« first day (2428 days earlier)      last day (2596 days later) » 

7:00 PM
I think the way I was thinking about it was in a similar way with homotopy groups
 
So I have to use this to show $\gcd(a/\gcd(a,n),n)=1$ ?
 
I am seeing concordance as currently the best way to form a group with "isotopy" classes
 
You can think of a concordance as a path between two knots, sure
 
AGH. You already decided that was false.
Let's keep going the way we're going.
 
False for $\mathbb Z$, but I thought it would be true in $\mathbb Z_n$ :p but alright
 
7:01 PM
Can it be thought of as an invariant of the space that the knots live in?
 
Do you know the Euclidean algorithm and what it tells you about the gcd?
 
A concordance is not a path between knots; that's an isotopy
 
and are there similar extensions to "higher" groups
 
@AliCaglayan The space of knots is Emb(S^1, S^3)
a path there is an isotopy, like Mike said
 
yes, it's the smallest positive value for which we can write $a$ and $n$ as a linear combination
 
7:02 PM
I dunno if it's a path in a literal sense
 
@BalarkaSen but that doesn't really have a nice group structure to it
you can weaken your topy from isotopy to concordance and it becomes a group
 
So that means you can write $k= sa + tn$ for some integers $s$ and $t$, right, @ShaV?
 
yes
 
What does that equation say when you bar it? :) i.e., look in $\Bbb Z_n$?
 
wait
 
7:04 PM
Who're the same?
 
they have the same remainder after division by $n$?
wait
 
In particular, it says $\bar k = \bar s \cdot \bar a$.
 
ofc:l
wait but
we're working in $\mathbb Z_n$
 
afaik there isn't really a good way to compose isotopies to give another?
 
how can we use multiplication of elements in $\mathbb Z_n$
 
7:06 PM
but that seems to be the case for concordances
 
I thought that was not defined
 
You haven't learned that yet?
 
or
Can't you say that we have:
 
Sure you can compose isotopies...?
 
Well, fine, you can do $\bar k = s\cdot \bar a$, where that means repeated addition.
 
7:06 PM
$\overline k=s\overline a=\overline{sa}$ ?
 
@BalarkaSen it won't be an isotopy though?
 
yes alright
 
Why not? Isotopy is a sequence of embeddings. Concatenate two sequence of embeddings
I don't see the issue
 
Now go back to your supposing that $\bar a$ had order $\ell<k$, @ShaV. What do you get?
 
You don't have inverse knots to connected sum until you mod out by concordance
 
7:08 PM
contradiction?
 
Well if I have two embeddings of S^1 in S^3 they can't really form another?
 
Ah, so the operation is connected sum not what Ali is saying
 
because $sa$ is a multiple of $a$
 
yeah
 
so they have the same order, and we've shown that $k$ has the same order as $a$
or $sa$ is a multiple of $k$, so you wish
 
7:08 PM
Yeah, I apologise I shouldn't of used compose
 
So $\ell \bar k = $ what?
 
I think you should probably start with a more general knot theory book
Rolfsen is a classic, and Lickorish is popular
 
So what I am trying to get at is: Is concordance the strongest we can go without losing the group structure?
 
ShaV: We are trying to show that $\bar k$ has the same order as $\bar a$. We're almost there. Do NOT assume it. So you had assumed $0<\ell<k$ and $\ell\bar a = \bar 0$, right?
 
@TedShifrin I don't even know what we're trying to prove :( Could you just tell me that, then I'll reread our conversation, because right now I have no single idea what I've done at all
 
7:11 PM
LOL ... I read your mind. :)
 
I still don't understand the question.
 
@MikeMiller Trying to see why you can do that upto concordance. You lift up in a dimension (which you can do in concordances) and try "cancelling" two crossings, or something?
Is inverse the mirror image?
 
I'm glad I'm knot involved in this knot discussion.
 
Knots are obviously a monoid under connected sum. You can just take the Grothendieck group of that.
 
7:13 PM
(Which is not very geometrically exciting...)
I don't think Ali has a question :P He's asking philosophy
 
Wait, Ali went from chemistry to math to philosophy?
 
Shit, when was I a chemist?
 
@BalarkaSen Take K x I and delete a tube to get a disc bounding K # Kbar in B^4
 
Weren't you a chemist in high school?
 
I guess I am all 3 and none at the same time?
 
7:14 PM
@MikeMiller Just like for M # bar{M}, yeah
(which is nullcobordant)
 
that reminds me, chemistry is pretty cool, why not make an experiment like now, in the kitchen?
 
The vampire strikes again.
 
cooking :)
 
because my landlord will kick me out if I do chemistry experiments in the kitchen
 
I do plenty of that, @Alucard.
 
7:15 PM
oh cooking right
 
@Ali: Did your wall ever recover from the violence?
 
yes, I have been huffing paint all day as a result
 
Ah, I'm doing well with prescience today.
 
what is prescience?
 
Knowing ahead ...
Mind-reading ...
 
7:16 PM
I thought it was science for kiddies or something
 
LOL, that too.
@ShaV: Are you caught up?
 
if you really want so, you can cook anything you like, nothing will happen if you keep your cool
 
@Ted Sorry, I'm rereading it finally :P
You don't have to wait for me!
I'm writing everything we just did out
 
So philosophy wise: are there "homotopy groups" where the loops can't intersect (ie knots and stuff)
 
Excellent, @ShaV. ... You will have it all proved when you're done with that :)
BTW, the Euclidean algorithm is super powerful. Never forget about it.
 
7:20 PM
I suppose you could study homotopy groups of Emb(S^1, S^3)... not sure how exciting that is.
 
Sounds like framed cobordism, @Balarka.
Or maybe not.
 
It's exciting. I think any object whose pi_0 is interesting also has interesting homotopy type.
But the homotopy type of each component of that is now completely understood, due to Hatcher and Budney. Components of prime knots are aspherical, iirc, and the fundamental groups are usually not exciting.
 
Wow, huh.
@TedShifrin Yeah, I know what you have in mind
 
Framed cobordism is related to the homotopy types of other things in more complicated ways.
 
Yeah, I mostly decided that, @MikeM.
 
7:24 PM
I only know the story for framed cobordisms inside R^(n+k)
 
time's up, faster
 
You can pretty easily make a simplicial set of framed cobordism classes inside a closed manifold M and prove that the homotopy type of its geometric realization is the same as [M,S^k] for appropriate k.
 
if such a "homotopy-like group" existed, I think it would be fair to say it would only be non trivial for S^3 as the other S^n can be unknotted afaik. So I don't think it would be a very interesting invariant
 
I think you've been smoking more than I have.
 
You can't ever accuse me of that one, @MikeM.
 
7:27 PM
i think its the paint
 
Hey everyone, I have a quick question regarding a proof
 
LOL, @Ali.
OK, lunchtime for this Bonzo.
 
1-knot theory in higher dimensional spheres are not interesting. Codimension 2 spheres can knot.
 
Bon appétit @Ted
 
@ShaV: Please let me know when you're done. I want to make sure you understand it all. :)
 
7:29 PM
That sounds like "cannot" ... ugh
 
Merci, @Astyx.
 
Problem : Suppose $f$ is a continuous real valued function on a metric space $X$ and that $f(a) > 0$ for some $a \in X$. Show that there exists some $\delta > 0$ such that $f(x) > 0$ for all $x \in B(a, \delta)$
 
@Ted yes, I'm writing it out!
If I miss one essential detail in the beginning, my entire brain melts:(
 
yeah @BalarkaSen its due to Haefliger right?
 
mmmh, apples
 
7:30 PM
I didn't understand anything you told me, and I had to reread it 5 times now
 
I dunno the names
 
j-knots in S^n are non-trivial when 2n-3j-3>0
 
Hey everyone!
 
with writhing it out, I mean as far as we've gotten
 
I think there was something about the embeddings being smooth but I'm sure its not that important
 
7:30 PM
Hi @Daminark
 
Hi @Daminark
 
Hi @Daminark
 
Is anyone familiar with simulating Monte Carlo values in R?
 
@ozarka I might be if its something more specific?
 
!
 
7:31 PM
How's everything going?
 
So a proof for the problem would go as follows. Let $\alpha = \{a \in X | f(a) > 0 \}$ and let $\epsilon > 0$ be the smallest distance such that $B(f(a), \epsilon) \subseteq \alpha$, then continuity guarantees that there exists a $\delta > 0$ such that $B(a, \delta)$ contains $x$ such that $f(x) > 0$ for all $x \in B(a, \delta)$
 
@Ali Suppose we are interested in testing the null hypothesis that the mean of a normal population is 10 against the alternative that it is greater than 10. A random sample of size 20 from this population gives 9.02 as the sample mean and 2.22 as the sample standard.
deviation.
 
whats your significance thing, 5%?
 
I need to estimate the p-value using Monte Carlo simulation. I know that my null distribution is t-test (with $n-1$ df).
@AliCaglayan, significance is 5%, indeed.
I can set the CDF of t-distribution equal to $U$ (uniform random variable) and do nsim in R but that seems very difficult to do.
I meant runif in R.
 
Sorry @ozarka I am not sure how to do that in R
 
7:39 PM
@AliCaglayan
No problem.
 
@ozarka Have you tried looking for answers on cross validated?
 
@Ted Assume $d\mid a$. So far I've only shown that $\operatorname{order}(\overline d)\geq\operatorname{order}(\overline a)$. I don't see in our conversation how we prove that it is an equality:(
 
I think you can make a concordance simplicial set and I was interested in that at some point but I'm pretty sure the homotopy groups of every component are the same and classically known except maybe pi_1.
 
That's not generally true @ShaVuklia
 
then I give up. I'm going to look for a proof online
 
7:44 PM
For instance in $\Bbb Z_8$, $\text{order}(\overline 4) = 2$, $\text{order}(\overline 2) = 4$. But $2|4$
Don't give up now !
 
@MikeMiller could you elaborate?
 
@ShaVuklia but, that would be cheating?!? is that still ok?
 
but I don't understand where Ted is going
this is what I wrote up:
Our goal is to prove that $\operatorname{order}(\overline a)=\frac{n}{k}$, where $k=\gcd(a,n)$, and where $1\leq a\leq n$.

We've shown that for any $t$ that divides $n$, it holds that $\operatorname{order}\overline t=n/t$. Now I have to show that $\operatorname{order}(\overline{mk})=\operatorname{order}(\overline k)$, for any $m\in\mathbb Z$. The harder part is in showing that it can't be true that $\operatorname{order}(\overline{mk})<\operatorname{order}(\overline k)$. Assume the inequality holds, and that $\operatorname{order}(\overline{mk})=l$ and $\operatorname{order}(\overline k)=r$.
It took me half an hour to write this nonsense:l
 
Not on my ipad.
 
nonsense, in the sense that I have no idea what I'm even doing
 
7:46 PM
@MikeMiller no worries
 
Learn some knot theory :)
 
Nonsense is sexy :D
 
I'll learn it with Ali if he does
 
Perhaps give it some rest and get back on it with a clearer mind ? There's not need in exhausting yourself on this problem
Just be sure this is not laziness :)
 
but my final test is tomorrow:l
i hate math
i hate that i study it
 
7:49 PM
No you don't :p
 
I have Lickorish's book on Knot theory atm
and I have glimpsed at some knots in hatcher
but thats about it
 
draw a picture at your exam and see what happens LOL
 
Why don't you read Rolfsen together?
 
just be sure you use the right color(s)
 
What are you stuck at @ShaVuklia ?
 
7:50 PM
I have no problem with Rolfsen if A wants to do it.
 
sure lets do it
 
@Astyx what Ted is going for, like: the "main" thing he is trying to prove
to prove the whole thing
Then I at least know why I'm proving what exactly
 
you could probably use the geometry and topology room
 
Good idea.
 
I know that I need to show that $\operatorname{order}(\overline a)=n/\gcd(a,n)$.
 
7:52 PM
That's getting frozen again
 
It seems @Ted is back, so I'll leave it up to him ? (I don't want to spill my nonsense on his work)
 
@ShaV: You gave the proof that the order of $\bar k$ is precisely $n/k$. Do you understand that part?
 
yes
 
Cool. Now we want to show that the order of $\bar a$ is equal to the order of $\bar k$.
 
ohh..
now I finally see why you want $\operatorname{order}(\overline a)=\operatorname{order}(\overline k)$
 
7:56 PM
You gave the argument that since $a=km$ for some $m$, it follows that the order of $\bar a$ is at most the order of $\bar k$.
 
yes
 
Then I suggested the Euclidean algorithm. This showed that $\bar k = s\bar a$ for some integer $s$.
 
oh yes ok ok ok....
I'm so sorry I didn't see it. I should have known that I wouldn't be able to follow it if I was confused in the beginning
 
Now, in one more line, you should be done :) If you had $0<\ell<n/k$ with $\ell\bar a = \bar 0$, what would that tell you?
You don't need to apologize, @ShaV. I wasn't trying to hoodwink you. :)
Did you sort out what confused you at the beginning?
 
lol, yea what we were proving in the first place
 
7:59 PM
(But I continue to emphasize that working out concrete examples is a good way to understand a lot of algebra stuff.)
 
I had no idea, I thought it would become clear to me along the way
 
I guess that was the point of looking at $a=8$ and $n=12$, $k=4$. :P
Are you solid on it now?
 

« first day (2428 days earlier)      last day (2596 days later) »