« first day (2422 days earlier)      last day (2617 days later) » 

12:00 AM
We want it so that when we add each corresponding power, each sum is prime
Weird.
 
I wish Laci was teaching algebra
 
Yeah @MeowMix idk if it's tractable to do something direct, the problem is actually linear algbera magic
idk if that's giving away too much
 
Oh hey @Ted
 
oh hey
 
Hey @Ted!
 
12:02 AM
lol
 
rehi Demonark
 
heya Ted
 
hi @Eric
 
The coefficients don't really matter.
 
Hiya chat.
 
12:03 AM
Hey @Fargle
 
Hi @Fargle
 
heya Fargle
 
How goes it everyone?
 
Goes it meh.
 
I'm sorry to hear, Yoda.
 
12:04 AM
It goes aight
 
How much longer do you have to be a redneck Texan, Demonark?
 
I take great offense to that, @Ted!
I say "aight" literally daily and I've never set foot in the Lone Star State.
 
You can take whatever you want, Fargle.
Oh, it had nothing to do with what he just said.
 
No it wasn't about "aight", I'm in Texas right now
 
Oh.
 
12:05 AM
lol
 
backpedals at relativistic speeds
 
I think Fargle owes Ted an apology.
Without any twin paradoxes.
 
I'm heading back on Sunday
 
Glad to have an unbiased opinion. :) Apologies, Ted.
 
Accepted.
 
12:06 AM
I think Zach needs to take a nap
Maybe not, perhaps I can be productive like last night
 
This is why you don't sleep well, Zach. Too many naps close to bedtime.
 
I sleep in the same room with two brothers who are snorers, so I have a half-excuse
 
Ear plugs?
 
OK, I won't take a nap then.
 
You know what might help? Look up the Hoffman-Singleton theorem
 
12:08 AM
Not like I make the rules or anything ...
 
It's total witchcraft
 
Hrm. I wish more radio music would play with time signatures.
 
@Ted I'm not saying you do either. :P
 
@TedShifrin wanna see an intriguing problem (not one I intend for you to solve, just an intriguing problem)?
 
@Fargle I was listening to a 5/4 song a bit ago
 
12:09 AM
Heya Duck
 
Might be something to left your spirits slightly
 
It wasn't on the radio or anything
 
hi
long time no see
 
@MeowMix Am doing so right now. Intervals is a good band.
 
What's the intriguing problem?
 
12:09 AM
@Ted I'm still thinking about Balarka's [modified] circle problem.
I don't see a way to solve it, and not even sure if it's true.
 
I dunno about that one.
 
nevermind. Stupid latex.
 
What was the problem, @MeowMix?
 
Show that for every $n$, there exists a circle where exactly $n$ lattice points lie on it.
 
You're still on this weird derivative jag, Duck.
 
12:11 AM
The regular problem is "contains", which is much easier
 
Just one letter, on versus in
 
@Fargle: By lattice, they mean the standard square lattice.
 
(The original problem works for any lattice of any dimension [where we consider $k$-spheres instead of circles], which is cool too.)
 
@TedShifrin Ah yes, I remember. That's hard.
 
@MeowMix Did you see what I wrote here?
 
12:12 AM
Yes
 
Indeed, the original problem works for any countable subset of any Euclidean space
 
Oh, yeah, that's right
Much more general.
 
@AkivaWeinberger There are n-balls that contain exactly $k$ points of $\Bbb Q^{n}$?
I find that surprising.
 
@Fargle Uhh, I feel like that's an exception
 
Have you guys ever had this weird feeling where you're in break and the topics you should be working on are orthogonal to what you feel like doing?
 
12:15 AM
…Not dense @Fargle
 
Ah.
 
Actually, let me think
 
What is the definition of "Dense"? I've heard it before, but I don't know an exact definition
 
A dense set is one whose closure is the whole space
 
@MeowMix A subset of a space is dense if every non-empty open set intersects it.
Wait, no.
Daminark has it.
 
12:16 AM
Yeah, I think "not dense" is all that's needed.
 
Or that, the two are equivalent
 
Oh yeah, duh.
 
what you said is fine, Fargle
 
I'm told mine works too.
Don't drink and math, kids.
 
disagree
 
12:17 AM
In fact, intersecting every non-empty open set is how we worked with them in order to do Baire Category
 
Actually, I'm not sure about "not dense". Stick to no limit points for now
 
@AkivaWeinberger Yeah, I was thinking it needed to be discrete.
 
Well, it does work for some spaces with limit points
 
Right, that's the word @Fargle
 
e.g. $\frac{1}{n}, n \in \Bbb N \cup \{0\}$
 
12:18 AM
Right, yeah
 
Maybe local compactness is part of it?
Because it also works for just the $\frac{1}{n}$s.
 
See, the thing is, for discrete sets, the problem is equivalent to a related statement. And the related statement is true for all countable sets.
 
Hi @Mike
 
The trouble is that the equivalence doesn't work for all countable sets.
 
12:21 AM
You lost me, DogAteMy, but that's OK.
 
@Ted Do I need to check a GPS to find you?
 
Too many pronouns I guess
 
braces for smack
 
None of those are pronouns, actually
 
probably wouldn't help, Zach
 
12:22 AM
Actually, @Ted has been rather smack-free as of late. Maybe I just haven't witnessed any.
 
better ask Demonark about that
 
@MeowMix Speak for yourself.
 
Are we doing anything fun?
 
He hit me with a yard-stick the other day. A yard-stick
 
OH, I remember now
You got imperially ruled
 
12:23 AM
I proved that the proposed counterexample I had to the "two cosimply connected sets have cosimply connected union" thing works
And I say "cosimply connected" because it's shorter than the alternative
even though it's not a type of connectedness
 
I was gonna ask what it meant, but I guessed
 
"space with simply connected complement"?
 
> CAN Richard Funk
 
Are spaces with connected complement called nnected spaces
 
12:25 AM
glares at DogAteMy
 
@AkivaWeinberger As long as we can have mpact and mplete spaces as well.
 
And untable
 
lol
mpletely normal
 
what is even happening
 
I had fun with the nvergent topology on $\{\frac1n\}$ a while ago
 
12:26 AM
Don't even ask, Eric.
 
no one ows.
 
I tab in for a second and suddenly everyone is speaking gibberish
 
Category theory: every joke is really two jokes!
 
@Eric To a mathematician, a coconut is just a nut
 
lol ok that one is pretty funny
 
12:27 AM
But double duality sometimes fails. So all of DogAteMy's jokes are reflexive, I guess.
 
It would make coco puffs a much more confusing cereal.
 
Hm. Better terminology: simply coconnected
Wait, that sounds too much like coconutted
which isn't necessarily a bad thing
 
To say nothing of Coco Chanel.
 
So that's why we just call her Chanel!
 
Right.
 
12:29 AM
There's a Math SE question on cocohomology somewhere. I forget what the verdict was, though
 
So a ntinuous function is a bijection whose inverse is continuous?
 
I think they decided it was different from homology
@Fargle I think there are too many math words that start with "co"
 
@AkivaWeinberger Rrect.
 
7
Q: Is co-cohomology the same as homology?

sifsaSuppose I have a chain complex of chains $C_n$. Then one can obtain the homology groups of this complex. Now if I choose any abelian group $G$ and I consider the cochain group $C_n^*=Hom(C_n,G)$ then I can obtain the cohomology groups. Now the question is: If I form the cocohomology group by cons...

 
Crazy for cocohomology
puffs
 
12:30 AM
Thirty second unskippable ads...
 
"Chain complexes in every bite!"
 
I like the ability to say ${\rm trig}(\pi/2-x)={\rm cotrig}(x)$
 
What's a nice place for a mathematician to have a vacation? Cocoa beach.
 
lol
 
I actually did that officially in some notes, DogAteMy. In proving the mnemonic for derivatives of cofunctions.
 
12:32 AM
And $\rm(cotrig)'=-co(trig')$ as well
Right
 
A coconut is just a nut
 
No, DogAteMy.
Don't forget chain rule.
 
Also, if we're going more advanced: A comathematician is a device for turning cotheorems into ffee.
 
Look, I knew what I meant…
:P
 
12:33 AM
You fool yourself frequently, DogAteMy.
Not usually mpletely, but ...
 
Related: Is uninformed just formed
 
What about ununiformed?
This is getting to be worse than Hebrew.
 
Unununium
 
@AkivaWeinberger Is there a natural isomorphism between incorrigibility and coinrrigibility?
 
Uncountable -> Ununtable -> Table
 
12:36 AM
Suppose that for $n \geq 1$, $X_n$ is uniformly distributed
on {1, 2, ..., n}. how to show
$\lim\limits_{n\to\infty}P(\frac{X_n}{n}\leq y)= y$ for $y\in(0, 1)$.
 
oh thank god actual math.
 
ROFL
Oh, Astyx had a good physics question we needed you for, @Semiclassic. Did you see my ping?
 
When did double negation start cancelling out? I feel like it was fairly recent for english
 
Saw it, forgot to look at it.
Lemme find it again.
 
He was interested in stability of a physical system and which forces needed to be considered.
 
12:37 AM
@Eric It doesn't quite cancel out always, based on context. "Not impossible" has a very different meaning in context from "possible".
 
@Eric It would probably be different times for different dialects (if I understand your question correctly)
 
Well, there seem like two issues.
1) Inertial v. non-inertial forces.
2) Velocity-dependent forces.
 
sure, but most of the time when there is cancellation from double negation it's coming from an expression that's an early hold out from old or middle english, where there is never any cancellation of double negatives
when there isnt cancellation*
 
I consider stuff like Bernoulli's principle to be in the latter category.
 
12:39 AM
@Semiclassical do you know how to use riemann sum to the question I just posted?
 
You shouldn't need to use a Riemann sum.
Actually, I may be wrong. Riemann sum does seem to be a sensible approach due to the limit.
 
The problem ask to use it, the idea I have is $\lim\limits_{n\to\infty}P\left(\frac{X_n}{n}\leq y\right)=\lim\limits_{n\to\infty}P\left(X_n\leq ny\right)=\lim\limits_{n\to\infty}\frac{\lfloor{ny}\rfloor}{n}=y$
 
@MeowMix Want a topology puzzle?
 
Sure
 
which you may or may not be ready for, but whatever
 
12:41 AM
I don't see why Riemann sums are needed, but I truly don't want to think about it.
 
I'm afraid I'm in that stage of mind right now as well.
 
A space is called "homogenous" if, for every pair of points $p$ and $q$, there's a homeomorphism $f$ from the space to itself such that $f(p )=q$.
Essentially, all the points "look the same".
 
@Simple: That looks right.
 
The shape of the letter T is not homogenous, for example.
 
You might need a bit of math to justify the very last step (the limit).
 
12:43 AM
It might be that the Riemann sum might be the most convenient way to justify the last step?
 
Nah.
 
@TedShifrin sandwich theorem
 
Precisely.
 
Well nuts.
 
@MeowMix Wait - do you know what a manifold is?
 
12:43 AM
I don't know a definition, but I have an idea
 
You define two charts to be compatible if...
 
If Zach ever gets through my book, he'll know.
 
Like surfaces, but possibly with lots of dimensions
@Daminark A topological manifold
 
You don't need to define abstract manifolds, guys, good grief.
 
I kid, I kid
 
12:44 AM
Let him think about a submanifold of Euclidean space.
 
Am I thinking of the wrong word?
I just want a space where every point has an open set around it homeomorphic to $\Bbb R^n$
 
That's a manifold.
 
@AkivaWeinberger Yup.
 
Like a sphere, or a torus, or a line, or $\Bbb R^n$ itself
 
LOL this guy pushed off his HDD platters while it was spinning youtu.be/fJ3lGKLDHk4?t=2m12s
 
12:45 AM
Oh we're dealing with topological manifolds, so yeah second countable/Hausdorff (can be done away with) and locally homeomorphic to $\mathbb{R}^n$
 
Alright
So it's like
if you zoom into it close enough, it looks like $\Bbb R^n$, right?
 
Informally, sure.
 
Yeah
So the question is, are all connected manifolds homogenous
 
Or an open subset of $\Bbb R^n$, yeah.
 
(and now I'm realizing that my proof might miss a subtlety?)
 
12:46 AM
I can prove that easily for smooth.
 
(so I'm not quite sure I know how to prove it, right now)
 
I'm gonna do some Ted-cercises, but I'll think about it
 
Perhaps "easily" is a misstatement.
 
@AkivaWeinberger Oooh, this is hard. Any counterexample I can think of seems like it would cease to be a manifold.
 
What's a homogeneous space?
 
12:47 AM
Clearly they need to be connected, since, like, the disjoint union of a sphere and a torus isn't homogenous
5 mins ago, by Akiva Weinberger
A space is called "homogenous" if, for every pair of points $p$ and $q$, there's a homeomorphism $f$ from the space to itself such that $f(p )=q$.
All points look the same @Daminark
 
Oh right... My attention span is like, $\mathbb{R}$ bad
 
OK, I need to eat dinner and go play bridge. Y'all misbehave without me.
 
@AkivaWeinberger Then again, it's debatable whether you'd call that a true manifold.
 
Wait is there a whitney type theorem for topological manifolds? (non-compact)
 
Yeah, Eric, there is.
 
12:48 AM
oh sweet
 
Often manifolds are defined so that the dimension to which it is locally homeomorphic is the same for every point.
 
Bye, all.
 
Bye @Ted!
 
I still don't know how to use Riemann sum to show limit, does $\lim\limits_{n\to\infty}P(X_n/n\leq y)=\sum_{k:k/n\leq y}\frac{1}{n}$?
 
Bye @Ted
 
12:49 AM
See you around @Ted!
@Eric Can every manifold be embedded into a higher dimensional analog of Schlag's surface?
 
Schlag's surface is diffeomorphic to $\mathbf{R}^{2}$ dude.
higher dimensional analogues are just $\mathbf{R}^{n}$
 
@MeowMix So apparently it's important to specify that the definition of manifold should include them being Hausdorff
because there's this really weird space called the "line with two origins"
 
@AkivaWeinberger Ohhhhh yeah.
 
That is true
 
so the answer is yes
 
12:51 AM
Yeah @Akiva I've seen in Lee's book that it's defined to also be second-countable
The idea is that you want unique limits, and smth called "partitions of unity"
 
Essentially, it's $\Bbb R\cup\{0'\}$, where the open sets are either the open sets of $\Bbb R$, or the open sets of $\Bbb R$ with $0$ replaced by $0'$
 
@Akiva: that space fails to be homogeneous because $0$ and $0'$ aren't equivalent to every other point?
 
(or the open sets of $\Bbb R$ with $\{0\}$ replaced by $\{0,0'\}$, I think)
@Fargle Yeah
So, in that space, the sequence $1,1/2,1/3,1/4,\dots$ has two limits, $0$ and $0'$.
So, this is not a manifold, and that's because we're supposed to include "Hausdorff" in the definition
(The line with two origins is not Hausdorff)
 
@Daminark wait Schlag surfaces are like hyperbolic paraboloids right
 
@Daminark Oh. God. Is the counterexample for this the long line?
 
12:55 AM
Basically @Eric
 
@AkivaWeinberger This is just $(\Bbb R \times \{a\}) \cup (\Bbb R \times \{b\}) / \sim$ where $(x,a) \sim (x,b)$ when $x \neq 0$, right?
 
That's homogenous, though, so I think I can safely ignore this @Daminark
 
@Akiva I have no idea what partitions of unity are, I just know they're a thing that exist and are important if you want to translate integration to manifolds, so you tend to ask that the topology is second countable
 
@Fargle Yeah.
 
Sorry for the edit-ping spam, couldn't remember how to typeset $\sim$.
 
12:56 AM
Crush two lines together except at the origin.
 
oh yeah so if you want the higher dim analogue to be like, simply connected, complete, and with everywhere negative sectional curvature you still have a diffeomorphism with $\mathbf{R}^{n}$.
 
This also gives us an example of a path-connected but not arc-connected space.
(Arcs are injective. What arc connects the two origins?)
 
Partitions of unity are a collection of smooth functions compactly supported inside the sets of an open cover that sum everywhere to one
they're used to smooth locally defined objects into global ones
 
@AkivaWeinberger So is the (or a) path just $[0,0.5) \mapsto 0$, $[0.5,1] \mapsto 0'$?
 
Dank
 
12:59 AM
That's not continuous. (The inverse image of $(-1,1)\cup\{0'\}\setminus\{0\}$ is not open.)
You need to leave the origin for a bit and come back.
 

« first day (2422 days earlier)      last day (2617 days later) »