« first day (1757 days earlier)      last day (3265 days later) » 

6:14 AM
Well, I'm just reading bits of Munkres right now when I'm free, sort of following this.
Aaaand Balarka is here to shout at me
 
Hello, @Kaj.
 
Hey @BalarkaSen
I've been reviewing some linear algebra recently via MSE
 
Yes, reviewing things via answering questions is a good idea.
 
Trying to teach other people is probably the best way for you to discover what you do and don't have a good grasp of.
 
Yeah. I have been trying to do that with algebraic topology.
I am also learning a lot from other people!
 
6:18 AM
I hope I do well in my algebra course next semester so I can potentially take algebraic topology in the spring. Then we'll have much to talk about.
 
What'd you get done since morning, @BalarkaSen?
 
computed fundamental group of wacky linear groups.
that uses both algebraic topology and some linear algebra (Gram-Schmidt and whatnot), so those are good exercises.
found them at some qual test papers.
 
@BalarkaSen great, thank you
 
no problem.
 
that clears up part of my concerns
 
6:22 AM
@iwriteonbananas for the cone thingy you have posted, I guess all you need is the homology long exact sequence for triples.
you need to exploit the fact that $H_n f$ is an isomorphism by saying "kernel of this map and image of this map is zero" instead of looking at $H_n(pt)$s.
 
but we still dont really know the map $(i_*, j_*) : H_1(S^1 \sqcup S^1) \to H_1(S^1\times I) \oplus H_1(S^1\times I)$ yet
 
but of course, I haven't tried it.
 
@BalarkaSen okay, i'll try it again after i finish this
 
@iwriteonbananas we do, don't we?
 
what is it? why cant it be trivial?
 
6:25 AM
the maps $i_*, j_* : H_1(S^1 \sqcup S^1) \to H_1(S^1 \times [0, 1])$ are both isomorphisms, aren't they?
 
why is that?
 
user147690
What does $[\Bbb Z/(n)][x]$ refer @Balarka? Polynomials taking values in $\Bbb Z_p$ even when $p$ is not prime, or should I assume we are talking about solely when this is a field(p prime)
 
$n$ is not prime. polynomials with values in Z/nZ
 
user147690
Okay thanks
 
Mathematical troll: use "p" in your work, but make a note that "p is not necessarily prime".
 
user147690
6:28 AM
Hahaha
 
"Let $\epsilon$ be a large number . . . "
 
whoops, @iwriteonbananas, not an isomorphism.
 
user147690
Let $\epsilon \lt 0,\epsilon\in\Bbb R$
 
Then later on, say "but now let $n$ be a prime. Then $\mathbb{Z}_p$ is not necessarily a field, but $\mathbb{Z}_n$ is." or some such thing.
 
user147690
Lmao
 
6:31 AM
"For all $\epsilon$ sufficiently large, <result>." :P
 
@BalarkaSen $i_*$ should take the generators of both circles and map them to the same thing
 
yes, that's why i said "not an isomorphism" :P $\Bbb Z \oplus \Bbb Z \not \cong \Bbb Z$
 
:P
but i dont know how to prove that that's what $i_*$ actually does.
 
well, we've just proved that
 
6:34 AM
enlighten me
 
the inclusion maps $S^1 \times \{1\} \to S^1 \times I$ and $S^1 \times \{0\} \to S^1 \times I$ induce the same map in arbitrary homology...
 
proof that fact doesn't need any singular/simplicial theory
 
yeah, we've proved that
but why cant the induced maps be trivial for example?
 
induced from which map?
 
6:40 AM
the inclusion $S^1 \times \{1\} \to S^1 \times I$
 
OK, assume $H_\bullet(i_*)$ is trivial. Compose that with the isomorphism $H_\bullet(S^1 \times I) \cong H_1(S^1)$. It's still trivial.
 
ugh, I keep thinking about cycles in torus.
 
me too
 
Damn you, Munkres
 
6:51 AM
@iwriteonbananas I am not sure what kind of wacky self-map $S^1 \to S^1$ wouldn't be nullhomotopic, yet would be zero on homology.
I guess I need some coffee.
 
@BalarkaSen yeah, on the other hand we dont really know much about the $H_n$ functors
 
Why does Munkres say that a topology on a set $X$ must be closed under any union of the subsets of $X$ in it, but then say that it must be closed under finite intersections only?
 
@SohamChowdhury do you know the definition of a topology?
 
That is the definition.
 
exactly
 
6:54 AM
no point in questioning the definition
 
Is there any desirable property that we lose without the "finite"? That's my question.
Is it impossible to intersect an infinite number of subsets or something?
 
the topology on $\Bbb R$ is no longer a topology. These definitions are inspired by metric spaces. It will be hard to understand the motivation for anything if you don't know much about metric spaces.
 
Which comes a while later.
Okay.
 
That's why I prefer Simmons.
 
He does metric spaces first?
 
6:57 AM
Yep.
 
Should I use Simmons instead?
 
Definitely.
 
(in your opinion)
OK.
 
i was typing a question to mike but then he left
he must've known
 
ping him again :P
 
6:59 AM
I actually did do a bit of Simmons earlier, because finding a pdf of Munkres without weird lines through the pages was hard. I got as far as Schroeder-Bernstein (and I typed it up in LaTeX for fun).
Okay, that's what I'll do then.
 
you did Simmons but you didn't know it starts with metric spaces?
 
@MikeMiller care to tell me why the map induced by the inclusion $S_1 \sqcup S^1 \to S^1 \times I$ is not trivial? it should be the map $R \oplus R \to R, (a,b) \mapsto a+b$
 
It was quite a while back.
I didn't remember.
I looked at a few topology books that day (because, apart from Munkres, there are few "definitive" recommendations). I'd been wanting to understand Schroeder-Bernstein for a while, so I did the proof and forgot about the book.
 
ok, I see.
 
Anyway, I gotta run, as you say.
 
7:02 AM
bah, @iwriteonbananas, no good questions on the mains.
 
i might ask this question on main, it's really annoying me
i feel like it should be clear and im just being stoopid lol
 
ok, let me have a serious try at it first.
 
that's how I always feel about my analysis homework
 
or, well, I did
I guess I finished it now :D
I'll get the questions that feel like they should be hard, and trip myself up over the details of the "easy" questions
 
7:11 AM
hmm ok
i guess the map $i_*$ is the same as the composition $$H_1(S^1 \times I) \to H_1(S^1 \times \{1/2\}) \to H_1( S^1 \times \{0\} \cup S^1 \times \{1\})$$ where the 1st map is an isomorphism and the 2nd map is $(id, id)$
 
no, the other way
 
I agree it's the map $(a,b) \mapsto a+b$. What's the issue?
 
proving that the inclusion map $S^1 \to S^1 \times [0, 1]$ isn't zero on arbitrary homology theory.
 
@MikeMiller i deduced that using cycles, but we're in an arbitrary ordinary homology theory now
@BalarkaSen what do you mean?
 
Each map $S^1 \times \{pt\} \to S \times I$ is a homotopy equivalence, hence induces an isomorphism on homology. The map from the disjoint union is the direct sum of those maps.
 
7:18 AM
oh right.
bah.
 
right
@MikeMiller i think that's pretty much what i wrote right before you came back
 
What more is there to say?
 
nothing, nevermind me
 
OK.
 
not my best day, I guess.
 
7:20 AM
+1
 
your map $H_1(A \cap B) \to H_1(A) \oplus H_1(B)$ is just a linear extension of $(1, 0) \mapsto (1, 1)$ and $(0, 1) \mapsto (1, 1)$.
 
yes, exactly
so, it has kernel and image $R$
 
yeah.
and the rest is easy
 
I'm gonna try to prove your cone thingy now.
 
7:23 AM
cool, i'm gonna work on that after i finish writing this one down
 
7:34 AM
@iwriteonbananas are you sure nothing about $X$ being a subspace of $Y$ is said anywhere?
'cause if it's so then it's easy to derive the result
 
@BalarkaSen it doesn't say that anywhere. how did you derive the result in that case?
 
long exact sequence. $Cf \simeq Y/X$.
i.e., you can contract loops in $X$ by sliding to the top of the cone.
 
hmm ok
right.
 
i'm trying to throw a bunch of general nonsense (naturality, long exact sequence of pairs, triples, etc) at it.
working with abstract homology theory is not fun :(
 
it's a bit nasty, i prefer more geometric things
the fact that the homology theory doesnt satisfy dimension axiom is annoying
 
7:42 AM
i guess all this will pay off when you do bordism homology.
or k-theory
 
...which probably wont be in the near future
have you tried using MV w/ the A= CX, B=Y?
$... \to H_n(A\cap B) \to H_n(A) \oplus H_n(B) \to H_n(X) \to H_{n-1}(A\cap B) \to...$
 
doesn't help
 
$A\cap B$ is $X$
 
@BalarkaSen, it's "abstract nonsense", afaik
 
people call it general nonsense too
 
7:48 AM
and the 1st map is (inclusion, f)
unfortunatly, we cant conclude that $H_n(A) = 0$ from the fact that $CX$ is contractible
 
right
 
i meant to right $C_f$ in the above sequence instead of $X$...just noticed
 
Simmons really should say [surje|bije|inj]ective, "one-to-one and onto" gets tiring after a while.
I sort of like the book, though, @Balarka. Thanks for the reco.
[sur|bi|in]jective*, sorry.
 
8:09 AM
@Soham most mathematicians say surj/inj/bijective
For some reason, "one-to-one" and "onto" are seen as more "intuitive" terms for beginners, and so they've gained popularity in intro-to-proofs courses and various texts, and they've caught on a bit from there.
 
For everyone who's interested in the Sage CAS: the Sage Proposal at Area 51 needs some more questions with 10 up-votes before moving to the next phase. (If that doesn't happen before june 1st, the proposal will be closed...)
4
 
@BalarkaSen can we say anything about the image or kernel or the map induced by the inclusion $x_0 \to C_f$?
 
@KajHansen it's an old book, so I understand. My school textbooks also do that.
@BalarkaSen Mmm, lots of nice set theory proofs I'd forgotten are here. :)
 
8:36 AM
First chapter of Simmons over. Metric spaces appear now. That's why I didn't know.
 
user147690
8:55 AM
Many of my old highschool friends are getting married and/or having/have-had a kid. This seems very strange to me. Pretty much people either went to uni or the above happened. Is this the same in America??
 
@BalarkaSen does Simmons treat the "quotient topology" (whatever that is) anywhere in Part 1?
 
9:12 AM
@SohamChowdhury This doesn't help you much, but I know Munkres treats it fairly early.
 
9:25 AM
I just checked and see that Chris's sis has not been here for 3 days and r9m has not been here for 5 days. Did something happen? I was gone from last Friday.
 
@robjohn Both did visit MSE recently, see here and here.
 
@barto I am talking about chat.
 
@robjohn I know.
 
@barto I did see that Chris's sis was active on main within 30 minutes ago. However, she had been pretty consistently on chat, too. So I was curious.
 
@robjohn: look at the older starred messages. @AlexC reckons his one from a few days back might've put her off or something.
 
9:42 AM
@SohamChowdhury You mean this one, I assume.
 
10:02 AM
Hello@SohamChowdhury
Doing Simmons....
Nice😃
 
10:17 AM
@robjohn Yes, he thinks so, at least.
 
10:49 AM
Hey @barto, I contributed a few questions to the Sage.SE proposal a while back.
Is $d(x,y) = |\{x\} - \{y\}|$ a pseudometric?
$\{x\}$ is the fractional part of $x$.
 
Ahh now its fine
 
Of course it's not a metric.
@Rem? Any idea? I'm just not sure about the triangle inequality here.
 
Yes... So what have you got till know
 
Where have you reached in the problem
 
10:53 AM
Well, I just have to check the three requirements.
 
Yes
 
I'm not sure about the triangle inequality. Does it hold for this function?
The first two are easy to prove correct.
 
First two are trivial stuff
 
Yeah.
So, is $|\{x\} - \{y\}| \leq |\{x\} - \{z\}| + |\{y\} - \{z\}|$?
Hey, Balarka.
 
Hey,Balarka
 
10:55 AM
I think it is.
Because $|\{x\} - \{y\}| = |\{x\} - \{z\} + \{z\} - \{y\} | \leq |\{x\} - \{z\}| + |\{y\} - \{z\}|$, does that look correct?
 
$d⁢((x_1,x_2),(y_1,y_2))=|x_1-y_1|$. I think this is a good psuedo metric
 
hello, @Soham, @Remember
 
Yes, but your metric is $\mathbb{R}^2 \times \mathbb{R}^2 \rightarrow \mathbb{R}$.
Mine is simply $\mathbb{R} \times \mathbb{R} \rightarrow \mathbb{R}$.
 
Yes it is so any probs
 
@SohamChowdhury I don't recall : don't think it does.
 
10:59 AM
No, just defined on a different metric space.
 
Quotient topology is better done from Armstrong.
 
Yes so its a bit more fancy.. :p@Soham
 
@BalarkaSen Basic Topology?
 
yeah.
 
@Rememberme It's the wrong space, duh.
 
11:00 AM
Whats the space
Given to you
 
I edited, look up.
@BalarkaSen What did you do all day?
 
Oh okay
Let me think of your example then
 
@Balarka, any help with my question?
 
Today I really feel like kicking CBSE
 
4 mins ago, by Soham Chowdhury
Because $|\{x\} - \{y\}| = |\{x\} - \{z\} + \{z\} - \{y\} | \leq |\{x\} - \{z\}| + |\{y\} - \{z\}|$, does that look correct?
 
11:02 AM
@SohamChowdhury nothing much other than computing fundamental group of linear groups.
had some schoolworks to do.
what do you want to prove?
 
I should really get started with my EVS project, btw.
That inequality. Is that correct?
$\{x\}$ is the fractional part.
 
He wants to find an example of a pseudo metric which is not a metric @Balarka
 
there are loads of them
 
No, I want to know if mine is correct. I already know an example.
 
Yup
 
11:04 AM
what's your metric?
 
Difference of fractional parts.
 
that works.
congratulations.
 
We have the trivial psuedo metric also
 
Mmm, thanks. Wanted to see how wacky one could be :P
 
We have some others which I cant remember now
 
11:06 AM
For the set of functions from $\mathbb{R} \rightarrow \mathbb{R}$, $d(f, f') = f(0) - f'(0)$ works.
Not a metric because $|\sin, \sin^2| = 0.$
 
Yup .....thats one I think so
Just sin??
what about the x :p
 
$\sin$ is the function.
 
Yes but bring the x
 
brings x
 
dan dan da da da da dan dan da....
 
11:08 AM
@BalarkaSen does this weird metric confer any structure on the space?
Also, aren't the p-adics defined by a sort of metric?
Hey @AlexC
 
user147690
Are you two still obsessing over doing topology? :P
 
yes
 
No, I'm doing topology.
 
Hey hey @AlexC
 
user147690
11:09 AM
@SohamChowdhury :P
 
user147690
Hey @Soham @Rem
 
@BalarkaSen: "These kids, they never learn."
@BalarkaSen Was that for me?
 
@SohamChowdhury metric closure of $\varprojlim \Bbb Z/p^i\Bbb Z$ (topology being the subspace topology from the product topology on $\prod \Bbb Z/p^i\Bbb Z$) is homeomorphic to $\Bbb Z$ with the $p$-adic metric.
yes, it was.
 
What of my previous question?
 
11:10 AM
yes
p-adic thingy
 
no, before that.
 
it sounds pretty nonsensical.
 
does every (pseudo)metric confer some structure on a set?
oh.
i guess so.
 
define "structure"
 
Start defining @Soham:p
 
11:12 AM
@BalarkaSen goodness.
 
and having structure on a set doesn't make sense. a set is just a set. what makes sense is to say "having structure over some set"
 
mea culpa
 
@AlexClark How's it going?
 
anyway, i'll get back to work.
 
Simmons?
The last question is nice
Of the chapter which you are doing
 
11:13 AM
@BalarkaSen "confer on" is a phrasal verb.
 
Have a go at it@Soham
 
Chapter, or section?
 
user147690
@BalarkaSen Somewhat alright. Still having trouble getting my head straight on all of these things, PIDS, UFDs, EDs, rings, principal/maximal/prime ideals
 
I suppose the question which you just gave an example about... Let me check once @Soham
 
@BalarkaSen i've never heard of "open spheres". Does Simmons mean "open ball"?
 
11:16 AM
@Soham $\mathcal{C}([0, 1])$ be the set of all continuous functions from $[0, 1]$ to $\Bbb R$. Prove that this a metric space equipped with $d(f, g) = ||f - g||$, where $||f|| = \sup_{x \in [0, 1]}\{|f(x)|\} < \infty$ (prove that it even make sense).
yes, open spheres are the same as open balls
 
@BalarkaSen Is that an exercise also for me ?
 
What's the $<\infty$ for?
 
sure, you can try if you want.
 
I don't understand the definition of your norm.
 
@SohamChowdhury I saw; thanks for the support!
 
11:18 AM
@Soham just to emphasize that $||f||$ is not infinity. you have to prove it, though!
 
(this is really an exercise in analysis)
 
I understood.
So it's out for me, I guess?
I'll have a go at it anyhow.
 
Assume $||f||$ exists and prove that it's a metric if you don't want to think about the analytical bit.
But the proof of $||f|| < \infty$ is fun, though!
 
I suppose the analytical bit follows from continuity?
 
11:20 AM
it's nontrivial.
"follows from continuity" is not a proof
 
i know, but is that a significant part of the proof?
 
'course it is. great, now you've found yourself another exercise : find a discontinuous function $f : [0, 1] \to \Bbb R$ for which $||f|| = \infty$
this one's easy-peasy though.
 
I think symmetry and triangle inequality are both trivial for this metric.
Because the norms are always reals.
$d(f,f') = 0 \Leftrightarrow f = f'$ is hard
 
I was thinking the same thing @Soham
 
no, that is very very trivial.
and why is triangle ineq trivial? prove it.
 
11:25 AM
Ah norm of something is 0 iff its the 0 vector Is that what you mean @Balarka
 
it's definitely not hard, but it's underestimation to say that it's trivial.
 
Well, $||f||$ is always a real, and the triangle ineq is easy for reals.
 
That's not a proof of anything, @Soham :P
 
$|x - y| = |x - z + z - y| \leq |x - z| + |z - y|$?
 
So what?
 
11:26 AM
Case analysis required there.
1. "If all are positive . . . "
2. "If one is positive, wlog, let it be x . . . "
3. If two are positive, negate them and see 2.
4. If all are negative, negate them and see 1.
 
Okay I have got to run Bye!!
 
You have to prove that $\sup_{x \in [0, 1]}\{|f-g|\} \leq \sup_{x \in [0, 1]}\{|f - h|\} + \sup_{x \in [0, 1]}\{|f - h|\}$. To do that you have to prove a simple lemma about sup.
@SohamChowdhury huh?
 
Wait, I think I got something terribly wrong.
 
do you even know what sup is?
 
Sorry, I thought something else.
Yes, I do.
 
11:28 AM
hey sup
 
Well last thing balarka doesnt the triangle inequality imply from what we know as the definition of norms
 
I thought $d(x,y) = |\;||x| - ||y||\;|$.
 
you have to prove that $||f||$ is a norm, @Remember. you can't just say "it's a norm" and sniggle out of the proof.
 
Greetings
@robjohn hi and welcome back. I was out since both you and @r9m weren't around these days.
 
eh, no, I have just defined $||f||$ up there, @Soham
 
11:30 AM
Here, have a go at this:
$$\int_1^{\infty} x^{-\Gamma(x)} \text{d}x$$
1.04 something.
ISC fails on this.
 
@SohamChowdhury Does it have a closed form?
 
I don't know.
I hope it does.
 
@SohamChowdhury I don't know either (at first sight).
 
Try to give it a go when you have nothing to do. :)
@BalarkaSen Yeah, I'll try.
 
@SohamChowdhury These days I found some amazing ways of computing integrals that are incredibly hard.
To give you an example
$$\int_0^1 \frac{\log^4(1+x)\log(1-x)}{x} \ dx$$
 
11:33 AM
once you do a bit more on metric spaces, I'll tell you about the p-adic metric, @Soham :)
 
@SohamChowdhury or $$\int_0^1 \frac{\log^7(1+x)\log(1-x)}{x} \ dx$$
 
@BalarkaSen Hm. As of now, I won't try to solve your problem. My memories of sup are closely linked with Rudin's first chapter. :P
 
@robjohn thanks for asking. I also stayed away from chat since I'm not in a great mood either.
 
@BalarkaSen Isn't it just that two numbers are close by if their difference is divisible by a large (or small?) power of p?
Or something similar.
But I'd love to learn more when I'm done with this chapter.
I won't even bother contemplating those, it's just unimaginable :P
You can post the answer, if you like, though!
I'll get back to my book for now.
 
@SohamChowdhury I plan to do that in my book, but not sure I'll include the version with $7$ as a power.
 
11:37 AM
@SohamChowdhury yes, but I don't think just knowing the definition lets one have any intuition about p-adic numbers.
 
@robjohn I send you something in db.
 
speaking of which, I wonder if @AlexClark is studying anything about inverse limits :P
 
Hey, can't I have an open ball with radius zero?
 
no.
a ball with radius zero by definition is a point
and a point is closed
 
@robjohn Sent.
 
11:41 AM
Hm.
 
user147690
@BalarkaSen I am going to be destroyed when I am meant to present at this rate surely, but it is worth more of my grade to get the assignment done
 
@SohamChowdhury Because every distance is nonnegative, the set of points at a distance strictly less than 0 from a given point would be empty.
 
@AlexClark right. you'll learn a lot by doing your assignments, too, as you seem to have some nice exercises there.
 
user147690
@BalarkaSen Hopefully some will be done by tomorrow lmao
 
So as you wish, an open ball with radius 0 is empty
 
11:44 AM
depends on how you define an open ball with radius 0
the definition I have in mind involves a choice of point, in which case it's a single point.
 
i believe an open ball of radius 0 should be the empty set
 
I don't.
Every point is distance zero from itself, so it should be part of the set.
Dunno if a single point is open yet.
@Balarka says no.
 
I am not discussing this unless somebody comes up with a definition.
@Soham Prove that a single point is closed.
 
a single point is a closed set as a metric space is hausdorff
 
11:54 AM
If we take the definition from Wikipedia, an open ball with radius 0 would be empty
But come on, this is not worth the discussion ;)
 
best thing i heard all day
:P
 
@SohamChowdhury Balarka said a single point is closed, but he didn't insinuate that it can not be open ;-) (it could be both)
 
@BalarkaSen For a space $P = \{p\}$, there must exist an open ball centered on p with radius > 0 for the space to be open, so there must be points in $P$ such that $d(x,p) > 0$, but there is no such x.
Correct?
 
if we have a metric space such that all point sets are open, then the space is in the discrete topology and every set is open and closed
 

« first day (1757 days earlier)      last day (3265 days later) »