« first day (4730 days earlier)      last day (586 days later) » 

00:01
You’re just a yak man, @Xander.
@TedShifrin Yakkity yak.
(Don't talk back.)
00:28
Back into your rucksack.
00:54
$$\sum_I f(r_i)\alpha_{1\cdots n}^{i_{1\cdots n}} + \sum_J f(r_j)\alpha_{1 \cdots n}^{j_{1\cdots n}}= \sum_{(I,\mathbf 0) \cup (\mathbf 0, J)} f(r_t)\alpha_{1\cdots n}^{t_{1 \cdots n}}$$
right?
$I,J$ are multiindexes, $(I,\mathbf 0)$ is the multi index like $I$ which has zero entries afterwards, and $(\mathbf 0,J)$ is the multi index like $J$ which has zero entries before its first nonzero entry
so what I mean is that $(I,\mathbf 0) \cup (\mathbf 0, J)$ counts through all of $I$, then counts through all of $J$, but doesn't do both at once
i.e., in $(I, \mathbf 0) \cup (\mathbf 0, J)$, there is no entry $(x_1, \cdots, x_n, y_1, \cdots, y_n)$, only $(x_1, \cdots, x_n, 0, \cdots, 0)$ or $(0, \cdots, 0, y_1, \cdots, y_n)$
maybe there's some more natural notation to say this?
01:16
forget the above, made a mistake in setting up the multi indexes
proper setup would have been $\sum_{(I, \mathbf 0)} + \sum_{(\mathbf 0,I)} = \sum_{(I, \mathbf 0)\cup(\mathbf 0, I)}$
01:37
whole idea is inconsistent, starting from zero heheh
I had cheese added to my ramen. It tastes amazing, and the texture is so much better. Did I discover something new
01:52
Yuck. What kind of cheese?
Ramen with soyish broth is not like Italian pasta with parmigiano reggiano.
It didn’t have any broth it was mostly dry
02:09
Not what I’m used to for ramen. What cheese?
hades: i've done that with parmesan
And sushi with pecorino romano.
02:37
make your favorite brand of chicken ramen and throw a handful of white cheddar cheez-its in there.
it's an old ted shifrin family recipe.
03:33
Barf.
 
7 hours later…
10:22
@TedShifrin just shredded mozzarella
11:26
Both will be under a lot of pressure while fighting.
16 hours ago, by Ted Shifrin
@sunny That is a total mess. You need to say this: Suppose $f_n$ are continuous and $f_n\to f$. If $f_n\to f$ uniformly, then $f$ is continuous. This is the statement which you want the contrapositive of — the opening hypotheses remain outside.
Hey @TedShifrin, could you just clarify if by your first occurrence of $f_n\to f$, you mean $f_n\to f$ pointwise, right?
 
3 hours later…
14:07
I
would like someone to look at my purported proof
it's always good to get a second set of eyes on anything
if you're available to look it over respond with "I'm ready to look it over"
nevermind
I found a crucial flaw in the argument.
15:10
I'm afraid I wouldn't know what you'd be talking about
:)
15:29
I think you would actually. It's a very straightforward "proof." But it's wrong of course.
just can't figure out what's wrong
you sure? I don't have any experience with PDE or manifolds. I avoid all things differential
Anyone know anything about elliptic surfaces?
I'm trying to understand the Dynkin diagrams associated to singular fibres
Specifically I do not understand why some irreducible components get a factor of 2/3/4/5/6 in front of them in the case of fibres of type _*
16:15
Hi :) Please would someone give some feedback on my question below? It was downvoted just now and I'd like some idea of why and of how to improve it.
0
Q: Verifying my proof that $Y\subseteq B_X(n)$ implies $B_Y(m)\subseteq B_X(mn)$.

ShaunThe Details: Let $G$ be a group and let $X\subseteq G$. Let $n\in\Bbb N$. Recall that for $H\le G$, $${\rm conj}_H(X)=\{ hxh^{-1}\mid x\in X, h\in H\}.$$ Define $$B_X(n)=\bigcup_{k=0}^n\underbrace{{\rm conj}_G(X)^{\pm 1}\dots{\rm conj}_G(X)^{\pm 1}}_{k\text{ times.}} $$ Note that the idea here is...

In an infinite dimensional Hilbert space can 0 be an eigenvalue of a compact operator?
It's the second downvote of the question. I don't get either of them.
I think I did all I could to make it a high quality question. If you have any suggestions on how to make it a better one, please let me know.
@PNDas why not? The operator $T = 0$ is like that
Ah non-zero compact operator. Sorry
It doesn't really matter if it's zero or not, what matters is that this is a projection
just consider any projection $l^2\to l^2$ of finite rank
something like $(x_1, x_2, ....)\mapsto (x_1, x_2, ..., x_m, 0, 0, ...)$
16:34
A statement says: "Consequently, the theory of compact self adjoint operators implies that all the eigenvalues of S are real, positive and..."
They have proved that $S$ is compact, self adjoint and $\langle Sf,f\rangle\geq0$
I don't understand how they get "positive".
I thought self-adjoint gives real. The inequality says that eigenvalues are non-negative.
So compact must be giving the positiveness.
$S = 0$ satisfied all of those
perhaps they are using the French definition of positive
as in non-negative
yeah I think positive here means $\geq 0$
But the author wrote: $0<\lambda_1\leq\lambda_2\leq\cdots$
did they wrote that? I didn't see it anywhere
I got it. The author wrote $\langle Sf,f\rangle\geq0$ but in that theorem, we actually get $\langle Sf,f\rangle>0$
@Jakobian The author of the book I was reading.
I see. It's hard to help remotely like that
16:46
But thank you for the help.
17:21
Is a function from $\mathbb{N}$ to $\mathcal{F}$, where $\mathcal{F}$ is the set of all continuous, real-valued functions, a well-defined function? In other words, a function that maps a natural number to a continuous-real valued function.
The reason for the question is we can denote a sequence of real numbers $(x_1,x_2,\ldots, x_n,\ldots)$ as a function $f:\mathbb{N}\to \mathbb{R}$ belonging to the set $\mathbb{R}^\mathbb{N}$ (this last notation is simply the generalized Cartesian product, i.e. $\mathbb{R}^\mathbb{N}=\{f:\mathbb{N}\to\mathbb{R}\}$). Now, suppose I have a sequence of continuous, real-valued functions $(f_1,f_2,\ldots,f_n,\ldots)$. How do we denote the set of all those sequences? Is it $\mathcal{F}^\mathbb{N}$?
@sunny yeah why not
@shintuku cool :) I was a little unsure
well defined here depends on the explicit definition however
and also it will just not be a bijection of any sort
e.g. $n \mapsto n$, $n \mapsto x+n$ are two such functions @sunny
ok
but uh it's probably a more serious consideration to check whether the set of all sequences of real functions is actually a set
not a trivial question, it sounds like a pretty big set
you can probably ignore it by posing your question adequately, if you're working with pointwise/uniform convergences
17:38
hmm, ok. I should have probably added a domain of all the functions, like $\mathcal{F}(A)$, where $A$ is the domain of the cont., real-valued functions
@shintuku rephrasing: you most likely can and should ignore it by posing your question appropriately
18:13
@sunny wdym by continuous real-valued functions?
But to answer your question, $X^\mathbb{N}$ is the set of all sequences of elements of $X$ i.e. $(x_1, x_2, ...)$ where $x_i\in X$. Formally such sequence is a function $f:\mathbb{N}\to X$
there is no disctinction between a sequence $(x_1, x_2, ...)$ and a function $f:\mathbb{N}\to X$ defined by $f(i) = x_i$. They are literally the same
For finite tuples when introducing set theory, we define a pair $(a, b)$ for example somewhat differently, but ultimately once we have functions we can again say that those are just functions $f:\{0, 1\}\to X$
@Jakobian let $C \subseteq \mathbb R$. suppose by continuous real valued function we mean a function $f:C \to \mathbb R$ s.t. for all $c \in \mathbb R$ we have $\lim_{x\to c}f(x) = f(c)$, is the set of all sequences of these functions a set?
wait i think I got the quantifier for $C$ at the wrong place
2 sec
That's a somewhat sketchy definition
I love chocolate
the function $f$ satisfies the proposition $P$ if there exists a set $C \subseteq \mathbb R$ such that $f:C \to \mathbb R$ and for all $c \in C$ we have $\lim_{x \to c}f(x) = f(c)$. is the set of all sequences of functions satisfying $P$ a set? at Jakobian
missing a condition on $C$, two sec
I hate chocolate. But I ate some vegan ready-made food and it was good. Rice and carrot, with some cicer and a bit of sauce of some kind.
@shintuku is this discussion necessary
I'm sure sunny just meant continuous functions on some fixed set like $\mathbb{R}$ or $[a, b]$ or other kind of interval
18:25
no it's super not necessary that's why I suggested they avoid this heheh
well, if we understand it as all real-valued continuous functions $f:X\to \mathbb{R}$ then those don't form a set
if we allow $X$ to be any topological space
yeah, i'm thinking of a decent condition on $X \subseteq \mathbb R$
well say for a simpler problem we're talking about $C[0,1]$
is the set of all possible sequences of functions in $C[0,1]$ a set?
of course
I'm just avoidant of notation $\lim_{x\to c} f(x)$ when $f$ is defined on an arbitrary subset of $\mathbb{R}$
right that makes sense
Note that $\lim_{x\to c} f(x)$ is defined by demanding that $0 < |y-c| < \delta\implies |f(y)-L|<\varepsilon$ and for arbitrary sets $C$ we can just add that $y\in C$
now say $c$ is an isolated point of $C$
18:33
yeah, need some conditions to get rid of pathological domains
then this implication will always hold no matter what $L$ is
it's not a problem with the domain, but using limits in the first place
well, limits of the form $\lim_{x\to c} f(x)$
If you were to use sequences instead nothing like this would be an issue
or just use the standard epsilon-delta definition of continuity
$|y-c|< \delta \implies |f(y)-f(c)| < \varepsilon$
@Jakobian this is a nice definition to have for calculus where you won't really consider a continuous function which won't be defined on some kind of interval etc. but maybe not the best in general
2 hours ago, by Shaun
Hi :) Please would someone give some feedback on my question below? It was downvoted just now and I'd like some idea of why and of how to improve it.
0
Q: Verifying my proof that $Y\subseteq B_X(n)$ implies $B_Y(m)\subseteq B_X(mn)$.

ShaunThe Details: Let $G$ be a group and let $X\subseteq G$. Let $n\in\Bbb N$. Recall that for $H\le G$, $${\rm conj}_H(X)=\{ hxh^{-1}\mid x\in X, h\in H\}.$$ Define $$B_X(n)=\bigcup_{k=0}^n\underbrace{{\rm conj}_G(X)^{\pm 1}\dots{\rm conj}_G(X)^{\pm 1}}_{k\text{ times.}} $$ Note that the idea here is...

didn't you already get feedback?
> Please use the space in your post wisely. It is really tiresome to have to scroll as if one is reading those recipe webpages that have unnecessary padding before one gets to the actual content. ;)
> Sorry, but again your theorem is an obvious statement with correct notations. I agree with Pedro 100%.
at Jakobian, i see. thanks for the comments
> @Shaun It does, but it is not the answer you want: what you need is to use better notation which will allow you to bypass the step where you ask us if the proof is right, since no hand waving will happen.
They seem to be criticizing your style of post, and claim it's obvious if you simplify your notation and stop hand-waving
I can see why would someone downvote your question
especially someone coming in and reading those comments
I'm of course, giving an opinion as an outsider
well, now that I had some delicious food, I probably should make some coffee
also you've got to admit that ultimately it's a pretty elementary calculation
18:54
Fried rice^^
19:04
anyone good with..metrics?
$(\Bbb R^2_{\gt 0},g)$ with $g=2 \sqrt{\frac{r}{s}}K_1(2\sqrt{r s})drds$

I'm looking to understand this space a little bit.
$r,s>0$ as well. As for $K_1$

https://mathworld.wolfram.com/ModifiedBesselFunctionoftheSecondKind.html
Thank you for the feedback, @Jakobian.
@冥王Hades amazing
19:30
actually I meant to write the product metric $$ dt^2=(2K_1(2\sqrt{r})/\sqrt{r})dr^2+(2K_1(2\sqrt{s})/\sqrt{s})ds^2 $$
$dt^2=g(r)dr^2+g(s)ds^2$
this tells me that this space has 0 curvature
iirc
$$dt^2=\frac{2K_1(2\sqrt{r})~dr^2+2K_1(2\sqrt{s})~ds^2}{\sqrt{s}}$$

I like this one though.
19:49
Correct me if I'm wrong, but any closed discrete subspace of $[0, \omega_\alpha)$ is finite, right
Since if it were infinite then letting $\alpha_1 < \alpha_2 < ...$ we'd have $\sup_n \alpha_n\in D$
well I guess I do need $\omega_\alpha$ to have uncountable cofinality
I got a fine yesterday for honking unnecessarily. I was stuck behind an awfully slow Toyota Prius and, in frustration, I honked telling him to move it.
I was fined
Even if $D$ is infinite, we can demand that $\alpha_1 = \min D$, $\alpha_n = \min D\setminus \{\alpha_1, ..., \alpha_{n-1}\}$
Then if $D\neq \{\alpha_1, ...\}$ we'd obtain that $\sup_n \alpha_n < \omega_\alpha$ which again would contradict that $D$ is closed and discrete
so $D$ needs to be at most countable
so the extent of ordinal spaces $[0, \omega_\alpha)$ is $\aleph_0$
So it's not only that $D$ is countable but also that it's order isomorphic to $\omega_0$
kind of an interesting property
Actually no this isn't really interesting, $\omega_0$ is the only infinite discrete ordinal in the first place
actually maybe it is interesting because here $D$ inherits the topology from being an ordered space
@Jakobian it must be club in that case of course. So the only discrete closed sets are countable clubs and finite sets. The former doesn't exist when $\omega_\alpha$ is of uncountable cofinality
well instead of $\omega_\alpha$ I can just put any ordinal $\alpha$
20:20
If the limit as a circle's radius approaches infinity is a straight line, what is the line's slope?
@user10478 Depends on how the limit is taken.
dunno, but note that "slope" is coordinate dependent. even the same actual, non-limit-of-a-circle, standard-issue line will have a different slope in different coordinates.
Oh, that too.
I had assumed that we were just working in a given Cartesian plane.
yeah, where the question would still be, i guess, which circles and what point.
20:23
Would it be more accurate to say an infinite radius circle is a family of lines rather than one line, given by y = Cx where C is an arbitrary constant in a differential equations sense?
Yeah I'm thinking Caresian plane.
None of what you've said makes sense to me.
it would be even more accurate not to say "infinite radius circle" at all.
if you have a family of circles with a common tangent line, that line has a well-defined slope, but maybe you wouldn't call it the slope "of an infinite family of circles"
@冥王Hades Damn right! Good!
or maybe you'd say something like, "'infinite radius circle' is just my funny term for 'line' and here perhaps are other aspects of this funny naming system that i define while leaving other aspects not defined"
@TedShifrin You don’t understand. I feel the need, the need for speed.
20:25
I also question the initial statement: "The limit as a circle's radius approaches infinity is a straight line." It is possible to construct a sequence of circle's who's limit is, in some sense, a straight line. But that is a very different statement.
If I said "infinite radius circle in a limiting sense" would that help?
@冥王Hades They should throw the book at you!
@user10478 No.
@user10478 How are you taking the limit? (which was my original comment)
user: you might be getting at the idea that there are lots of points on any given circle, and depending on which ones you consider in a family of circles, you could get different "limiting behavior" (putting in quotes as we have not given this a definition yet) of the family.
So in my mind, I'm thinking of a geometric animation, where one point on the circle remains fixed, and everything else expands outward.
20:28
There is no geometric limit. The limiting curvature is $0$, so that says line. But …
it's not particularly unusual, or circle-specific, for a limit to depend on the thing that you're taking the limit of. e.g. you could consider the x-axis as some kind of limit of the family of circles that have centers (0,R) and radius R. you could consider the y-axis as some kind of limit of the family of circles that have cneters (R,0) and radius R.
@TedShifrin They barely reach my shoulders Ted.
so, different limiting lines, and different slopes, associated with different families of circles. maybe all that suggests is that how you think of a line as a limit of circles depends on which circles you start with.
If the circles are all tangent to a fixed kine, then … you win.
So depending on which point on the circle you hold in place, you end up with a different line. Is that the idea?
20:29
You need more than a point tixed.
How many points remain fixed? It seems to me that if you fix more than one point and try to expand a circle you necessarily deform it.
Read my comment about same tangent line at the point.
Okay, I think I get the problem, if I want to truly say I'm in the Cartesian Plane I have to treat multiple shifted lines as different. So the point on the circle I fix when I expand it lets me set the slope of the line, and the position of the circle in the plane lets me set the y-intercept of the line.
But I still only fix 1 point I think?
I dislike this
wrong attitude
21:45
@Jakobian it’s not what you think it means. It’s a jab at Intel, the company, for recently manufacturing dead-on-arrival CPUs
@冥王Hades The crime is not putting meat in it, like the Italian Arancini al ragù, which I would prefer any day over the Japanese Yaki Onigiri.
22:17
wow - been 6 years already
@冥王Hades I think that something you have maybe not understood is that a large number of folk here are extremely annoyed by your childish memes, and would prefer that you just stop. This is a semi-professional, semi-focused space. Some of your communication habits might be more welcome at... I don't know... 4chan?
@XanderHenderson A simple “please stop the memes, it’s a bit annoying” would’ve done the trick.
But you’re Xander.
4chan is so 2000s
@冥王Hades I've previously asked you to knock it off. It doesn't seem to stick.
22:32
a more modern reference might have been discord
@XanderHenderson I don’t remember but had you asked again I would’ve stopped even if temporarily
2/10, unaesthetic communication, dated references
So...no memes?
no 4chan, geocalc. no 4chan
You know I can’t be mad at him, I used to be like this too as a mod in discord. Berating whoever I like. It’d be hypocritical of me to criticize something even I’ve done
22:34
@shintuku lol
it's difficult to be a person of tolerance, because when one tries to act intently in kindness, negative feedback stings a lot more
I mean, 0 memes in the chat would be very refreshing honestly
7
and then, after that memes, so we are refreshed from having no memes?
an infinite loop of memes and no memes
no memes? what's next? no jokes? and then what? no capital letters? bye bye punctuation? gee, i wonder when was the last time this happened
Sehr schrecklich.
We will reduce to nothing, a collapse, and then from nothing, we would rebuild the Mathstackexchange chat till we reach a point where people start posting memes again
23:02
Hi everyone,
I'm reading Elementary Linear Programming with Applications by B. Kolman and R. Beck. The following definition caught my attention.

Definition. Let $\boldsymbol{x}_1$ and $\boldsymbol{x}_2$ be two distinct points (or vectors) in $\mathbb{R}^n$. The line determined by $\boldsymbol{x}_1$ and $\boldsymbol{x}_2$ is the set of points
$$
\{ \boldsymbol{x} \in \mathbb{R}^n | \ \boldsymbol{x} = \lambda \boldsymbol{x}_1 + (1-\lambda) \boldsymbol{x}_2, \quad \lambda \text{ real} \}
$$

In this case, I think the vectors should be linearly independent, but I don't understand what distinc
23:16
No.
Think of the line through the heads of the two vectors.
Hi Prof. @TedShifrin
This is my problem actually, the heads.
Why a problem?
How can we draw a line between the tips if they point in the opposite direction?
I think there is no line because the vectors are linearly dependent.
right?
but again I don't understand the exact definition of distinct points in this context.
No. Distinct just means they are not the same point.
I think that thinking of these objects as "vectors" may be an unnecessary confusion. Just think of them as points.
23:27
You can draw a line through $(1,0)$ and $(-1,0)$.
@XanderHenderson exactly my point. Using vectors here confuses me.
If you have two distinct points in the plane, then you can construct a line which passes through both points. That line is the set $\{x \in \mathbb{R}^n : x = \lambda P_1 + (1-\lambda) P_2, \lambda \in \mathbb{R}\}$ (among other parameterizations).
Except I don’t know how to add points.
This is worse.
@TedShifrin Yeah, I don't really like the "adding points".
You cannot!
In affine geometry there is no sum, only difference.
23:30
What I really want to do is define a subtraction of two points to give a vector.
@TedShifrin true but is this what is meant between tips? In this case, the line passes via the tail as well.
And then consider the set of points $x + \lambda v$.
@CroCo Nothing about your definition suggests that this is impossible.
Still not defined.
@TedShifrin I can use other notation if you like, but it is entirely reasonable to define $x + v$ to be the action of $v$ on $x$, i.e. the translation of a point $x$ along $v$.
@XanderHenderson If $x_1 = -x_2$,we get $x=2\lambda x_1 + x_2$, right?
23:33
It’s fine as long as the “point” $x$ is a vector.
But the definition no longer holds
@TedShifrin No. $x$ is a member of some space, which is being acted on by elements of some group.
@XanderHenderson well, I thought this was self-deprecating hence why I commented. But it's true that this is a semi-professional space
Have you never worked with parametric equations of lines using vectors, Croco?
I won’t continue this, Xander.
I've read it long time ago, not sure if this will help me understand this definition better.
23:36
Of course. That’s what this is.
I'm happy with the points though.
or should I forget the notion of vectors here and pay attention the points.
Worst moment of my life. I unironically thought that the long line is perfectly normal
You need vectors.
It’s perfectly abnormal.
@冥王Hades I agree that Xander was a bit harsh especially since your memes aren't that bothersome and are easy to ignore. But I wouldn't compare his status as a moderator on this site to a mere discord mod.
What Xander said is true though
@Jakobian I wasn’t drawing a comparison but I don’t wanna probe this further
23:41
@TedShifrin It's interesting how a banal definition can become a hassle.
:<
It’s not a hassle. You have a lack of understanding.
This is at the beginning of every reasonable linear algebra course.
@CroCo I think you have it backwards. It is easy to think that an idea is intuitive and obvious, but when you start trying to use that intuition, it turns out that things are actually kind of complicated, and you need more precise definitions.
And that the definitions make it much, much easier to both understand and communicate.
It just takes time to learn the definitions.
Is there any thing prevents me from assuming $x_1=-x_2$ in this definition?
@CroCo No.
And if $x_1 = -x_2$, you end up with a line through the origin, which is parallel to either of the vectors you started with.
My stupid mistake is now clear to me. There is a feeling of humiliation in my heart.
23:53
@CroCo Meh. It's part of learning.
I wouldn't feel shame about it.
I doubt that anyone here is judging you.
@XanderHenderson Your words are warm and welcome.
thanks.
Huh, they ran out of chocolate stuff dorayaki
Wonder why

« first day (4730 days earlier)      last day (586 days later) »