« first day (4508 days earlier)      last day (809 days later) » 

00:03
wait what did i do lol I forgot to leave the r's
i am steaming.
moving on
for something like $\int \int_R xydA$ with $R = [0,0]\times [3,5]$ I can use fubinis theorem and it won't matter which order dx,dy I do it in?
and reversing order in the case of regions where there are no bounding functions is just simply changing the order in which you integrate?
00:44
yes.
 
2 hours later…
02:38
computing mass of a solid Q is like computing its volume except you're also multiplying the volume by the density function inside the integration
 
1 hour later…
03:59
@TedShifrin $f(z) = z+ a_2z^2+a_3z^3+\cdots$ so $f(z^n) = z^n+a_2z^{2n}+a_3z^{3n}+\cdots$. Then I just let $g(z) = z+\sum_{k=2}^\infty b_kz^k$ such that $(g(z))^n = f(z^n)$?
$g(z_1) = g(z_2)\Rightarrow z_1^n = z_2^n$ so can't guarantee the injectivity of $g$.
You have to argue that $b_n$ exist and convergence, bit yes. I don’t follow.
how do you integrate something of the form $\int\frac{y}{x^2+y^2}dy$
I tried u-sub $u=x^2+y^2$ to get $\frac{\sqrt{u-x^2}}{u}$
then $dy = \frac{du}{2} - x dx$?
LOL yeah I'm lost.
it might be helpful to write out what du is, after the substitution but separately before fiddling with the integrand
I could also do IBP and have $y = u$ , $dv = \frac{1}{x^2 + y^2}$
a common thread here is not writing or thinking about dy
04:10
I thought you should think about what you want to differentiate/integrate first to simplify?
nvm
I forgot x was constant
so that u sub earlier would give $du = 2y dy$
yeah, and that helps
after some simplifying yeah it turns into $\int \frac{1}{2u}du$
grazi
@TedShifrin I found that $n =2$ case is proved in RCA. I just followed that proof.
04:37
would it be easier to solve volume of a given ellipsoid with spherical coordinates rather than rectangular?
getting the volume of an octant and multiplying by 8 still seems rough.
'solve' how? via an integral? i dunno. unless it's a sphere, there won't be spherical symmetry so it wouldn't come out that simply.
the vibe in any coordinates might be to make use of changes of variable whose effect on the volume are understood, that increase the symmetry.
hmm what if it's a sphere, then spherical coordinates are the easiest?
sure? you tell me. it's hard to make it complicated if it's a sphere.
if you want to do this via an integral, i'd focus on finding some integral that you can take and get the right answer. then maybe try it in other systems and decide what the best one is for the future.
it isn't fruitful (in my very humble opinion) to want to know in advance what the 'best' way will be, because it's an unnecessary side line to just solving the problem.
or just writing out the integrals and seeing what you like the best. sometimes people remember some techniques and not others and what's 'easy' for them might not be 'easy' for you or for me.
04:57
Okay so I came up with this: $4x^2 + 4y^2 + z^2 = 16$ an ellipsoid volume given by $\int_{-4}^{4}\int_{-\sqrt{4-x^2}}^{\sqrt{4-x^2}}\int_{-2\sqrt{4-x^2-y^2}}^{2\sqrt{4-x^2-y^2}}dzdydx$ does this make sense
advance warning that i am not rendering chatjax at the moment but with that qualifier, it looks OK to me.
oh okay, sorry it's very messy then.
on second closer look, i'm not sure x should be going from -4 to 4. so, something seems a little off.
but I was wondering if instead of circular cross sections spanning the z-axis, doing semi-ellipses spanning 0 to 2 pi might be easier but Idk how I'd convert the coordinates
this one has a cylindrical symmetry. the cross sections through planes z = k are circles (when they are not points or empty). that suggests one way of rewriting it or alternatively setting it up.
05:02
oh should be between -2 and 2 for x
the radius of the largest circle was 2 not 4
05:14
wait a parabola is a cylinder?
in R^3
thought they had to have depth I came across $z=1-y^2$ being called a cylinder
nobody's calling that a cylinder.
OK. this all works better if first you post the thing and then react to it, than the other way around.
i'm sorry lol
but I take things very literally, it sounds like they're saying bounded by the cylinder z=1-y^2 , standalone statement
sometimes people will call a 'cylinder' anything whose cross section through a series of parallel planes is the exact same plane curve. when that cross section is a circle, you get a traditional cylinder.
05:18
AND between the planes
implicitly, they're using the more general definition.
Oh okay, I think my prof mentioned something about planes being cylinders so that makes sense I guess.
if you take a single circle in a plane, and then put lines perpendicular through the plane through all the points in the circle, you get a standard cylinder.
here you're doing that with parabolas z = 1 - y^2. the cross section through any plane x = const. is the same thing, that parabola.
or maybe you're not doing that, they're doing that.
maybe it's distracting terminology. the problem is the same if you replace "cylinder" with the word "surface."
I won't complain, because adding more names for shapes and such just makes things more complicated.
I'm just used to the everyday notion of a cylinder.
okay but how is there even a volume for this surface if the "cylinder" is just a parabola in the yz-plane?
OH wait it spans the x-axis
yeah. just like {(x,y,z): x^2 + y^2 = 1} isn't a circle but a... what was that word again
if the constraint involves only two of the variables, the one that doesn't appear is the axis along which the thing is a 'cylinder.'
05:29
a can I think was the colloquial term
just as certain ellipsoids are footballs
i would have said 'generalized cylinder' or something else. it might be worth looking in the index to see if they define the term elsewhere and include a picture with examples.
I was joking :P you're right it's just a mathematical cylinder.
my prof brought a can of pringles (chips shaped like hyperbolic paraboloids) within the first few weeks when we were learning surfaces.
A delicious lesson that was.
 
1 hour later…
06:48
What does with mean in logic?
Is it and?
example statement pls
For all a ∈ R, |a| ≥ 0 with |a| = 0 if and only if a = 0.
I was thinking it means and.
the logic is this: for all a in R, if |a| = 0 then |a| is greater or equal than 0 iff etc.
it's a conditional
How did you know?
It was not written in my logic book.
you exemplify the statement. Suppose it is true, and let a = 0. Then, |a| \geq 0 is not implied if |a| is not equal to 0
i.e., suppose a = 0. then you don't get |a| \geq 0 unless |a| = 0
suppose it is true, and let a = 0
06:56
Oh I see.
I was wondering why even state redundant antecedent but now I see the purpose.
hm, wait a second i'm not sure of my answer anymore
I am confused too.
can you post the statement in your logic book?
It is analysis.
I will post proof too.
May be it will be easier to understand.
go with the "and" interpretation, it makes sense here
07:02
I would normally parse P with Q if and only if R as $P\land(Q\iff R)$
|a| \geq 0 and |a| = 0 implies a = 0 obviously, and a = 0 implies |a| = 0 and |a| \geq 0 obvious
so use "and"
So I was right?
yeah
but "with" is not unambiguous
07:03
I was confused with proof. May be I should try to read it carefully.
i would dare to say there are cases where it means "if", but in this statement it clearly means "and"
@KarlKroningfeld oh that's an interesting contribution
but it is a weird situation in any case, @KarlKroningfeld, i think in the book's statement, it means "if we have both P and Q, then we have R, and if we have R, then we have both P and Q"
It has for all a in R so if you choose a=1 then a=0 doesn't make sense to be called theorem. So it must be and.
no, that's fine, that's quantification over the reals
read that statement as follows:
for any a in the reals
if we have |a| \geq 0 and |a| = 0, then we have a = 0
but also, if we have a = 0, then we have |a| \geq 0 and |a| = 0
that two way conditional is the meaning of "if and only if"
@NotTfue note that "for all a in R, if P, then Q", does not imply that P holds for all a, and neither does it imply that Q holds for all a
was gonna write or.
The statement $|a|\ge 0$ applies regardless of whether $a=0$. A different example of the same wording: artofproblemsolving.com/wiki/index.php/AM-GM_Inequality
07:15
Makes sense now.
@KarlKroningfeld oh right, of course
that's why i would interpret the statement as (P and Q) iff R, though
but if P is always true, it is the same as P and (Q iff R)
gah, logic
logic dictates how your proof will appear :(
Indeed, this disagreement would be resolved with a case in which $P$ is not true.
@NotTfue for the question you posted, for all a in the reals (P and Q) iff R will work to prove the statement
@NotTfue you should worry only about that
I see it clearly now. Thanks.
07:22
so, what your proof requires is:
Prove (P and Q) implies R
and
Prove R implies (P and Q)
Wouldn't you actually want to show the inequality holds?!
you suppose the inequality holds in order to prove the statement
the statement doesn't comment on whether the inequality is true or not, but we know it is always true
Note that Q is stronger than P. So, it would have been more efficient to simply state Q iff R
Why did they include P at all?
weird eh, probably cause I can see a student be mystified by having |a| \geq 0 and |a| = 0, and not knowing that it means a = 0
These particular authors start the proof of (i) with a proof of $|a|\ge 0$ :)
Why?
07:33
huh, probably because they only just defined absolute values as $|a| = a \iff a \geq 0$ and $|a| = -a \iff a < 0$
so they're beginning to develop the algebra of absolute values
is my guess
these things aren't obvious at all to beginners heheh
@shintuku Yeah but strangely I don't see that pattern of proof on my book.
what's the book btw
Not that they had to prove it as part of the statement of Positive Definiteness :P
@KarlKroningfeld lol
I was gonna use the pattern you provided.
@shintuku William Wade.
07:36
@KarlKroningfeld man authors need to be careful about this sort of thing, especially since for someone new in analysis the logic is obliteratingly confusing
@NotTfue is it course assignment? i would suggest to start with Spivak otherwise
Authors these days
@shintuku Nope. I just wanted to improve my analysis it is book used by professor.
@shintuku Sure It is well known here I guess I should start to do spivak. I don't think Rudin and Wade is good for beginner like me.
yeah he does the properties of absolute values at page 11
Sounds good :)
the book is called Calculus, don't let it fool you it is what people do in a first real analysis class
07:46
May be naming it analysis would scare people away :P
 
1 hour later…
08:52
Hello people, how to solve x^2+4xy+5y^2-4x-6y+7. My initial idea was about pair of straight lines, and finding the M value for the homogeneous part, and so on. But turns out the homogeneous part doesn't have a real solution. Not able to get any other ideas, even completing squares seems to not work
@shintuku There is ambiguity in book. Now I realize book means i) For all a in R |a|>=0 ii)For all |a|=0 iff a=0 lol.
Book sucks.
I read it like 5 times already and it didn't make sense. Now it makes lol.
oh that's a reading that makes sense
@KarlKroningfeld Mistake.
i think it was KarlKroningfeld's original interpretation too
This "P with Q if and only if R as $P\land(Q\iff R)$"
I thought we all were thinking this.
09:05
my reading has lead us astray
I don't think so. The author should be held accountable. Author should use non-ambiguous word when making book.
09:52
Given the integral: $\int\limits_{0}^{+\infty} \dfrac{\sin{\beta x}\cdot \cos{\gamma x}}{x}\cdot e^{-\alpha x}dx$. I know that this is solved via integration or differentiation by parameter, but I'm not sure how to do that. Differentiating will leave us with two more parameters, and about the integration I don't know which function gives such an output after integration
 
4 hours later…
13:34
If $u(x,y)$ is a harmonic function on a disk $x^2+y^2<R^2$ then a function
$$f(z) = 2u\left({z\over 2},{z\over 2i}\right)-u(0,0)$$
is holomorphic on $|z|<R$ with real part $u(x,y)$. I know how to solve this but it uses some series expansion (double power series comes out) and is quite nasty. Is there any simple way to prove this? Maybe using Schwartz/Poisson integral formula.
 
3 hours later…
16:14
Let $u$ be a continuous function on $\Bbb D$ which is harmonic on $\Bbb D\setminus\{0\}$. If
$$\lim_{|(x,y)|\to 1}u(x,y) = 0,\quad\lim_{|(x,y)|\to 0}u(x,y)/\log(x^2+y^2) = 0,$$
then $u$ is identically zero on $\Bbb D$.

Since $\lim_{|z|\to 0}{u(z)\over\log|z|} = 0$, near $0$, $\epsilon\log|z|\leq u(z)\leq-\epsilon\log|z|$ for some small $\epsilon$. If $|\zeta| =1$ or $\zeta = 0$ then $\liminf_{z\to\zeta}u(z)+\epsilon\log|z|\leq 0$. Hence by the minimum principle of harmonic function, $u(z)\leq -\epsilon\log|z|$ on $\Bbb D\setminus\{0\}$. Similarly, if $|\zeta| =1$ or $\zeta =0$ then $\limin
Does this make sense?
Jam
Jam
16:29
Do i have a p-group for evey $p^k$ that divides the order of the Group? So for $|G|=2^{10}$ i have at least one 2-group of order $2, 2^2,2^3...$?
are you talking about p-subgroups of a fixed group or what? I don't completely follow
Jam
Jam
yes exactly
i know there exist a maximal p-subgroup
but can i get that inside this maximal exists another p-subgroup
16:46
yeah, this is true
I think rightmost equation should have it's limit reversed.
I did u substitution to the middle one.
Do it carefully.
u=t-tau.
well limit of middle one is tau=0 to tau=t and limit of right i get t-(tau=0)=u to t-(tau=t)=u
You're not done yet.
@TedShifrin ok I will try again.
@TedShifrin may be the integral is negative...
17:00
What is the most important thing you must do when you do a $u$ substitution?
Change limit I guess.
No.
There might not even be any limits.
or... derivative
Right. As soon as you write down $u=t-\tau$, you must write down $du = -d\tau$.
i took derivative in terms of t
was thinking something was wrong there
so you can reverse limit
17:08
Yes.
oh i see i made a dumb mistake :P
nearly stopped me from doing sin thank you
sinning*
17:39
dang wolfram alpha standard computation time exceeded :( I hope someone here can verify. Given solid $Q=\{(x,y,z): 0\leq z \leq 9-x-2y,x^2+y^2\leq 4\} \rho(x,y,z)=k\sqrt{x^2+y^2}$ computing mass with cylindrical coordinates answer should be $48k\pi$ but I got $72\pi k$ I tried to verify if the rectangular setup was correct $\int_{-2}^{2}\int_{-\sqrt{4-x^2}}^{\sqrt{4-x^2}}\int_{0}^{9-x-2y}dzdydx$ in wolfram alpha but yeah didn't give me the answer. my cylindrical form was
$\int_{0}^{9}\int_{0}^{2\pi}\int_{0}^{2}k\sqrt{2^2}drd\theta dz$
oh crud I forgot the extra r
welp I got it right now
carelessness is going to cost me on this test I can feel it ..
17:58
@Obliv Herre, you can have mine.
vey much appreciated, now I can replace the one I lost :P
18:30
@Obliv Where did the $z\le 9-x-2y$ go? And your density is incorrect. $r=2$ only on the boundary, not inside.
@XanderHenderson You're generous when you have a surplus.
@nickbros123 see if this convinces you. Let S be a subset of the reals. If it has a minimum, that's the infimum. Suppose it has no minimum, and suppose it has some lower bound.

Let L be the set of lower bounds of S. Clearly, L is bounded above by the elements of S, so it has a supremum, sup L. Suppose s is an arbitrary element of S.

If s is to the left of sup L, then sup L is not the supremum of L, which is a contradiction. s is not at sup L either, because since S has no minimum, that implies there is another element of S at sup L's left, which we've seen is a contradiction. Therefore, s
18:54
@TedShifrin weirdly enough I can't seem to explain what I wrote but it got me the right answer.. I thought of it as a circle shaded in given by $x^2 + y^2 \leq 4$ bounded by the z values $0\leq z \leq 9-x-2y$ so does the rectangular form I wrote make sense first?
z integral just becomes the bounds, y and x integrals follow standard circle integral bounds so you have a shaded in circle going from z=0 to z=9-x-2y
to convert that to cylindrical I just thought radius goes from 0 to 2, theta is the whole circle, 0 to 2pi, z is the independent variable with minimum value 0 and maximum value 9
actually thinking on it now, I feel like what I wrote is wrong and would give you extra volume..
19:12
The correct set-up should be $$\int_0^{2\pi}\int_0^2\int_0^{9-r\cos\theta-2r\sin\theta} kr^2\,dz\,dr\,d\theta.$$
you're so right, that's also evaluated to 48pi k
I have no idea why my integral worked
maybe there was a symmetry where the extra volume of the circle was balanced
like the plane intersects the cylinder in the middle?
i wonder if i'd get full marks if that were on the test :P
definitely not ted's test
19:27
@Obliv I would give you negative marks.
:P
there's got to be a reason why they're both 48 pi k.. I just can't figure out what
maybe wolfram alpha is broken
@Obliv Coincidence? Maybe.
I grade most on the set-up. The actual answer is only a point or so out of 10 or 20.
Indeed, on the multiple integration test, I often have the problem be just to set up the integrals.
@TedShifrin I do the same. (Both of the above!)
if I knew the explanation why they were equivalent integrals, and wrote the explanation, would I get the credit for the problem?
or would that be considered not following the rules?
19:36
There is no explanation.
I can't even get the same answer doing the $z$ limit as $9$.
Interesting. The integral of $1$ equals the integral of $r$. Does that mean you get credit for the wrong integral? NO.
I'd spend my time understanding how to do things correctly, not incorrectly.
20:06
@nickbros123 I am not sure what you mean by "solving", but you can complete the squares as (x+2y)^2+y^2-4(x+2y)+2y+7=u^2+y^2-4u+2y-7=(u-2)^2+(y+1)^2-12, where u=x+2y.
I assume we want the zero set.
@nickbros123 Maybe you should rotate the axes to get rid of the $xy$ term. Then you can tell it's an ellipse.
What is the context of this question? That is, what class are you taking in which this occurs?
if you want to compute the circumference, this could lead deep into complex analysis
What's one major difference between complex analysis by Churchil and Brown and by Ahlfors?
Ahlfors is a graduate-level math text. Churchill-Brown is a bit more aimed to undergraduates and science/engineering. Not so sophisticated.
During my UG, I studied complex analysis from C&B but I don't know much about Ahlfors'.
20:19
To do Ahlfors, you need some basic topology and notions from analysis like uniform convergence and more.
Both are a bit old-fashioned at this point.
thanks @TedShifrin.
Do you have any recommendations for a text on complex analysis?
apart from Ahlfors'.
Ahlfors is still a standard text. Students like Conway; I find it unsatisfying. People with a number theory slant like Stein/Stakarchi.
@TedShifrin I'm not Ted, but I like Freitag
it's very clear and induces some topology/algebra when it is appropriate
and the author is a great teacher :) still teaching at 80 years
I like for example that it interprets the Cauchy integral theorem in terms of homotopy and in terms of homology. Much more useful than just the statement about intergration over a circle
(but you don't need to know homotopy/homology beforehand)
I should note that it also has a number theory slant
it does modular forms and the zeta function
but modular forms occur naturally when you study elliptic functions
@nickbros123 I would not suggest reading set theory at your level as I perceived. You could try reading the first chapter of Zorich's Mathematical Analysis.
@Yai0Phah Zorich :-)
20:29
What's wrong with that?
Nothing. He's one of my favorites.
@TedShifrin Why is Ahlfors' book "old-fashioned"?
I've learnt a lot from Zorich's.
I regret that I did not read it when I was a freshman.
freitag has over 400 exercises with hints. Elliptic functions are an interesting topic that was pretty important historically
20:32
While studying Zorich's it feels like the author is sitting with you and explaining things to you personally.
oops I wanted to ping @koro with my recommendation of Freitag
I actually have never seen Freitag's book, so I can't speak to it.
@LukasHeger: Do you mean this book link.springer.com/book/10.1007/978-3-540-93983-2?
@Yai0Phah Some of his topological definitions were popular in the mid 20th century, but are definitely not what we would do now. When I taught out of the book, I certainly did not follow it word for word or section for section.
20:36
I ask because while typing the name, I got recommendation for Freitag 2 as well.
Actually, Lang's book is quite good. I also like Narasimhan for an advanced book.
well, vol 2 is more advanced and assumes you know vol 1 of course
Hille is another good book.
2 volumes, also.
vol 2 does Riemann surfaces, sheaf theory, several complex variables, interesting stuff, too :)
@LukasHeger: That makes sense. I asked because for example Herstein has atleast two books on abstract algebra and both seem pretty much the same to me.
20:38
I guess that it would be better to refer to both authors.
viz.1. topics in algebra and 2. abstract algebra
oh sorry I forgot Busam
train to busan
I forgot because the second volume is single-author (of course, that's no excuse!)
yeah I meant Freitag-Busam
Ahlfors' book costs only equivalent to 640 USD.
20:49
quite a steal!
yeah, you see. It is because it is imported.
@Koro Definitely not the same. One is the classic, and the other was a far watered-down version.
This chat is full of elitist accademcians. Thanks for throwing back the progress of humanity.
it's my pleasure
@DanielDonnelly ?
21:01
If $X$ and $Y$ are compact homeomorphic metric spaces, does it follow that their isometry groups are homeomorphic?
consider the unit ball with respect to the 1- or 2-norm
21:30
Is it true that $o\left(\frac{f(x)}{g(x)}\right)=\frac{o(f(x))}{g(x)}$ for $x \to x_0$? I believe yes, because $o(f(x))$ means that there exists a function $h$ such that $o(f(x))=f(x)h(x)$ with $h(x) \to 0$ as $x \to x_0$. Hence $\frac{o(f(x))}{g(x)}=\frac{f(x)h(x)}{g(x)}=\frac{f(x)}{g(x)}\cdot h(x)$, but $\frac{f(x)}{g(x)}\cdot h(x)$ with $h(x) \to 0$ as $x \to x_0$ means $o\left(\frac{f(x)}{g(x)}\right)$ for $x \to x_0$. Is this correct?
21:43
I have some vague memory about Donaldson's book on Riemann surfaces. I do not understand much.
@Sonozaki Not in general, no. What if $g(x)\to 0$ as $x\to x_0$?

« first day (4508 days earlier)      last day (809 days later) »