« first day (2053 days earlier)      last day (3264 days later) » 
00:00 - 15:0015:00 - 00:00

00:41
Is there a list of all named principles and laws in logic theory? Like the principle of explosion? $\perp \vdash \psi$
Max
Max
01:23
If det(A^-1)=1/det(A) and A^{-1}=\frac{adj(A)}{det(A)} then det(A^{-1})=det(adj(A))/det(A)=det(adj(A))*det(A^{-1})
so det(adj(A))=det(A^{-2})
\where am i making a mistake here?
Surround LaTeX with dollar signs.
\$det(A^{-1})\$ becomes $det(A^{-1})$
(assuming you have LaTeX enabled on this chat)
(and it looks better as ${\rm det}(A^{-1})$)
I just did a handstand
it was great
Oh, and ${\rm det}(kA)\ne k{\rm det}(A)$ in general, I think
@MikeMiller Pics or it didn't happen
(Just kidding, I believe you)
(I have the sudden urge to attempt a handstand now)
Max
Max
oh, you're right it is not equal
I was up for like three seconds, beating my previous record of zero
01:34
HOLY SHIT THAT'S HARD
ALL THE BLOOD GOES TO YOUR HEAD
AND I DIDN'T EVEN GET COMPLETELY VERTICAL. I just kinda got sort-of vertical-ish and fell back down
and then tried a few more times
did you try to do it using a wall to get up? that's what I tried at first but I had the same problem
you just gotta kick yourself up from the ground and strike the right balance, your head won't ache so much since you'll probably fall before the blood rushes
I'm doing it against my bed, so it's sort of like a wall
I have just now startled everyone in the house by failing to handstand.
But I did manage to pass the perfect balance point
So it's definitely possible to get there.
The kick-up method works.
i've been trying since yesterday
01:38
bang "the heck was that?!?" "sorry, that was me" "why were you trying to do a handstand?" "because the internet told me to, duh"
I've been trying since five minutes ago
To be fair, no one in here said "@Axoren Do a handstand."
@Axoren, do a handstand.
2
there we go
After that, go to your room and think about what you've done
...my head still hurts
I blame @Mike
01:41
You should stand up with your arms towards the ceiling.
It's the inverse operation.
yeah that happened to me when i tried the wall thing
took like 15 minutes to feel better
Would it be better to be warm- or cold-blooded for handstands?
@Axoren Types of standing does not form a group
@AkivaWeinberger I'd argue that they're a subgroup of body permutation group.
I expect that some body permutations are incredibly painful and, in practice, not invertible
01:47
Especially the ones that kill you
$(head, arm)$
I'm mortified.
Or just plain morte.
(They're almost invertible, but you get a zombie instead of the identity)
What if you start with a zombie?
More zombie-like
I'll never understand zombie morphology at this point.
02:04
Today I learned that "dig" (in the sense of "to understand/appreciate") may have been borrowed from the West African language Wolof
Or possibly Irish
 
1 hour later…
03:17
@AkivaWeinberger I dig it.
Is there a characterization for the finite graphs in which there is the same number of u-v paths for all pairs (uv)?
@TheKindCat Like a full bipartite graph?
@Axoren those don't work
take for example $K_2,2$
I don't think I follow behind what it means "u-v paths for all pairs (uv)"
I mean no matter which two vertices you pick, there is the same number of paths between them
03:28
I see, now.
A lapse in my reading comprehension.
@TheKindCat Can you think of a graph for which it's path count would be 3 or higher?
Rather, one that already exists in another well-known family?
$K_n$ works, with $n\geq 4$
$K_4$ has more than $4$ paths between each pair.
Are you only interested in minimal paths?
Oh, nevermind. Again, I made another mistake.
No, all of the paths
03:33
$4$ has no importance with the number of paths.
Simply that there be the same number for each pair.
Hmm, it's an interesting property.
Haven't heard of it before, though.
Cycles and trees work.
In fact, trees are the only ones that work when we want exactly one path between them
Oh, and it suffices to classify the connected ones
$K_n \cup \mathcal T \cup C_n$ so far
Where $\mathcal T$ is the class of trees.
Don't think there's a symbol for it normally.
Are these the only classes?
I don't know, I haven't found any more
03:45
What if you subdivide some of the edges in a $K_n$ graph
We know that for $n = 1$, the graph must be a tree.
Consider $K_4$ with one of the edges subdivided; does that work?
For $n = 2$, does this necessarily make a cycle?
@Axoren is $n$ the number of paths between two vertices?
:28337143Yeah, that seems to work.
03:47
Yes. Sorry, let's give it a proper name under a different variable
I can see that overloading fairly quickly.
@Axoren No formal proof but I can't imagine anything else
I can give you a proof right now, actually.
Just thought of it.
Let's call it $n$-pathed, 'cause why not
And we want to show that all $2$-pathed graphs are cycles
Consider there are two paths from $(u \to v)_a = (u, a_1, a_2, \dots, a_k , v)$ and $(u \to v)_b = (u, b_1, b_2, \dots, b_k , v)$
And that no $a_i = b_j$ exists.
$a_1$ might equal $b_1$, no?
03:49
You almost caught me :)
These two paths by themselves form a cycle.
Imagine the graph with vertices $\pm2,\pm1,\pm i$, connecting $\pm2$ with $\pm1$ (for each of the two choices of $\pm$) and having a cycle from $1$ to $i$ to $-1$ to $-i$
If these existed any other points $c$, then there would have to be a path from any $a_i \to c$. Likewise for $b_i \to c$.
There would suddenly be a third path between some $a_i$ and some $b_j$ through $c$.
Therefore, no other vertices $c$ can exist. The only valid 2-pathed graph is a cycle.
@AkivaWeinberger I'm not clear with your counter example. Is that a connected graph?
My proposed graph isn't $2$-pathed, since there's only one path from $-2$ to $-1$, but letting $u=-2$ and $v=2$ shows that the statement $a_i\ne b_j$ doesn't need to hold
I see, but on the contrary, you misunderstood the point of that declaration.
I simply stated it that way to enforce that $u$ and $v$ lie on a subcycle of the graph.
Rather, a simple subcycle
And, letting $u=-1$, $v=1$, and $c=2$, we don't get "a third path between some $a_i$ and some $b_i$ through $c$"
03:56
I'm trying to make some problems for a competition, but my last 5 problems have been either false or unsolvable by me :(
$(u \to v)_b = (u = b_1, b_2, \dots, b_k = v)$
A restatement that might help.
You have two paths between $u$ and $v$ in the cycle.
And $2$ exists outside that cycle.
Oh, the $K_4$ graph with a subdivided edge doesn't work
Sorry
OH I see the flaw in my argument.
And why you're able to connect $2$ like so to counter the argument.
Still, though, there's only one path between $1$ and $2$…
then there would have to be a path from some $a_i \to c$ Likewise for some $b_i \to c$
03:59
…so it's not a true counterexample.
Instead, pick $u = i$ and $v = -i$ and pretend it's still a counterexample.
I didn't have sufficient evidence to claim ANY pair $a_i$ and $b_j$ satisfied it.
So, we know that $n$-pathed graphs for $n = 1,2$ have known families
What of $3$-pathed graphs?
I don't think we actually proved it for $n=2$
Or I missed something
Can we go over it again?
It may be clearer without vertices.
Draw a circle.
Then, draw an arc from one point of the circle to another above it.
Pick two points on the circle. How paths can you take between those two points?
Wrong.
You can travel one way around the circle or the other
04:04
Wait. Did I misunderstand the "draw an arc" bit?
The third way, is travel around the circle until you hit the base of the arc
And travel over the arc to the other side of the circle
Oh, the arc connects two points ON THE CIRCLE
Then, travel the rest of the way to the destination point
Yes.
This arc represents points not on the circle that are connected to it.
This is one such case.
04:06
Oh, I see; it has to connect to the circle on both ends so that there are at least $2$ paths from points on the arc to one of the bases
Got it
We clearly see that some points have $3+$ paths between them in this case, so the graphes similar to this shape cannot be $2$-pathed.
Another possibility is to draw a circle and then erect a tower at some point.
There are only $2$ paths between points on the circle, but for points on the tower, there's only $1$ path from them to points on the circle.
Which has only one path from points on the tower to its base
Yes.
Arright, I think this works.
So, for $3$-pathed graphs.
What kind of graphs work?
I suspect that not even a circle with an arc works for this.
Because I can imagine $4$ paths between any two points.
04:09
Note that $K_4$, the next-smallest one we know, is $5$-pathed (right?).
$3$-pathed actually.
Label the points 1, 2, 3, and 4
Points labeled $\{1, 2, 3, 4\}$
Between 1 and 2 we have:
12, 132, 142, 1342, 1432
Right.
My mistake.
For the circle with the arc, we have the following paths:

Clockwise along the circle,
Counterclockwise along the circle,
Clockwise until arc, follow the arc over, take the shortest way to the destination
Counterclockwise until arc, follow the arc over, take the shortest way to the destination
Lets flatten the arc to a diameter-length bisector. This gives us more descriptive power.
04:13
So, θ
$\theta$
If the two points are on opposite semi-circles, there are 4 paths between them
And if two points are on the same side, or both on the diameter, we have 3
If they're on the same semi-circle, there are 3, because traveling one direction around the circle will reach the destination point before reaching the base of the arc/separater
It's disqualified as a $3$-pathed graph
I can never get that right the first time
And from here on in we have to be careful
because subdivision doesn't preserve $n$-pathèdness
Adding a point (to which this arced cycle is an allegory for) immediately transitions from a $2$-pathed graph to a strictly more-than-$3$-pathed graph.
04:16
So we need to remember which points are vertices and which points aren't
Let's see if adding an edge works?
What do you mean "adding an edge"
From a graph that is $2$-pathed, connect currently unconnected vertices with an edge.
We can see that this is equally fruitless.
It has the same effect as added an arbitrary amount of points along the diameter.
04:18
The $θ$ graph becomes the $K_4$ (and thus $5$-pathed) when adding an edge
Consider a diameter/arc with no points on it that could still be traversed.
The $\theta$ graph?
Shaped like the letter θ
Right, but it doesn't become $K_4$
Imagine $K_4$ minus an edge;
it becomes the theta graph
I think you misunderstood what I mean circle.
The number of vertices on the $\theta$ graph could be upward of $1000$ to nearly uncountable if it was a stretching of the real interval $[0, 1]$
The arc/diameter's vertex count was always arbitrary.
I don't know exactly what you imagined the $\theta$ graph to be.
Was it the $C_4$ with extra point?
$1 \to i \to -i \to -1$?
04:23
$C_4$ plus an edge connecting two opposite points
I see.
Consider a larger cycle, because we don't need to stay that small.
Consider $C_{999}$
But if we take $C_4$-plus-an-edge…
…and add another edge…
We don't achieve $3$-pathedness
we get a $5$-pathed graph.
$K_4$.
We want to show that if we take ANY $2$-pathed graph, we don't get a $3$-pathed graph.
So, you can't use that single case as the counter example.
04:25
Oh, OK
Consider a circle. Connect any two points with a portal.
Yeah
Then that's not $3$-pathed
Even if that portal does not consist of additional points, the same effect occurs.
There are no $3$-pathed graphs.
Adding edges and adding points are the only ways to increase the pathedness of a graph
So that's that
For $3$-pathed graphs
Now, let's consider $n = 4$
Maybe the idea of a spanning tree is useful?
Maybe.
What does a spanning tree give us outright?
04:29
To connect $a$ and $b$, we can either go along the unique path on the spanning tree, or we can do modifications of that (occasionally using additional edges)
I'd find it hard to utilize that in the general case where we have an arbitrary graph.
I'm not sure if it helps a lot, though
Yeah
Let's consider a general method:
Take any $5$-pathed graph.
I actually need to go to bed
:(
Same with me, but just wait a moment to catch this last bit
You can have nightmares about it and come up with a solution
Removing edges is the only way to reduce pathedness.
Is there a way to reach a $4$-pathed graph starting from some $5$-pathed graph?
Orphaned points can be discarded, too.
Doesn't seem like it
But we don't know what all $5$-pathed graphs look like. We only have the one.
The next question is then to be are there graphs other than $K_*$ for $n \ge 5$?
Good night
04:32
Good night
wait, I thought subdividing an edge worked
is this not the case?
If we take an n-pathed graph and subdivide an edge it is also n-pathed right?
@TheKindCat If you take $K_4$ and remove the path between $1$ and $2$, you affect all paths between $u$ and $v$ that pass through $1, 2$
Between $1$ and $2$, there are now 4 paths.
Between $1$ and $3$, however.
How many are there?
subdividing an edge (u,v) means adding a vertex w
that is only connected to u and v
and removing edge u,v
Subdividing an edge works in theory.
I think
But let's look at the paths from $w$ to some other points
04:44
For instance, how many paths are there from $w$ to $u$?
there is a direct one
There's $1$ + $\text{paths from $v$ to $u$} - \text{$1$ direct path from $v$ to $u$}$
exactly
so $n$
So, subdivision works.
Because we can subdivide again and the same formula will hold again
For arbitrary subdivisions.
well, we need to analyze the other cases, but yeah, it checks out.
04:46
That works.
But we still don't have a $4$-pathed graph
And we can't prove that the only $5+$-pathed graphs are subdivisions of $K_n$
Rather, we haven't.
:(
I'll let you know if I can solve it.
Let's discuss the notion of minimal $n$-pathed graphs to stay this troublesome idea of being able to subdivide edges as much as we want.
That way, we know removing an edge will not preserve pathedness.
Is the minimal $n$-pathed graph $K_m$ for some $m$?
minimal in which sense?
That you can't replace a series of tree edges with forward edges in the spanning tree and preserve pathedness.
oh
I'll work on it first thing tomorrow, my parents insist I go to sleep :(
04:54
Good night.
thank you very much for your help !
I'll fidget with it more.
 
3 hours later…
Max
Max
07:46
Is there any relationship between between and adjugate matrix norm an matrix norm?
between an adjugate matrix norm and matrix norm*
is it true that for, let's say, supremum norm |adj(A)| < |A|^{2n}
or anything like that
?
08:02
I have to prove that if a group $G$ is of even order then it contains one element of order 2

Proof 1: We know that the identity $e$ does not have any inverse (it has itself, I mean to say it cannot be paired) that is $e=e^{-1}$.Assume there is no element of order 2(i.e , for some $a\in G, a^2=e \implies a=a^{-1}$) But every other element has an inverse . So we can makes pairs for each of the elements that is the element and its inverse. All such elements add up to an even number, but since there is the identity which is till left , the total sum will be an odd number. This is a contradicti
@Albas That starts with several wrong statements
the identity does have an inverse (itself)
Sorry for that . Wait
and $a^2 = e$ always implies that $a = a^{-1}$. This is unrelated to the existence of elements of order $2$.
Sorry I do not get you @Tobias
@Albas You assume no element has order $2$ and then write i.e. ...
but your i.e. has nothing to do with that assumption
08:07
I just wanted to show that if $a^2=e$ then what would have happened . I should not have written i.e
Anyway, the general idea is fine
What about proof 2?
Proof 2 is pretty much complete nonsense, yes
I had felt that. Sorry
You correctly show that the group would not be isomorphic to a very specific group of even order. But plenty of groups of even order are not
08:10
Hmm... So is there any way I could prove what was stated with the help of isomorphisms
Thanks for checking @TobiasKildetoft
Wait... I had one more question
in general what is the relation between the order of a group and the order of an element?
@Albas The order of an element is the order of the subgroup generated by that element and it divides the order of the group.
Also, less trivially, if $p$ is a prime which divides the order of the group, then it has an element of order $p$
What if the order of the group is a prime?
I leave that as an easy consequence of the above
So the order of any subgroup that will divide the group of prime order will have the same prime order?
@Albas Well, or order $1$
08:20
Yup.
 
3 hours later…
11:11
@TheKindCat @Axoren Subdivision does not work. Take $K_4$ with vertices labeled $1,2,3,4$, and add a vertex $x$; remove the edge $12$ and add edges $1x$ and $x2$.
Then there are $5$ paths between $1$ and $2$, as usual, but
there are $6$ paths between $x$ and $4$:
x14, x24, x134, x234, x1324, x2314
 
1 hour later…
12:16
Why does $\mathbb{Z}^{\times}_{20}\cong\mathbb{Z}^{\times}_{4}\times\mathbb{Z}^{\times}_{‌​5}$ imply that $\mathbb{Z}^{\times}_{20}\cong\mathbb{Z}_{2}\times\mathbb{Z}_{4}$? I know the first bit is the Chinese remainder theorem, but why does the implication follow?
@DeMoivre Just check this directly
they are small groups
12:29
@robjohn, Polya's Problems and Theorems in Analysis says, 'In mathematical induction the result to be obtained and the means available for its proof are proportional, they stand in the ratio of n + 1 to n. Hence, strengthening the statement to be proved may also be advantageous, for we strengthen at the same time the means available for its proof.' How is that ration n+1/n? I do not get this.
@Silent That is a terrible formulation, which you can probably ignore. The important thing is that strengthening the thing to be proved also strengthens the means available for proving it
(this does not always mean that one can just strengthen the statement, but it does mean that smetimes it makes things easier)
@TobiasKildetoft, thank you very much
13:20
|x>3| = |x| > 3 = x > 3 | = | 3 =/ x+_y?
 
1 hour later…
14:23
is there any name for this group:
For a prime $p$
$$\Bbb{Z}(p^{\infty})=\{\overline{\frac{a}{b}}\in \Bbb{Q}/\Bbb{Z}| a,b\in \Bbb{Z}, b=p^i, i\geq 0 \}$$
where the operation is usual addition as it is for the rationals modulo integers
it might just be superficial, but that seems reminiscient of p-adic stuff
Prufer group
@Semiclassical: it's superficial
welp, there we go
note that everything in this group is torsion!
what do you mean by torsion @MikeMiller
14:33
dinite order
though, wikipedia's page on prufer also mentions at the bottom that p-adic integers can be defined as the inverse limit of the finite subgroups of the Prufer group
so that doesn't seem entirely superficial
@AkivaWeinberger Oh yeah, my bad.
sure, one is a limit and the other is an inverse limit, but that's a pretty big difference
to be fair, i don't really understand inverse limits
IMO less superficially is that this is the pontryagin dual if the p-adics
14:35
nice
not that I really understand pontryagin duality.
i don't really know pontryagin duality either, alas, aside from it somehow encompassing the usual notion of position space - reciprocal space duality in Fourier analysis
though when i use the word reciprocal space i'm really thinking in terms of applications to solid state
let's take another crack at this question I wasn't able to answer last night
back in a bit
14:39
How would I graph something implicit and 3D
Specifically, $x^2y^2+x^2z^2+y^2z^2=xyz$
Wolfram Alpha doesn't zoom in enough
tell it to zoom in
I don't know how
whisper zoom in
possibly "enhance" if you want the real CSI feeling
I think I got it, actually
I typed in plot x^2 y^2+x^2 z^2+y^2 z^2=xyz, for -1<x<1, -1<y<1, -1<z<1
guys, I did poorly on an exam. should I kill myself?
probably not
hmm, alright, guess I won't
but fuck am i angry
@iwriteonbananas wtf
14:50
lol :/
go smoke or something and pass out for the rest of the day
i can't will myself to feel better the day something shitty happens
ugh ok. guess i'll start smoking
idk i'm just giving worthless advice
undoubtedly you know what to do when life is terrible for you better than i do
00:00 - 15:0015:00 - 00:00

« first day (2053 days earlier)      last day (3264 days later) »