« first day (2515 days earlier)      last day (2804 days later) » 

14:05
Hi chat
Damn, not all math is commutative
We wouldn't have the word "commutative" if it was
We are however lucky that duality is involutory. That saves us from having to discuss cocococococolimits.
True but it makes you bat an eye whenever someone says "coconuts"
cocommutative would be just mutative then ? Or mmutative ?
14:12
The latter, I'd say.
Surely the latter
It's the only reasonable option
@Balarka I think doing atop through the mindset of categories is a wears sunglasses categorial imperative
@Daminark I Kant with this shit.
I'm still proud of my "nvergent topology"
I also like the untable topology.
14:21
(I actually forgot all of its properties. But I could probably recover them with some thought)
I think we've just turned into a bunch of memers since @Ted's not around to rein us in.
(And/or digging through messages I've posted on this chat and places)
goteem
And lol wait till he gets back and finds us all making memes and only discussing $\infty$-categories
14:23
starts the Aspirin for Ted fund
Lmao
Also @Akiva kek
What's the untable topology?
Unstable?
@Daminark Cocountable.
Open sets are those whose complement is countable.
Oh I see
The nvergent topology on $\{1/n:n\in\Bbb N\}$ is the one whose open sets are the complements of sets with convergent sum
(plus an empty set)
Hi. I need help in binomial theorem
0
Q: Find the remainder when $2^{2013}$ is divided by $17$.

FawadThis question was asked me in binomial theorem chapter. I don’t know if I am doing it right. Here are my steps Let $(1+x)^n=C_0+C_1x+C_2x^2+\cdots C_n x^n$ Let $x=1,n=2013$ $2^{2013}=C_0+C_1+C_2+\cdots C_{2013}$ Let $2^{2013}=17k+r$ from here I am lost. Help.

14:26
Notably, for uncountable sets with the cocountable topology, all compact sets are closed, but the topology is not Hausdorff.
And for the nvergent topology, if I recall correctly, all compact sets are finite.
Wait hold on
Yeah I think that's right
Also the intersection of any two (nonempty) open sets is nonempty
14:45
@Semiclassical i got into UWO
:}
I AM NOW A WESTERN MUSTANG!
@Dodsy YESSSSSSSS
:D
:D
m'dudes
I was literally at the end of my rope
the rope is infinitely long, both at the end and at the start
literally?
14:48
lmfao :)
yes, literally
what is your rope?
Like a physical rope? Merp
not sure if that's too comforting but
yep
hm
this is just the best day of my life.
hi chat
@Dodsy wooooooo
14:51
I've spent 4 years trying to get here
and now I'm here
Congratulations.
ha ha
you got accepted into a mathematics program?
Bastille pompie
song
I dont think you tried 4 years to get into the chat :D
14:58
@Dodsy Congrats man!
@s.harp Plot twist...
@Secret you might also like to focus at the same time on the motivational part I used to post.
which ones?
ah this one, yeah, you used to post these 2 years ago which is where I am now as I tried to crawl the maths chat to find the very first maths chat post of mine
Btw I have been dreaming about a lot of stuff that is a mix of integrals, recursive functions and ordinals recently. however, with the exception of the most recent one, the others are not even mathematically consistent
Outside of these dreams, I also came up with some integrals which unfortunately I have not though about how to approach them yet, but some conjectures were being proposed which will be investigated later (one of these are the tetration integrals that I tagged you and simpleart in)
15:10
6 hours ago, by Secret
However, after some random pondering and free association about the dream itself after I woke up, I then came up with something that might be a little bit more sensible:

Let $I=\int f(x)dx$. There exists functions $f(x)$ such that the following holds for all $n \in \Bbb{N}$

$$\int f(x) dx =I^2 = \left(\int f(x) dx\right)^2$$
$$\int I^n dx =I^{n+1}$$

Whether there are such $f(x)$ I don't know, and I will deal with it later...
6 hours ago, by Secret
$$\int I^n dx =I^{n+1}$$
Ok it turns out we can do something about this: Differentiate both sides to get
$$I^n =(n+1)I^n f(x)$$
We can rearrange this equation to get
$$I^n (1- (n+1) f(x))= 0$$
This give us two conditions:
$$I^n=0\text{ or } f(x) = \frac{1}{n+1}$$
Now consider the base case:
$$f(x)=0\text{ or } f(x) = 1$$

Therefore the only consistent solution is the zero function
6 hours ago, by Secret
In general, the only function that is nilpotent under indefinite integration is the zero function
I say it is invalid to go from $I^n (1- (n+1) f(x))= 0$ to $I^n=0\text{ or } f(x) = \dfrac{1}{n+1}$
why? functions are like numbers, you don't have zero divisors in functions
so you cannot have nonzero functions that producted to zero
(you might have those if you restrict the domain, due to various things like periodicity and so on)
function multiplication is pointwise multiplication
@Secret Happily these days I manage to bring into reality what in other times I only dreamt like you. Particularly, I'm studying some results entirely special, pretty similar to some obtained by Ramanujan in his second Notebook.
As you know, there is always some difficulty about anything related to the last results one gets. Some mystery is always present to some extent, and this is not bad at all (it's part of the fantastic flavour of mathematics).
15:18
I see. Meanwhile, as it takes qutie a bit of time for me to get number theory background solid, which is why mostly the integral investigations I am currently dealing with are mostly symbolic manipulations.

I do have a preference about tetration due to my interest in ordinals, which is why you seemed to see quite a bit of my integrals often involving a tetration integrand. However to deal with these properly, one needs good Lambert W function manipulations
Most of the time, I and my dreams came up with a lot of ideas and I don't have any idea what they mean yet, thus I will often just store in notebooks or posts bits of it so that when other users react to them, they will indirectly inform me what I can do about them
For me, my preference is rather than asking them to solve the integrals, I often just post the integral and leave it as is. Most users tend to react with them in different ways, and from that, I gained some insights on what I can do about them
This way I handle mathematical obejcts is very similar to how fine art is presented: You just leave your artwork in the venue, and the audience will came up and connect with their own interpretations
@LeakyNun So you suggests there are nontrival solutions $f$ that will satisfy that integro recurence relation?
@Secret no, I don't
@Daminark There's actually a cool way to prove $\Omega K(G, n)$ is homotopy equivalent to $K(G, n - 1)$.
Do you know Yoneda lemma?
Not quite, how does it go?
@LeakyNun because if you said it is invalid to split it up into these two cases analogous to polynomials, it means the equation has to be dealt with as a whole, which might provide gaps for a nontrival f to solve it
15:24
Hey @Alessandro
@Secret no, they would be discontinuous
This is because once you split that up into those two cases, the trivial solution will jump out
@Secret Yes, that's fine, but at the same time you might like to find a way to solve the problem, a way which defines you in a particular way, creating your mathematics there, not like in books, papers, or like suggested by others on chat. That way which defines you mathematically in an unique way.
@Daminark Suppose $F : C \to \mathrm{Set}$ is a functor from a small category $C$ (ie where the hom-sets are actually sets). Then for any fixed object $A$ of $C$, natural transformations between $F$ and $\hom(A, -)$ are in one-to-one correspondence with the elements of the set $F(A)$.
hello
15:27
You really need to prove it for it to make sense, at least in my experience
I have a bit of a question about a paper i'm reading.
they define a flux vector is a vector $\mathbf{u}\in\mathbb{R}^R_{\geq 0}$ where $R$ is a finite set of reactions over $\Lambda$ which is a finite set of species.
That sometimes happens, but it is currently rare in the maths domain for me. If I recall, the only truly original stuff I have done so far is proving theorems that showed that almost all algebraic systems that involve division by zero must be nonassociative (It is original because the maths community think it is crazy to mess with pathological things like division by zero, but I am kinda a reckless rule breaker in some way.
One of the things that attract me most are precisely those pathological mathematical objects that most fear to work with, because they are weird and helps to flex the mind)
Hm, maybe it's useful to think about the contravariant version (you know what contravariant functors are, right?). That says natural transformations between $F$ and $\hom(-, A)$ are in one-to-one correspondence with $F(A)$, where $F$ is a contravariant functor, ie, a functor $C^{op} \to \mathrm{Set}$ instead.
@SohamChowdhury It's also super-easy to prove. You can prove it.
(this is a paper about continuous chemical reaction networks; in other words, a system of chemicals where reactions are used to get from one state to another as a way of solving problems.)
@Secret That's OK, a good amount of courage is needed to go there and try things that most wouldn't dare or think of in mathematics. That makes a huge difference.
15:31
it then says that the support of $\mathbf{u}$ is the set $\text{supp}(\mathbf{u}) = \{\rho\in R\mid\mathbf{u}(\rho)>0\}$ where $\rho$ is a reaction $\in R$ (of which I can give details if necessary).
I'll try it at some point
@LeakyNun What's your suggestion to approach that equation so I can took account of the implied discontinuity?
@Daminark So you understand the statement of the lemma, yes?
it says that a flux vector is applicable at a state if every $\rho \in \text{supp}(\mathbf{u})$ is applicable at the state and if (the state is $\mathbf{c}$) $\mathbf{c}(s) + \sum_\limits{\rho\in R}\mathbf{u}(\rho)\Delta\rho(s)\geq 0$ for every $s\in\Lambda$ and finally that a flux vector sequence $\mathbf{U}$ is a tuple of flux vectors (sorry for so much detail; I am unsure what's important here).
In which case, consider the "homotopy category of CW-complexes", $C$, whose objects are CW complexes and hom-sets are $\hom(X, Y) := [X, Y]$.
15:34
I think so, yeah
the paper then says in the main algorithm to "compute the max support flux vector sequence $\mathbf{U}_{\mathbf{c},\epsilon}$
so first: what do the two subscripts mean here? second, what is the reference to "max support" seeing as it defines no such thing as a support flux vector sequence or a max flux vector sequence or anything, just a flux vector sequence?
@Daminark Maybe I should add one thing: Suppose $F = \hom(-, B)$ for some other object $B$ of $C$. Then natural transformations between $\hom(-, B)$ and $\hom(-, A)$ are in one-to-one correspondence with $\hom(A, B)$.
Applying this to our homotopy category of CW complexes: First, a fact: $H^n(\Sigma X; G)$ is naturally isomorphic to $H^{n-1}(X; G)$ for any $n$ (you will soon prove this). Using that, you get $[\Sigma X, K(G, n)]$ is naturally isomorphic to $[X, K(G, n-1)]$. Or, $[X, \Omega K(G, n)]$ is naturally isomorphic to $[X, K(G, n-1)]$.
I see
(Though we may end up using this to prove that :P )
This gives a natural transformation $[-, \Omega K(G, n)] \to [-, K(G, n-1)]$ which is also an isomorphism. By Yoneda this should correspond to an isomorphism $\Omega K(G, n) \cong K(G, n-1)$ in the homotopy category (which is precisely a homotopy equivalence), if I am not wrong.
@heather Without reading the paper in detail, I am guessing you have a sequence of flux vectors (which is stored in $\mathbf{U}$) for a given state $\mathbf{c}$ (thus that might explain the subscript $\mathbf{c}$). So the max support flux vector sequence probably means to maximise $\mathbf{U}$ given $\mathbf{c}$ under some parameters (given by the $\epsilon$ subscript). But I felt like we might need to read that paper in more detail to deduce what it is doing
15:43
Yeah, it definitely should. Yoneda should say $\text{Nat}(\hom(-, B), \hom(-, A)) \cong \hom(A, B)$ is not only a bijection but sends "natural isomorphisms" to isomorphisms.
Bashing out the proof explicitly should do this.
What I'm worried about is I'm not really working on a category, but a homotopy category, where "eveything is coherent upto homotopy". I guess I can handwave this off.
@Daminark yeh peter may is doing everything backwards man
@Secret Just to be aware of one more thing: one needs a huge (hard to measure) investment in mathematics to get some results, some sacrifice that needs good consideration, to be sure you want to do it. I mean not just going half a way, that means a lot of your time was wasted.
@Secret sometimes the results simply do not come, not that fast, and maybe you even need to crawl at some point.
It took me 5 months to prove those theorems I stated to you earlier, and I expected when I get my nonassociative algebra and cateogry theoric background solid, the rest would definitely took many years to prove as division by zero enters the nonassociative regime
Lmao, it's fun, let's see how this experiment goes
I hope what I said so far is right.
@BalarkaSen I know it now.
I know it for now, rather. If I don't do abstract nonsense for a month I'll forget.
15:54
@MikeMiller I am right in saying that in the homotopy category of CW complexes, $[Z, X] \cong [Z, Y]$ for all objects $Z$ implies $X$ is homotopy equivalent to $Y$, right? I guess I just want to know if the Yoneda lemma works in the "homotopy localization" or something.
The interesting thing about that division by zero algebra personal project of mine is its nature means it cannot fail (the only way for it to fail is when it is shown to be equivalent to a halting problem without oracles, which I highly doubt it will happen since the axiomatic systems are quite consistent)

Currently the direction suggests at least one nonasssociative division by zero algebra exists. If we can find that, everything we knew about division by zero will be revised

But of course, the alternate outcome is that we instead found a series of no-go theorems ruling out the existence
@BalarkaSen that looks like full faithfulness preserving isos for the $h^Z$ functor (but htpy instead of vanilla)
@Secret are you aware of wheels?
Yeah, it works in plain vanilla categories because Yoneda is exactly that.
My philosophy of answering any questions thrown by people or nature itself is to find the most perfect answer possible, an answer that once given, is already all that is need to know thus closing the question completely.
I want to know if it works in homotopy-localized categories.
15:56
Hm.
@SohamChowdhury I do, I am aware and had read the papers on Wheels, Meadows and so on
Akiva and co. also helped on the analysis of the Wheel papers
but Wheels only defined zero terms, not zero inverses (because they showed any attempt in doing so, the wheel becomes trivial)
(btw, for those who are unfamilar with the terminologies I made up for my research, zero terms are basically things like x/0, 0x, x0, 0x0, x0y 0^n etc. that does not collapse to 0)
Ugh I am doing too much abstract nonsense right now.
It's easier to just give Yoneda the fingers and use the path-space fibration $\Omega K \to P K \to K$ and the fiber long exact sequence.
Like an honest topologist.
But anyway, without at least a solid abstract algebra background, I cannot continue my analysis on nonassociative systems, which is why the project is temporary put on hold
16:00
Oh wait no. Oh god.
That does not give me a weak homotopy equivalence $\Omega K(G, n) \to K(G, n - 1)$
It just gives me something at the level of homotopy groups.
@Secret no idea
I just discovered today that there's a whole stackexchange for martial arts
@s.harp yeah 4 years to get into university :P
Yeah, ok, so I can construct a map $\Omega K(G, n) \to K(G, n-1)$ but that involves the representability too. Basically choosing the identity element from $H^{n-1}(\Omega K(G, n); G)$ and representing that. Then that's a homology isomorphism, which gives a weak homotopy equivalence, I think.
Seems like Peter May wins on this one, @Daminark.
Hi @Ted
Hi, Balarka, Demonark, Nate, et al.
16:11
TED
I was accepted!
:D
Balarka, you're sounding like true demons have taken over your brain.
Yea! Great news!
Well they have
Thanks Ted.
16:13
To the second one?
to UWO
The only thing is that it's part-time, the lady said I applied a little late and that the program was full, but that I could be admitted for part-time. i said alright and she said okay you're in. :}
But I think it'll be good, I haven't been in a school day program in 4 years.
good to ease my way in.
Just got back to it, and there is actually a problem with that I think. But will see how I can get it around that.

Hi Prof. @Ted
@Dodsy Cool ! I am really glad for you ? But why did the decision change ?
Hi Ted
How is/was Croatia ?
@Astyx so basically at first it was "sorry the program will be full by the time we review your application"
then "well you could apply to part time it will cost 50 bucks"
turns out it costs 145 bucks so i said "well i don't want to pay 145 bucks"
so then she said "i'll email you back tonight" she never emailed me. I called her today and she said "good news, if you want to do part-time I can admit you right now" and I said "alright!" and now I am a Western Mustang.
16:22
What's the difference between part time and full time ?
Or whatever the other option is
I take 2 less classes a semester.
so, I just take them during the summer of my first year
and then for second year I'll be full-time.
How many classes do you take a semester again?
3.
instead of 5.
why do people even take full time years then
part-time is less pressure
yeah, but for instance, the math department only offers first year math courses in the summer.
and the summer is probably going to kill me.
16:27
how many courses do you take on the summer then
I'll have to take 4 classes I guess.
tit for tat lol
so you make up for the less classes
Yeah haha
but summer goes from may to august
4 months.
Right, so more pressure
yeah
it'll be brutal
This is just weird, I started losing hope, thinking I'd have to wait another year
16:31
you might think that that was the better option on those 4 months i guess
but it'd be fun
total fire
Hi, I'm trying to find a question I saw here a week or so ago. It was about three 3-dimensional points and showing whether or not they form a right-angled triangle
I can't find this question though :(
Hi @Alessandro
I'm wondering if it was here or mathoverflow (or maybe math educators). Here seems most likely, but I've searched as best as I can on all 3
Hi @Astyx — it's gorgeous! Except for my dental catastrophe. The dentist office was out of the 70s.
@TedShifrin Hey!
16:46
@DigitalTrauma You could use dot products. Or you could see if the smallest two of $\|x-y\|^2$, $\|y-z\|^2$, and $\|z-x\|^2$ add up to the largest one
Just messaged elsewhere. ;)
(a là Pythagoras)
Mr @Pedro!
DogAteMy!
Ted!
$\det^{-1}$
17:13
[Chemistry]
I have one bad news and one good news
V: Finally get the standalone program to work, thus I can finally get the answers for a certain section of the paper after 2-3 weeks of lack of progress)
X: The H2Se calculation still refuses to converge. Upon advice of my peers, I think I will file a error report to the help centre of the software soon
Meanwhile, I need to figure out how to reorganise my too maths saturated brain to back to chemistry, so I can read through the literature faster to prepare my proposal
Hi, could anyone help me with the calculation of the cohomology with compact support of R^n, as done in Hatcher's book on page 244?
I dont understand why $H^i(\mathbb R^n, \mathbb R^n \setminus B)$ is zero for $i=0$. To me it seems that $H^i(\mathbb R^n, \mathbb R^n \setminus B) = H^i(S^n)$, so $H^0(\dots)=\mathbb Z$.
Anybody simulates stuff in mathematics with computer hair ?
@BalarkaSen Yes, as long as there's a map inducing that equivalence.
@Perturbative thanks! Sorry, I didn't see your message!
my computer has been very slow lately.
17:28
@AkivaWeinberger Thanks - yes - I'm not specifically looking for methods. I wanted to find a specific .SE question
@Semiclassical I received a 74% on my functions test...
don't ask me how.
@MikeMiller Ah ok thanks
so a natural isomorphism of [-, X] and [-, Y] I guess
@Dodsy What weee you expecting?
80-93%
:P
but I finished the course with an 87%
which isn't terrible.
17:33
I just don't know where I would've lost those extra 6 marks.
I forget what I had problems on.
I have to help my grandma move concrete. Thanks guys for all of your support.
@klirk I think $H(A,B)=\widetilde H(A/B)$ (under most conditions)
meaning it turns into reduced cohomology
@TedShifrin really? Apparently there's quite a few people in northeast Italy who go to the dentist in Croatia because it's cheaper but still good
@AkivaWeinberger Ha - after a careful search through my browser history I found it :) math.stackexchange.com/questions/2324551/…
17:57
2
Q: Factor the two polynomials into a product of irreducible elements of $\mathbb{Q}[x]$

ALannisterI need to find a factorization of both $f_{1}=2x^{2}+4x+6$ and $f_{2}=2x^{2}+4x-6$ into a product of irreducible elements of $\mathbb{Q}[x]$. I already was able to do so in the case of $\mathbb{Z}[x]$: $f_{1}=2(x^{2}+2x+3)$ $f_{2}=2(x-(-3))(x-1)$ Note that for $f_{1}$, $x^{2}+2x+3$ has no r...

18:13
@AkivaWeinberger thx, i forgot about that
Hello@BalarkaSen
@ALannister already answered in the comments. 2 is irreducible in Z but not Q.
@arctictern how would you factor that polynomial, then?
you would factor it into irreducible elements
Try rational roots right away without factoring out the 2 first?
Haha.
18:16
have you heard of the AC method?
I teach it in intermediate algebra for factoring quadratics ax^2+bx+c using integers.
Look it up.
Why not try the rational roots theorem?
who said don't try the rational roots theorem?
there's nothing wrong with that
It's also not irreducible in $\Bbb Z[i]$
18:17
Dunno. I kept suggesting it and people keep ignoring me.
So I just figured it was a silly idea.
the point is, if you claim 2x^2+4x-6 is "the product of the three irreducible elements 2, x+3 and x-1" then your claim is incorrect over Q
some people would read your factorization 2*(x+3)*(x-1) as claiming that
I see. What you write in the comments was helpful. I'm going to try stabbing it again
so, just add some more words to make it unambiguous, or absorb the 2 into one of the elements, or interpret 2(x+3) as a single element
All right then
also, it's nice to know the ac method
18:33
@Albas hi
Its been a long time. Hows it going@BalarkaSen
good, good
Good to hear.
Hey
0
Q: Relation between a ringhomomorphism and the induced map on the spectra

latticeLet $\varphi:A\to B$ be a ringhomomorphism. Show that the induced map $$^a\varphi:Spec(B)\to Spec(A),\ \mathfrak{p}\mapsto \varphi^{-1}(\mathfrak{p})$$ satisfies $$\overline{^a\varphi(V(\mathfrak{b}))} = V(\varphi^{-1}(\mathfrak{b}))$$ for any ideal $\mathfrak{b}$ in $B$. Here, for a ring $A$ we ...

Can someone help me with this?
Sorry for bothering@BalarkaSen, but I wanted to know. My uni is giving category theory with just some basic intro to topology. So should I need to do a bit more topology before category theory?
18:38
Nah
which uni did you end up going
Okay. Thanks!
@BalarkaSen Well I have applied for the IISER's waiting for results... There is something called SNU.
@BalarkaSen yep, and the proof is literally just the Yoneda lemma
Thats where I end up going
i see
@MikeMiller right, i just wanted to make sure it applies to homotopy categories
so that's the only way i can prove $\Omega K(G, n) \cong K(G, n \! - \! 1)$
18:53
hi chat
Hi semi
19:22
@Balarka a homotopy category is a category
20:18
Differential Geometry question here: If I want to follow the flow of two vector fields $X$ and $Y$ over a manifold, but $[X, Y] \neq 0$, for instance trying to reduce two different varieties of error, $\mathcal L_X Y$ can be used here to account for that, right?
nvm, I just broke my chatjax plugin at some point.
I'm still trying to wrap my head around exactly what the lie derivative $\mathcal L_X Y$ is in this case.
Actually, the example I gave of two different sources of error commutes, ignore that.
Does 503 | y^3 + z^3 ever happen except for the trivial cases 503 | y, z?
20:44
Hi @TedShifrin
@Ted What parts of Croatia did you visit ? Which would you advise me to go to if I were to travel there ?
Or anyone esle for that matter
21:08
Can someone give some link/resources recommendation to practice curve sketching? Thank You.
21:24
It's crazy how you can have a ton of bad days in a row
and then it's like the clouds part
@AliceRyhl well, 503|y+z?
Guys I need help
How would I show that $\sum\frac{\sin kx}{\ln k}$ converges for each $x$
I mean, it clearly converges for $x=0$. And Desmos seems to show that it converges everywhere (albeit not uniformly)
$\displaystyle \sum_{k=2}^\infty \frac{\sin kx}{\ln k}$?
$k=2$, but yeah
@AkivaWeinberger I missed the first trap and stepped into the second
21:35
?
Oh :P
the first trap being $k=0$
I have no reason to expect it to converge at all :p
Probably by using that the sine is the imaginary part of a complex exponential and using dirichlet's test or so
$\displaystyle \sum_{k=2}^\infty \frac{\sin kx}{\ln k} = \Im \left[ \sum_{k=2}^\infty \frac{\exp(ikx)}{\ln k} \right]$
now we've upgraded our problem up one dimension.
The reciprocals of the logarithms form a decreasing series converging to 0
@SteamyRoot you mean sequence
21:45
All you then have to show is that $\left| \sum_{k=2}^N \sin(kx) \right| < M$ for some $M$, for all $N$
Yeah, sequence
@SteamyRoot that's equivalent
I'm just going through the conditions of the Dirichlet test...
@SteamyRoot Ohh!
There we go, thanks
@AkivaWeinberger I'm not seeing how that is true
you see, I'm clearly incompetent.
You can legit sum $\sum_{k=2}^N\sin(kx)$
There's a formula
Right?
21:48
right
And you can see that it's bounded
Oh, wait, hold on
You're not asking how we can show that it's bounded, you're asking why it implied convergence of the series
@AkivaWeinberger That's where I find it easier to switch to complex numbers. Makes summing them easy as hell.
7 mins ago, by Leaky Nun
now we've upgraded our problem up one dimension.
@LeakyNun So?
nothing, just a cheesy remark
21:51
There's plenty of times were moving from $\mathbb{R}$ to $\mathbb{C}$ ends up making things really easy ^^
Or definite integrals with integrands that do not have a primitive in terms of elementary functions
@Akiva: Abel's theorem (summation by parts)
Or this question that was on the exam I had to supervise today: find $a$ such that the following integral vanishes:
$$\int_0^\infty x^3 e^{-ax} \sin(x) dx $$
@TedShifrin Not sure how that helps
21:58
$1/\ln k$ decreasing, partial sums of $\sin kx$ uniformly bounded
flips through Rudin
A-ha!
Theorem 3.42.
22:10
Hey there.
22:23
Hm, I wonder - if I know that $T$ is bounded, is it sufficient for me to conclude that if $ran(T)$ is closed, $ran(T^n)$ is also closed?
T being some linear operator.
15 hours ago, by Leaky Nun
I only do linear algebra if it is of finite dimension :p
Haha yeah
That solves many things
Hey, welcome back @TedShifrin
22:38
meow
Hi @Mike, long time no see
22:54
Can I do a quick sanity check?
@AkivaWeinberger just ask; don't ask to ask
The function $f:[0,1]\to\Bbb R$ defined by $\begin{cases}n,&x=1/n,~n\in\Bbb N\\0,&\text{otherwise}\end{cases}$ is not Riemann integrable, right?
Because any bit of the partition containing $0$ has a supremum of infinity
Does anyone here have access to the following: "Asterisque, Vol. 126 (1985); Géométrie des surfaces K3: modules et périodes"?
so the upper Riemann integral is infinity
@AkivaWeinberger you would benefit from asking literally anyone in this room other than me.
22:57
Pinging @anyone
hm?
I'mma go with no
oh I have no clue.
I think Riemann integrable functions have to be bounded, as a rule
23:15
It gets hard to define upper sums for something unbounded
Yeah
That's so frustrating, though, because the discontinuities are measure zero
so it feels like they shouldn't matter
Welcome to the realm of Lebesgue integration.
I assume that when the function is, indeed, bounded, they don't matter (for Riemann).
@SteamyRoot Sigh.
Why the sigh?
The idea is still "the surface under a curve can be approximated by a bunch of rectangles" - just in a smarter way
Yeah
I always knew that I would have to switch to Lebesgue for some integrals
Just... not like this

« first day (2515 days earlier)      last day (2804 days later) »