« first day (4760 days earlier)      last day (557 days later) » 
03:00 - 18:0018:00 - 00:00

03:59
@Jakobian Something my master's advisor told me, when I was in my late 20s, after I had declared a distaste for applied mathematics: "You don't know enough applied mathematics to have an opinion. Until you have had worked through some actual papers in the field, you need to keep your opinions to yourself."
@XanderHenderson this is a good attitude
@Semiclassical Totally agree. He and I were very different people, and I didn't always get along with him, but I learned so much from him.
in general i feel the modern world (social media etc) seems to default to "you need to have an opinion on X"
sometimes that makes sense---opinions on certain issues indicating your values etc
One of the best courses I taught in my long career was a year-long applied math course, based on Strang’s (then) brand-new book. I learned so much (having never taken such a course).
I have not read any of his books, but he came to UCR while I was there, and gave a couple of really good talks (a talk intended mostly for undergrads on singular value decomposition which really helped me to see a point to some of the undergraduate linear algebra that I never saw much value in), and a higher level expositional talk on some of his work.
04:09
SVD is underrated in physics too
save for quantum information theory where it's absolutely central
The higher level talk mostly went over my head (it was more for the applied math folk, and I don't speak that language very well), but the talk on SVD was fan-f'king-tastic.
(albeit in a slightly different form)
And just having said that one should study more before having an opinion:
These kinds of questions totally puzzled me when I first studied them in abstract algebra. "Why exactly this definition and not this other one which is almost the same one?" It may take a while until you get to category theory but the definition of a field in categorical terms is very very succinct: "Ab-enriched abelian group". Both definitions of Ab-enriched and of abelian group we can agree are very "natural". So with this definition it feels more like math gave us the definition and not that we had to make a bunch of arbitrary choices to get to the definition we wanted. — Julián 12 mins ago
Spoken like a true category theorist.
My impression of a category theorist: "That very complicated object you are studying? Yeah, it's just an object in the category of Foos, where the morphisms are Bars."
"feels more" seems like the key weaseling there
4
Q: Show that $f'(x)=0$ for all $x$ implies $f$ is constant.

Powder Problem Let $f:I\rightarrow \mathbb{R}$ be differentiable in an arbitrary interval I. If $f'(x)=0 $ for all $x\in I$, then $f$ is constant. My idea If $x \in \operatorname{int}(I)$, we know that $\exists$ $n_0\ge0$ such that $x\in [x-\frac{1}{n_0},x+\frac{1}{n_0}]$ then by the Intermediate...

In the first answer by José Carlos Santos I think he used Lagrange's Mean Value Theorem(LMVT). But how did he know, that the function $f(x)$ is continuous on $(a,b)$ which is a necessary condition for using LMVT ?
04:15
Category theory seems interesting. Most category theorists seem like arrogant asses. But maybe I just hang out with the wrong crowds...
I was very proud that Strang agreed to teach one of the recitations when I lectured the 350-person multivariable calc class at MIT.
@TedShifrin Wow. That's a real feather in the cap.
@TedShifrin oh dang
I genuinely love that man. One of the best mathematical expositors I've ever had the pleasure to hear speak.
We gave him the students who had struggled most in first semester. I don’t remember quite how we did that.
04:16
@XanderHenderson i can respect category theory as a way of licensing the silly diagrams i draw for quantum computation
I find some of his teaching/writing a bit annoying.
But, yeah, a good expositor.
i think he just finally retired. Must be well into his 90s.
Strangely, the answer got 6 upvotes. Did I miss something?
but honestly i've never been able to read much of any category theory
@TedShifrin 88 is what a quick Google search indicated.
Wow, only 18 yrs older than me. I thought more.
04:20
Oh, leave it. Indeed, I missed a minor detail.
@ThomasFinley Yes. It’s continuous on any closed interval in that interval.
So constant on any closed subinterval. Hence constant on the original interval.
04:58
0
A: Show that $f'(x)=0$ for all $x$ implies $f$ is constant.

Thomas FinleyThis can be done more easily. We have $f'(x)=\frac{df(x)}{dx}=0\implies df(x)=0dx.$ Now, integrating both sides, we have, $\int f(x)=\int 0dx\implies f(x)=C,$ where, $C$ is an arbitrary constant. This seems a one-liner.

@TedShifrin I think you should check this out. Isn't it more trivial ?
Others may give their opinions as well.
mm, no. your solution implicitly uses the connectedness of the domain, when a likely goal of an exercise like this is to make the use of the hypothesis on the domain more explicit.
example: if f(x) = -1 if x < 0 and f(x) = 1 if x > 0, then f is differentiable {x in R: x is nonzero} and f'(x) = 0 for all x in this domain, but f is not constant.
05:14
@leslietownes but that's never mentioned explicitly
a lot of the "+ C's" that pop out of formal methods of solving ODEs implicitly use the connectedness of the domain. this sometimes leads to confusion. e.g. if f(x) is the function just given then f(x) + ln(x) is an antiderivative of 1/x, but does not have the form "1/x + C."
i'm not going to debate this with you, you asked for a comment and you got one.
@leslietownes Sure, I got a comment. But I couldn't agree with you completely. Well, let's see what others feel about this.
you might get more responses to your questions if you didn't follow-up with everybody who responds, demanding more, more, more, more.
please tag me another five times.
@ThomasFinley No. The proof that antidifferentiation results in $+C$ is indeed the identical MVT proof. This argument totally begs the question.
@leslietownes Here’s one. Munchkin can do 4 more.
ted: very tempted to ask another question involving the inverse function theorem. to add to your parade of examples.
05:27
Orangeskid wrote the answer copying mine :)
someone should have. it is a tragedy to leave good math in the comments.
@TedShifrin i finally found a rigorous statement of the section-projection duality for convex bodies
Proposition 2.6 from here:
"Let $K$ be a convex body in $\mathbb{R}^n$, and let $E \subseteq \mathbb{R}^n$ be a subspace. Let $P_E$ denote the orthogonal projection in $\mathbb{R}^n$ onto $E$. Then: 1. $(K \cap E)^\circ = P_E(K^\circ)$; 2. $(P_E(K))^\circ = K^\circ \cap E.$

In both parts, the operation of taking the polar set in the left hand side is considered in the subspace $E$."
@leslietownes I had already written that proof four times or more.
@Semiclassical Cool!
here $X^\circ$ is the polar set of $X$
next step: figure out how to actually apply it to my case of interest
the nice thing is that my convex bodies all contain the origin, so the duality is an involution
@leslietownes How about good insults in the comments? I just left a few.
05:38
so fewer annoying comments to make
@leslietownes It's true, that I do tagging , but that's only to notify the interlocutor about my response else they might miss it. If you think that's annoying for you, you might want to block me (there's an option in here) or stop responding to my comments here, and that'll really solve the issues amicably.
ted: i like the idea that there are more solutions to this question from 2019 yet to be posted.
@ThomasFinley leslie is politely suggesting that you might not want to wear out your welcome quite so quickly.
@leslietownes I thought you would. You should leave comments more insulting than mine.
@TedShifrin maybe. I just gave him a suggestion, in the most polite manner in all probability. But let's drop the topic, anyways.
But can you explain what do you mean by "This argument totally begs the question " ?
05:46
i think it's generally true that a lot of people would comment more on MSE, if people who posted on MSE did a marginally better job of attempting to digest the comments that they do receive.
the taylor series idea is a good one but i think there are even bigger theorems that could, without too much contortion, also prove the same result under even stronger hypotheses.
so this is probably just a "it's too late in the evening to think correctly" qualm but
@leslietownes Except for a yuge gap.
is the operation $(x,y,z,w)\to(x,w,(y+z)/2)$ (from $\mathbb{R}^4\to\mathbb{R}^3$) properly speaking a projection?
@ThomasFinley To prove the $+C$ “fact” is precisely this result/proof.
my impulse is that it's "really" $(x,y,z,w)\to(x,(y+z)/2,(y+z)/2,w)$
05:53
You have to define it with codomain $\Bbb R^4$ and compute the square.
and then we're just adopting a more useful coordinate system
@TedShifrin yeah, it's the codomain being written as $\mathbb{R}^3$ that seemed off
(if i ignore $x,w$ it's just the projection matrix $\begin{pmatrix} 1/2 & 1/2\\1/2 & 1/2\end{pmatrix}$, so yeah. no issue there)
i lost 60 rep today for serial voting whatever that it. i'll never get my jump suit.
i even got dinged on travel.se for this serial voting. not quite sure how, given that i have only 2 answers there. weird
I lost 30 yesterday. And I have yet to receive anything. I suspect jumpsuits are a thing of the distant past.
copper: what's that? does it mean some 'suspicious' account was upvoting your posts en masse?
because i distinctly recall setting my bot to downvote your posts en masse
The account needn’t be suspicious. Only its activity.
06:05
sounds like something stalin would say
Or Comrade Tsar Donnie?
 
1 hour later…
07:09
so i'm not crazy, i thought I lost 5 but didn't pay enough attention to be sure
this represents a loss of nearly 33% of my rep, i could sue them for this
07:20
is there any methodical way to construct a differential equation with several solutions that satisfy some set of initial conditions? (e.g., say you have four initial conditions that you want four solutions to your differential equation to satisfy - how could you construct such a differential equation?)
07:38
Hi :)
08:11
o/
08:43
auden: at that level of generality it is difficult to say. what do you mean by 'initial conditions' (e.g. what is it that a function is supposed to satisfy, other than the differential equation)? do you have any restrictions at all on the form of the differential equation (and if not, why not, i.e. what is motivating the question)?
08:56
OK, this is kind of funny. both most of the setup of the question and the attempt are large images. math.stackexchange.com/questions/4753474/…
what is happening there
i think the question is that they don't see how "2.19" follows from the integral formula, but i can't be sure because they were too busy posting images to ask.
anyway, i'd never seen that before. that's my new favorite post.
they're a discoverability nightmare and an accessibility problem and all sorts of other things, but image posts are sometimes also funny.
my favorite subgenres are "image with tons of whitespace in it for no reason," "screenshot where the math thing is only one of many visible things on a cluttered desktop," and "'screenshot' that is a photo of the screen of another device."
if you combine all of those, you win forever, but only if your attempt is also an image of the same kind.
09:12
I have done the last one in the chat :)
@leslietownes yeah, i'm looking at finding an f(t,y) such that y'' = f(t,y) has solutions corresponding to four different initial conditions
at auden: hmm what is the motivating problem?
a lot of differential equations have solutions for any possible initial conditions
auden: is an "initial condition" a specification of a pair of values y(t_0), y'(t_0) [for a particular and fixed value of t_0] or something else? given any numbers a, b for example, something like a cos(t) + b sin(t) solves y'' = -y with y(0) = a and y'(0) = b. and as shin notes this is a special case of a more general theorem when this kind of "initial condition" is involved.
the motivating problem is:
consider initial condition A where y(0) = 0, y'(0) = 0; initial condition B where y(0)=1, y'(0)=0; initial condition C where y(0)=0, y'(0)=1; and initial condition D where y(0)=1, y'(0)=1. find a function f(t,y) such that the solution to y'' = f(t,y) gives y(1)=0 for initial condition D and y(1)=1 for initial conditions A, B, and C.
i've never encountered a problem like this, and frankly i'm not sure how to approach it beyond guess and check (which so far hasn't been very fruitful).
09:31
@leslietownes My all-time favourite on MSE was a photo of a problem on an exam paper, which included the OP's real name. Apart from the image, there wasn't much text in the question body, apart from a desperate plea for a quick answer.
@copper.hat Someone upvoted a bunch of your posts, and those votes got reversed. That happens automatically if someone gives you too many votes on one day. But votes can also get reversed manually when they've been cast by members of a voting ring. Such rings primarily upvote each other's posts, but they also upvote random posts of other people, as a form of camouflage.
 
1 hour later…
10:45
@PM2Ring sounds like reputation mafia
11:03
@Jakobian There's a black market in rep points, especially on SO. See meta.stackoverflow.com/q/410831/4014959
2
@PM2Ring when did this all start?
11:27
Why would you buy something you can get for free
This is ridiculous
if you would pay for rep you deserve to pay for rep
@user726941 Ages ago. Ever since SO became famous. In some places, sufficiently high SO rep can help in job interviews. OTOH, in some other places, if your SO rep is too high, they may not want you because you spend too much time on SO. ;)
oh snap, nvm that changes my opinion of rep
In some enlightened workplaces, they expect you to spend several hours of your working week contributing to the community. Doing stuff like writing SO answers, or code for Open Source projects.
A smart IT employer should look at your SO rep. But the quality of your answers is more important than the raw rep number, because it's a better indication of your coding skill, and your communication skill.
11:42
bastards always trying to slip their sticky tendrils into the private sphere
makes sense, it's the smart thing to do
That's the altruistic approach to creating a global community.
It would be pretty weird for a professional coder to have never used SO. But that doesn't mean they need an account. I was reading SO answers for several years before I decided to join and start contributing.
@user726941 eh, it privatizes the criteria for what should be a global community
more problematic even is that standards of community would be contribution to a private community such as stackexchange. already a pain to make the execs bend to the will of the community
They're starting an OverflowAI.
They'll have chatrooms full of bots chatting with each other to draw in real people.
They're probably "learning" off our transcripts as we speak.
12:12
@PM2Ring can someone create a second account and upvote all his answers? Just asking for a friend
@SineoftheTime You are allowed to have multiple accounts. But they must not upvote each other, or interact in any way that awards benefit to each other. Really, it's best for them to avoid all interaction. Eg, you could use one account to ask a question, and another to answer it. But plenty of people would be suspicious of that, even if the two accounts didn't vote for each other, etc.
If you did create such a "sock-puppet" account that upvoted your main account, it's likely that the mods would eventually notice. Then the fraudulent votes would be invalidated, the sock account deleted, and the main account would be suspended.
12:34
0
Q: Round the following number to $4$ significant figures: $794834.$

Thomas FinleyRound the following number to $4$ significant figures: $794834.$ For the number $794834$, I simply observed that all the digits in here are significant. The $4th$ significant digit in this case, is $8$ and the $5th$ digit is $3$. Now, as $3\lt 5$ so, we can represent the given number as $7948\tim...

Need a little help with this.
794800
@user726941 Thanks! I would suggest you to look at the whole post tho.
I did, and it looks good 👍
@user726941 Oh! Thanks a ton!
12:46
$7949\cdot 10^2$ is not scientific notation.
Yeah, you need to move the decimal point.
Depending on the scientist :P
In standards scientific notation, you write $x \cdot 10^n$, where $x$ is some number between 1 and 10 (specifically, $x \in [1, 10)$---$x$ could be $1$, but not $10$), and $n$ is an integer..
So $794900 = 7.949 \cdot 10^5$ in scientific notation.
^@ThomasFinley
@XanderHenderson ok...
Noted
Engineers like to use a variation, known as engineering notation, where the exponent is a multiple of 3, to make it easier to work with the standard SI prefixes, milli, kilo, etc. However, engineering notation doesn't unambiguously express the number of significant figures unless there are at least 3 significant figures.
12:59
Ew...
I asked a question about the scientific notation of zero, but I wouldn't suggest a beginner read it until they feel very comfortable with the usual rules.
7
Q: What is the Scientific Notation of Zero?

user 726941This question was asked here, where the answer uses this description. The last line reads: "The special case of $0$ does not have a unique representation in scientific notation, i.e., $0=0×10^0=0×10^1=..."$ My question is the value of $a$ cannot be $0$ since, as they state: $a$ is a $\color{blue}...

@user726941 The things is that zero doesn't exist in the "real world". Nothing is ever really zero. So not having a way of expressing zero in scientific notation isn't really a big deal.
You just write $0$, and have done with it.
'Cause ain't noone got time for dat.
@XanderHenderson , @user726941 and @PM2Ring But how do you round off 630 to 4 sig figures ?
Should it be, $6300\times 10^{-1}$ ?
@ThomasFinley If that number actually had four sig figs (i.e. if your measuring tool is actually that powerful) you would write it as $630.0$ which, in scientific notation, is $6.300 \cdot 10^{2}$.
Note that you have not "rounded off" $630$ to four sig figs. There is no rounding involved, since the number you started with doesn't seem to have four sig figs.
@XanderHenderson so, what'll be ur answer? $6.300\cdot 10^2$ ?
13:10
Correct 💯✓
@user726941 which one ? $6.300\cdot 10^2$( as Xander Henderson) suggested or $6300\times 10^{-1}$?
Work on quickly moving the decimal around with numbers. The answer is as Xander suggested.
@ThomasFinley It depends on how, precisely, you are defining "scientific notation", and what your goal is.
Something I have noticed about your questions in here is that you seem to be locking into the extraordinarily prescriptivist mindset---you seem to believe that there is One Right Answer™. That is not typically how mathematics works, and the things that you are asking about aren't really mathematics (they are more closely aligned with the sciences).
You should be thinking more about the goal of all of this. Why do we care about sig figs? What is the point of scientific notation?
If you understand why these things are, you have a better chance of understanding what it is that you are supposed to do.
In this case, the why is all about communication. The notation exists to clearly and concisely communicate how precise the tools used to make measurements are. So $6300\cdot 10^{-1}$ is not wrong (since it expresses the same number as $630.0$ or $6.300\cdot 10^2$),...
...but it is not the most clear way of communicating that quantity in most contexts, because it is not the expected expression (which will generally be a number of the form $x \cdot 10^{n}$, where $x \in [1,10)$ and $n\in\mathbb{Z}$).
This is a topic covered in middle school, no?
Algebra 1 for the US.
Have you tried Khan academy @ThomasFinley?
@XanderHenderson this is a deep philosophical question, because they have plenty of time for that :-)
13:40
"It's unclear to me what the Mellin transform of the sum over that spectrum of the family of metrics means geometrically. Is there any nice geometrical interpretation of this? Does a geometric interpretation extend to the critical strip?" -Geocalc
@XanderHenderson (following up about your comments yesterday about Weyl-Berry-Lapidus)
"I believe that Michel would say that it measures the "geometric oscillation" of the set. Very, very roughly speaking, this is related to the roughness, periodicity, and scaling of the boundary."-Xander
So the boundary in my case is $\Gamma= \lbrace xy=1 \cup +x,+y \rbrace$
@XanderHenderson I 'll try to keep them in mind. Thanks for your suggestions! I liked them.
that is $xy=1$ along with the positive $x-$axis and positive $y-$axis.
that is $\Gamma$ precisely because the space of metrics are bounded (tuning the parameter) between $xy=1$ and $+x,+y$. This is of course an unbounded region and Lapidus seems to require open bounded regions of $\Bbb R^n$
@geocalc33 Yes, boundedness is a major hypothesis.
and that brings me directly back to Berry-Keating lol
Openness is, ultimately, less important. But his theory doesn't handle unbounded sets.
13:50
where "bounding the space is an outstanding problem"
so it seems boundedness is a very important part of this
Yarp. The tools just don't work well in unbounded cases.
Michel introduced an object he calls a "relative fractal drum" which, I think, is kind of intended to deal with unbounded sets, as well as some other pathologies.
I know he talks about those in FGCD... let me see if I can pull up a more easily accessible reference.
Oh, doi... his paper with Goran and Darko: arxiv.org/abs/1603.00946 .
Thanks
No idea if it will be helpful, but happy to share the citations.
@XanderHenderson my opinion remains the same
14:05
@XanderHenderson Yeah - and like you said Lapidus works with the wave equation, my framework is based on the heat equation so I think something more aligned to en.wikipedia.org/wiki/De_Bruijn%E2%80%93Newman_constant would probably be better
My last question on this topic will be a validation: Round off the number 0.08 to four sig fig. My answer: 0.08000*10^{4}
Did I get this right?
Is $0.08 = 0.08000 \cdot 10^{4}$?
And is $0.08000\cdot 10^{4}$ of the form $x \cdot 10^n$ for $x \in [1,10)$ and $n\in\mathbb{Z}$?
@XanderHenderson Ugh..I am so sorry for the typo. I wrote out an extra power of 10. My answer is: 0.08000
Also, I really dislike the instruction "round off". The number $0.08$ has, at best, three sig figs. Personally, I would assume that it has only one. When you round a number, you reduce the precision of the number, which means that you are taking a number with a lot of digits, and reducing it to some specified number of digits which are significant.
You can't round a number with one sig fig to a number with four sig figs.
@XanderHenderson Now, that was the instruction we were given
14:14
Perhaps "Express the number $0.08$ with four sig figs" would be a better instruction.
@XanderHenderson Might be
It's just a badly written problem.
All in all, my answer is 0.08000. This has 4 sig nos.
@ThomasFinley Is $0.08000$ of the form $x \cdot 10^n$ for $x \in [1,10)$ and $n\in\mathbb{Z}$?
@XanderHenderson yes
14:15
No, it isn't.
Oh! U wrote [1,10)
But wait
I saw you mention this rule(?) earlier, but I am not much confident in using it. Can I ask, a reference?
My ref is Scarborough 's book
@ThomasFinley Have you tried Googling "scientific notation"?
The name is, Numerical Mathematical Analysis
@XanderHenderson yeah, but it wasn't much helpful
The things were badly scrambled
I am in need of a good book that gives all these detailed instructions
above is perfect explanation on scientific notation
you might not find anything on scientific notation in college books
I like the aesthetic of that site
14:19
mathisfun is great for all basic stuff
@ThomasFinley According to the Google, that book is from 1930. I would suggest you find a more current reference. We think about things very differently now. Computers and the space race have had an effect on the way we communicate ideas.
en.wikipedia.org/wiki/Scientific_notation (see the bit on "normalized scientific notation"---my guess is that most people mean this when they say "scientific notation").
@XanderHenderson That is indeed a point to be noted, but unfortunately our professor just asks us to go through it. He talks high about the book. I mean, ok, the book is good and all, but yes, just as you mention, it's a really old one
@XanderHenderson ok, I definitely take a look.
Basically, Google "scientific notation", and look for sites with a .edu TLD.
And, again, think. What is the goal? WHY would anyone want to express a number in this manner? What are they trying to communicate?
Use that thinky stuff between your ears.
14:25
@XanderHenderson Ha ha, I genuinely like your explanations a lot. I wish I could have attended one or two lectures under you. Your students are lucky! Best wishes from my end.
@shintuku Thanks, I really found that simple. It matches completely with Xander Henderson 's perspective, which seems to be what the world uses now, unlike the 30s...
i often check mathisfun when i need refreshing on the basics
14:42
okay so this is my proof of continuity of roots of polynomial
Let $x\sim y$ iff $x = sy$ for some $s\in S_n$ and $q:\mathbb{C}^n\to\mathbb{C}^n/\sim$ be quotient map. Define $$d(q(x), q(y)) = \min\{d(x', y') : x'\sim x, y'\sim y\}$$ then $d$ is a metric on $\mathbb{C}^n/\sim$. Since $q^{-1}q(x+B_r)$ is open, it follows that $q(x+B_r)$ for $x\in\mathbb{C}^n, r > 0$ is a basis for $\mathbb{C}^n/\sim$ and clearly $q(x+B_r) = \{q(y) : d(q(x), q(y)) < r\}$ so that $\mathbb{C}^n/\sim$ is metrizable. Consider $\sigma:\mathbb{C}^n\to\mathbb{C}^n$ (continued.)
Defined as $\sigma(\lambda_1, ..., \lambda_n)$ maps to coefficients of $(z-\lambda_1)...(z-\lambda_n)$. From Viete's formulas, $\sigma$ is continuous. If $a^{(m)} = \sigma(\lambda^{(m)})\to \sigma(\lambda) = a$ then inequality $|\lambda_i^{(m)}|\leq \max(1, \sum_{j=1}^n |a_j^{(m)}|)$ shows that we can take a convergent subsequence $\lambda^{(m_k)}$
It follows that $\sigma$ is perfect as a continuous map between metrizable spaces such that convergence of $\sigma(\lambda^{(m)})$ implies that $\lambda^{(m)}$ has convergent subsequence
here's a question i may put on the main site. suppose i have some triangle in the plane which contains the origin. if i draw a line through the origin, then there are two sets i can construct: the orthogonal projection of the triangle onto the line, and the intersection of the triangle with the line
Thus $\sigma\circ q^{-1}:\mathbb{C}^n/\sim\to \mathbb{C}^n$ is a closed continuous bijection, so a homeomorphism
generally, the intersection will be a proper subset of the projection. but under what conditions are the two sets equal?
anyone has objections to my proof?
(i would think it depends on the choice of origin within the triangle)
14:59
@XanderHenderson After going through, all the links you attached, I have come to the conclusion, that "Round off the number 630 in 4 sig fig" makes no sense and the correct rephrasing of the problem would be, "express 630 in 4 sig fig" then, the answer would be, $6.300\times 10^2$. Similarly, if it was given to express 630 in 5 sig fig, then it would be, $6.3000\times 10^2$. Did I get this?
Similarly to express 0.08 in 4 sig fig, we can write it as, $0.08000\times 10^0.$ (from mathisfun.com)
I have a feeling that I have got it somehow...
15:37
Hi :)
in In the search of a question, 1 min ago, by Shaun
I'm looking for proof that an algebraic torus $T$, when defined as isomorphic to $\Bbb D_n$, is isomorphic to $(k^*)^n$.
I tried here
According to this search, it might not be here.
There'd be more than one proof, surely. It's a fairly standard theorem.
I could argue in terms of their dimensions . . .
D'you reckon I should ask a question on the main site about it?
Suppose that W is a pointed topological space. How can one show that [X,W]_* is associative? [X,W]_*= the set of homotopy classes of based continuous maps from X to W.
16:07
0
A: Continuity of the roots of a polynomial in terms of its coefficients

JakobianHere's a proof without using complex analysis, differential geometry or the projective plane, and only using results from general topology. Consider the equivalence relation on $\mathbb{C}^n$ given by $x\sim y$ iff $y = (x_{s(1)}, ..., x_{s(n)})$ for some $s\in S_n$. Give $\mathbb{C}^n/\sim$ the ...

0
Q: Defining a torus $T$ a certain way, prove that $T\cong (k^*)^n$.

ShaunThe Question: Suppose we are given the definition that a torus $T$ of a linear algebraic group over $k$ is a subgroup isomorphic to $\Bbb D_n$, show that $$T\cong(k^*)^m$$ for some $m\in\Bbb N$. Context: I have looked in Springer, Borel, and Humphreys; a proof is nowhere obvious in them. The text

16:22
@Shaun It is tautology.
What is an invertible diagonal matrix?
wait. No it doesn't
16:40
@TedShifrin It's essentially a collection of $n$ nonzero elements of $k$. Sigh. I really need to stop overthinking this stuff. Everything seems so much harder to me than it really is.
Thank you.
Yeah, I think you psyche yourself out sometimes :)
2
What I really need to do is solve more problems in the area to boost my confidence in my intuition.
Yes, working problems and interesting examples is essential for learning every part of math. This is why good exercises are for me the most important part of a good text. Sadly, many authors who write grad texts are too lazy to write exercises — verify the details if such and such a proof is not a good exercise.
this is also why it's important to start trying to pose problems to yourself
I mean, it’s a starting point, but one needs more creativity.
Yeah, that is hard for lots of students, Semiclassic. Even faculty are sometimes bad at generating new problems.
16:46
obviously you won't have as much insight into what a "good problem" is as an expert, but ultimately you can't rely on the author doing that for you
17:02
does someone mind checking what I wrote for correctedness
@PM2Ring I understand the general idea. But on the travel.se site, I only have 2 answers and no upvotes in a long while, so the serial upvote thing is a bit of surprise. Similar situation on mse.
17:19
@leslie is haunting you, @copper.
Did you know about the rise of Stack Overflow's rep black market?
@TedShifrin If they were downvotes I would understand, but I can't imagine Leslie serially upvoting me :-)
Yes, you make an excellent point. Nor me.
They're probably using leslie-coin.
One can only get so far with counterfeit.
17:24
@TedShifrin I'd add the requirement that the exercises are thoughtfully chosen. When I was reading the book "Topological groups and related structures" by Arhangelski and Tkachenko, I didn't feel like the exercises were chosen very thoughfully, Alex Ravsky even told me once that some of those exercises were or are open problems
Is getting your reputation points through the black market better than having a black reputation.
oh I guess this counts as "good"
including open problems isn't necessarily bad, if you're up front about it
@Semiclassical not if they're hidden as normal problems
if you're up front about it
putting them among normal problems is not being up front about it
17:27
the interesting thing with that book is that they include both open problems and normal exercises
so them including open problems in exercises is more of a result of being sloppy and negligent
they even include a theorem that's false there
Makes learning much harder.
@Jakobian as part of the text, or as an exercise?
as a part of the text, yes
17:29
wtf
i mean, books do have errata for a reason
so that's not unheard of
But doing it on purpose?
things like these is why I'm paranoid about correctedness of everything
@user726941 "it's just a prank bro"
17:31
>8(
@user726941 I don't think it was on purpose, I think they were, again, very sloppy about it
Time to find a new book "bro."
@Jakobian I don't disagree. But putting it as a last problem with a symbol that indicates "challenging, perhaps heretofore unsolved" is OK with me.
you see that with journal problems occasionally
On the other hand, overdoing labeling with asterisks or double asterisks (indicating harder or harder yet) I've learned from experience tends to dissuade students from attempting problems many of them would in fact solve.
So I didn't do this in my books. (Sometimes I put a parenthetic "challenging" at the beginning of a particularly hard problem.)
17:34
this is where an instructor's role of assigning problems is important
I think Concrete Mathematics does that too
I have an actual physical copy of that book
@Semiclassic Most instructors don't pay much attention. They just look casually and say, "Oh, that looks like a good one to assign."
Since I traditionally wrote a lot of my own exercises over the years, I had to think a bit harder than that ... especially teaching stuff that was new to me (like the applied math class or probability).
You, sir, have put yourself in the shoes of the first time learner.
17:45
I'm sure @D.C.theIII will offer plenty of comments on my choices of exercises in a certain text :)
i've been in the intro physics trenches long enough that i have "opinions" about certain points
e.g. "centripedal force" should be excised from all intro books
petal or pedal?
It's centrifugal that's fictitious, so your comment puzzles me.
petal, yeah
Certainly uniform circular motion is allowed to have centripetal acceleration.
So I think I totally disagree with your comment.
@TedShifrin yes, and therefore i'm fine with the phrase "centripetal acceleration"
it's "centripetal force" that i disagree with
it makes it sound like it's one of the forces that act on a body
17:52
Ah, so if there's a rope attached to said rotating object, I can't say that tension in the rope provides centripetal force?
The point is that the "centripetal" should refer to the motion of the object. So I see your point.
Fictitious forces, however, show up in everyday life.
well, it's more that a student will see "centripetal force" as being akin to other "X forces"
contact force, spring force, etc
but it's never correct to write $F_{net} = F_C+\cdots$
Yeah, I see your point. I now agree. What do you say about centrifugal?
it's probably best avoided in teaching it, because non-inertial frames are hard to deal with at the intro level
I took a sophisticated first-year course out of Kleppner/Kolenkow, and I thought they dealt with that well, but it came late in the course.
the physical 'sense' isn't exactly wrong---in a non-inertial frame you do have net force=centrifugal force + other forces
but myself i just stick to the phrase "centripetal acceleration", to emphasize that it's a condition on the object's acceleration
in the same way that an object sliding down an inclined plane has zero acceleration in the perpendicular direction, or two objects moving together have the same acceleration
@TedShifrin yeah, for some courses i think it'd suitable
i'm mostly thinking of the less-sophisticated instances of the course
17:58
Yeah, I have withdrawn my objection.
i have seen the phrase "centripetal force condition", and I see what it's trying to do
but ultimately i prefer my way :P
03:00 - 18:0018:00 - 00:00

« first day (4760 days earlier)      last day (557 days later) »