\Omega$ algebraically closed. then $\phi:R \to \Omega$ extends to $\phi': R' \to \Omega$
the above proof is a bishhop
@user858770 In that it can lead students to do sloppy and, ultimately, wrong manipulations.
I would hope that the word "cancel" is first used with numerical fractions without any difficulty.
it's lead me in the recent past to being hopelessly confused about the difference between an integral domain and a field for a couple of hours
integral domains have cancellation, fields have invertibility
i was like, huh isn't that the same thing
Hello World!
Any suggestions upon this:
0
Q: Find a power series solution about the regular singular point $x_0=0$ of the differential equation: $x^2\frac{d^2y}{dy^2}-x\frac{dy}{dx}+(x^2+1)y=0.$

Thomas FinleyFind a power series solution about the regular singular point $x_0=0$ of the differential equation: $$x^2\frac{d^2y}{dy^2}-x\frac{dy}{dx}+(x^2+1)y=0.$$ My Attempt: We assume the power series solution to be of the form $y=\sum_{n=0}^{\infty}c_nx^{n+r},$ such that $c_0\neq 0.$ Now, $y'=\sum_{n=0}^{...

17:00
@shintuku Um, no. :)
Can someone recommend me a Integration book?
@robjohn It's been like 6 years since I came to this platform.
@robjohn I still remembber you but i used a different account back then called "The Artist", not sure if you remember me. Anyway do you remember this was this lady from romania who was into solving definite integrals? She had this book, do you remember what it was.
not a book she has written but a book she used to recommend a lot.
17:16
@GCRClassic Are you talking about Chris's Sister?
@robjohn Yes the name had "sister" as I remember. how is she? She is a self taught mathematician.
@robjohn sounds about right. Let me check on the sites under users.
@robjohn can't find her searching the name. Do you remember the name of the book she used to be very fond of?
I don't
Do you have the link to her profile?
Is she still active here?
17:30
No, but her (or maybe his) book was published several years ago.
@GCRClassic If so, they have changed names.
@TedShifrin it was? I am so glad. Do you know the name of it
how much differential geometry do i need to determine whether an arbitrary function $f:\mathbb R^5 \to \mathbb R$ is convex
 
1 hour later…
18:40
does anyone know how to compute the mean of standard gaussian conditional on the value being greater than 0?
can someone help me here?
0
Q: How can I draw the image of this fractional linear transformation?

user1294729 Let me consider the map $$T:\Bbb{C}\setminus \{-i\}\rightarrow \Bbb{C}, ~z\mapsto \frac{z-i}{-iz+1}$$ I want to look at it as a map from the upper half plane to the unit disk, because there I know that it is an isometry if I endow the upper half plane with the Riemannian metrig $g_\Bbb{H}= \frac...

19:01
@shintuku Arbitrary functions are never convex,
Well this is kind of embarassing:

$$\int_0^1 \int_v^1 \frac{1+2u}{(u^2+u+1)^2} du dv$$

Let $k = u^2 + u + 1 \rightarrow dk = 2u + 1$. Therefore:

$$\int_0^1 \int_v^1 \frac{1}{k^2} dk dv = \int_0^1 (-1 + \frac{1}{v}) dv = -v \bigg]_0^1 + \log(v) \bigg]_0^1 = Problems$$


Did I mess something up in my limits of integration?
i meant a given function @TedShifrin
@shintuku Arbitrary functions are never convex. Why the word arbitrary? And what is your definition? Should just be positive definite Hessian everywhere.
a given, fully determined function
THe solution in the book is $\frac{\pi \sqrt{3} - 3}{9}$
19:07
@D.C.theIII Why that order of integration? When you substitute, don’t you have to change limits? Grr.
@shintuku What suggests that “differential geometry” is involved?
Fact: For every $\epsilon>0$, there exists a closed subset F of [0,1] such that F contains only irrational nos. and that outer measure of F i.e., |F|> 1- $\epsilon$.
Jesus...Talk about me being sloppy.....I didn't even do the "lazy" way doing the subsititution correct.
@TedShifrin seems like it's what comes next after a multivariable calculus class
and haven't seen anything beyond n=3 in the multivariable calculus texts
I chose that order of integration because I wanted to take advantage of the substitution that could be done. But this was in my fit of absolute laziness
Maybe multivariable analysis would be relevant, shin. But you still haven’t told me a definition.
19:12
shin: does mathjax work on discord?
yep @Koro, but the server you're on must have a latex-translation bot
@TedShifrin continuous functions in 5 real variables
say, what would come after your text, if i'm interested in higher dimensions?
Huh? Continuous? Huh?
yeah
My text is not limited to $3$, but there aren’t any foundational differences between $3$ and $n$.
You’re supposed to define CONVEX.
oh, well i'm interested in the convexity on shapes in 5 variables, in the sense that, increasing a variable in one dimension locally leads to an increase in the image
19:16
This is not a definition of convexity.
yeah, i know
that's why i'm looking for treatments of the stuff
i'm wondering to what extent we can use intuition in 3d or 2d images to define geometrical notions in higher dimensions too
If you understand convexity of a function of three variables, it won’t change.
well in three variables it's simple, cut the function with a tangent plane everywhere and make sure the function stays above it
then it's convex
maybe hyperplanes or something can translate this in 4d dimensions
but 5?!?!?
You’re talking about graphs or level sets? You have been sloppy.
there are a billion characterizations of convexity and most of them generalize in very predictable ways, but you do have to pick one.
19:22
Careful multivariable calc in $n$ variables is valid for any $n$, in general.
@shintuku I see. Thanks.
shintuku one way to think about the stuff about lying above tangent lines or planes is that it's deriving from a characterization of closed convex subsets of R^n (i.e. something you could potentially consider without having a function in mind at all). this prevents you from maybe having to bake the "tangent" part of "tangent plane" into a definition of convexity.
convex things are often automatically differentiable or close enough to that to have well defined things that we usually associate with differentiability, but you don't want that in the definition of 'convex'
or you might not want it. i dunno. it's up to you.
more than convexity in particular, i'm trying to find a geometrical property in 3d or 2d space that could be of use to begin classifying shapes in higher dimensions
with your comments, well i'm seeing convexity might not be it
that's why i'm after a book studying shapes in higher dimensions, and well i thought that it could be a book using analysis or calculus since i know a bit of those
convexity is a useful property, meaningful for e.g. any subset of any vector space (a subset S of V is convex if whenever u, v are in S, so is the line segment between them, i.e., so is (1 - t)u + tv for all t in [0,1]). whether it's useful for "classifying shapes" might depend on what you mean by classifying and shapes.
a helpful and less galaxy brained way of approaching this is from the bottom up, with examples. e.g. are a square (with its boundary and interior) and a rectangle that isn't a square (with its boundary and interior) going to be thought of as 'the same shape' or not. same question with rectangles of different 'aspect ratios.'
that question can have more than one answer even in R^2.
what about just within squares, are they all 'the same' or do you care that one has more area than another, can tile another, etc.
it's fine to be galaxy brained about it if we're just ruminating about the possibilities, but to get more specific, you often have to make choices about this kind of thing, because they matter. even in R^2.
hm yeah i understand
19:41
If you really are talking about “shape,” then, yes, this could be Riemannian geometry, but start with surfaces in $\Bbb R^3$, as most notions of curvature, etc., are based in notions you learn there.
oh, so say i want to talk about shape, then i should go through differential geometry of surfaces first, and then read something about Riemannian geometry?
@TedShifrin can you maybe help me in the question I asked above
20:08
What is the question?
@TedShifrin I'm a bit confused here math.stackexchange.com/questions/4701289/…
because if I let wolframalpha draw it, it does not look as my picture
20:23
What are you drawing?
at the moment I drew something like this
But now I wanted to consider maybe the vertical line through -1+i but I'm confused since the imace of this should also be orthogonal to the blue half circle and it should pass through $-1,2/5+i/5, i-2$ if I'm not mistaken
The real axis maps to the whole unit circle , no?
Ah yes it does sorry
Vertical lines will map to circles orthogonal to the unit circle and passing through $i$.
but I mean the image of $l_2$ is not a circle
20:32
$l_2$?
yes so my red line
i denoted it by $l_2$
is $l_2$ a line or a ray? (It looks like a ray in the first picture but a line in the second)
Right. It maps to the vertical line. Lines and circles are basically interchangeable in this world.
Well, it’s easier to think about the whole plane. Restricting to the half-plane, it maps to a line segment.
okey but do I then really need to compute the images of some vertical lines or is there an easier was
oh. line segment. ignore me
20:36
the vertical lines other than l2 map to literal circles. this phenomenon of having one line or circle being somewhat exceptional in the usual R^2 sense, but having the exception vanish if you treat lines and circles as the same, is general to moebius transformations.
user: in asking what you "need to" do, how are we supposed to know that? is there an underlying problem here that has yet to be stated?
same question. i don't know what it means to "draw the image" of a fractional transformation, or what is specifically meant by "I want to find out where the coordinate system is sent around the origin." i can't sketch an infinite number of circles, for example.
one could come up with an equation satisfied by the image of the line Re(z) = t, in terms of t, but i don't know if that's what is asked for. we're suddenly talking about "the images of some vertical lines," well which ones?
no I only would like to think about this map how does it look on the coordinate lines, therefore I wanted to start with vertical ones and look where they are mapped to, but my qestion is do I need to compute this explicitly for some lines to find a pattern or is there an easier way, so yes I mean they must pass i since every vertical line contain infinity and they must be orthogonal to the unit circle but is this enough to draw it
You can construct the circles easily geometrically because of the orthogonality to the unit circle.
for example, it isn't clear to me if this isn't a question about graphing this in software, or if geometrical guidance about what the lines get mapped to is enough.
20:41
It’s more natural for differential geometry to work in polar coordinates in the disk, anyhow.
@TedShifrin so you mean like this
ah okey I see perfect thanks!
$\infty$ maps to $i$; that’s how we figured this.
a few more branches and you'll have a pretty tree
20:45
but why does wolframalpha print me this one here?
Is it Christmastime already? I always called the Thomae function the Xmas tree function cuz of its graph.
what wolfram alpha is drawing isn't wrong, it's just not very useful here
if troubleshooting or revising whatever you are inputting to wolframalpha is really what you want to know about, it might help to provide that input. you might not find that the same people are interested in it
if i wanted to do this myself, i would manually work out the parametric plot
The horizontal lines also pass through $\infty$, right?
I don’t find that plot very helpful. It’s toooo cluttered at $i$.
20:49
it's not even especially clear which subset of the z-plane it's using in the first image
yes the horizontal lines also pass through infinity
Good point, Semi.
@user1294729 what was the input you used to get that?
Crazy man integrals being assigned by this man......😅😅
is there an argument why only the imaginary axis is mapped to a line and the rest is mapped to circles?
I put in the map $\frac{z-i}{-iz+1}$
20:53
i see it
Yes. How many lines through one point of the unit circle are orthogonal to the circle?
in general i think the 'complex map' feature in WolframAlpha does a pretty bad job
@D.C.theIII Or you messed up again.
infinitely many no?
No..I got it
20:56
@user1294729 Try again.
I had to get an arctan and use a "sneaky" subsitution
ah sorry lines, I thought about circles, no then there is only one line
The type of sneaky substitution I hadn't done since Calc II days of yonder
Well, that is all expected to be used. You need partial fractions, I guess.
No partial fractions didn't work
20:59
what integral was it?
am I right?
Used a sneaky substitution and then got an arctan expression
Of course it works.
@Semiclassical $\int_0^1 \frac{1}{x^2 + x + 1} dx$
That”s a trivial integral.
You get $\int \frac{u(2u+1)}{(u^2+u+1)^2}du$.
21:03
Ahhhh..... That's where I could've done partial fractions
Whereas doing it in the order I did it.....I created chaos
When I did it in the other order I had $\int_0^1 \frac{-1}{3}dv + \int_0^1 \frac{1}{v^2 + v + 1} dv$
and then to go through the second integral I did the substitution $k = \frac{2v + 1}{\sqrt{3}}$ and it eventually worked out
With a lot more frustration. When I should've recognized the partial fractions in the original order I had planned on doing it.....
oh I also had to complete the square before I did the substitution
yeah, there's something general there. completing the square is usually what tells you which substitution to do.
It was putting in the $\sqrt{3}$ that made it tricky and not straightforward for me
$\sqrt{1-1/4}$ isn’t all that mysterious. You’re rusty.
@user1294729 here's a nicer looking image done with mathematica:
Much nicer!
21:18
there's something going weird when i save it as an image, so it's not perfect
lemme just do the screenshot version instead
nicer
the trouble with such maps is that anything near the z=-i pole tends to stetch the plot out a lot
kinda looks like the wu tang symbol
as it should, but it just maps the pictures a bit annoying
" Da Mysteries of Mathematica Plottin' "
project ya neck
replacing the square grid with a polar grid is the simplest tho:
21:25
do all field extensions $E/F$ induce an $F$-module structure on $E$?
Not if we want to understand the half-plane only.
that’s called a vector space
Oooo, Circles of Apollonius
oh modules are vector spaces when they're built on a field
cool thanks
shin: yes. one way to think of this is just taking the field structure of E and 'forgetting' that you can also multiply elements of E by elements of E that aren't elements of F.
noted, ty
21:30
or, maybe not forgetting that entirely, but no longer thinking of it as "scalar multiplication." you might still want to think of it as an F-linear map.
alrighto
21:59
going into number theory
3
Q: Representation of integers by $A = B^{\frac{5}{4}} + C^{\frac{5}{4}}$

mickLet $n>-1$ be an integer and $f(n)$ be the integer part of $n^{\frac{5}{4}}$. And to avoid confusion define $f(0)=0,f(1)=1$ and all $f(n)$ as the integer part of the positive real values for $n^{\frac{5}{4}}$. ( no complex or negative branches for the power ) Conjecture : Every positive integer c...

any ideas ?
 
1 hour later…
23:14
Got an earful in the nurse's office because I stay up late during the night and always have my eyes glued to a screen
23:33
@D.C.theIII Turning into me is full of disadvantages. Not worth it just to 'master' a subset of a particular topic.
$F$ some field and $p \in F[a_0, \cdots, a_n]$. is $p = \sum_{i_0, \cdots, i_n} f_{i_0+\cdots+i_n}a_0^{i_0}\cdots a_n^{i_n}$ an adequate description of $p$?
$f_i \in F$
shin: your definition of F[a_0, ..., a_n] should tell you that. if it is one of the standard ones (which is not me saying that it's OK for you to not have a definition in mind), in general, the coefficient of the "monomial" a_0^{i_0} ... a_n^{i_n} might depend on all of the indices i_0, ..., i_n individually, and would not depend only on the sum i_0 + ... + i_n.
you might also clarify what those indices i_0, ..., i_n, are ranging over.
e.g. if F = Q and n = 1 your current description of "p" for polynomials having monomials up to total degree 1 looks like "f_0 + f_1 a_0 + f_1 a_1" which does not capture the situation where the coefficients of a_0 and a_1 are not the same rational number.
oh no
you're right
people sometimes introduce 'multi-index' notation for stuff like this, e.g. think of, and write, both the coefficient c_{i_0, ..., i_n} =: c_I and the monomial a_0^{i_0} ... a_n^{i_n} =: a^I as functions of the "multi-index" I = (i_0,...,i_n) and write sum_I c_I a^I
23:53
i don't know if my algebra texts are too noobish but i'm not given an explicit definition of $F[\cdots]$ when there is more than one element in the brackets
will look elsewhere and see if i can find stuff on the multiindex notation
the proof i'm reading actually skips over giving an explicit characterization of the polynomial but it struck me that it should exist
this isnt just an algebra thing incidentally PDE people love them some multi-indices too
sometimes the 'list of coefficients' aspect of a polynomial is useful in a proof or argument, and sometimes it's not
its not uncommon for people to gratuitously name coefficients when it turns out that maybe the notation for all of them isn't needed or helpful

« first day (4669 days earlier)      last day (647 days later) »