« first day (2783 days earlier)      last day (2228 days later) » 

12:00 AM
1 hour ago, by DarkVampiric AbstractArtist
What is the smallest positive integer n for which 147x+105y = n = 606u+909v for some x, y, u, v ∈ Z? For this n, give values of x, y, u, v that satisfy the requirements. My attempt: I think n is 3, since 3 is common between all of them and divides them all. Then it requires some messy Euclid's algorithm
 
Semiclassical is this supposed to do with Bezouts coefficients?
 
it's more basic than that.
What's something you can definitely say about 147x+105y if x,y are integers?
 
that they have a common divisor?
 
Okay. What common divisor is that?
 
12:02 AM
Aha.
 
Right. So $147x+105y=n$ has to be a multiple of 21.
What about $606u+909v?$
 
Right. So n has to be a multiple of both 21 and 303.
And you want the smallest such n...
 
So not at all the gcd of all 4.
 
They share 1 and 3.
 
12:04 AM
multiple, not factor!
 
Ted :D
@TedShifrin Hey Ted ._.
Did you watch federer game?
 
About Euclid's lemma: if $p$ is not prime, then it will be false?
 
12:19 AM
Is that it semiclassical? We find that 3 is a multiple of both 21 and 303, so 3 is the smallest positive integer?
 
No.
3 is not a multiple of both 21 and 303.
It is a factor of both of them, but it is not a multiple of either of them.
 
21 = 3 x 7 and 303 = 3 x 101?
 
Yes. But that doesn't make 3 a multiple of 21.
It makes 21 a multiple of 3, but not vice versa.
 
LCM(21,303)=2121 and since 2121=3x7x101?
 
Right.
So the smallest n is 2121.
 
12:27 AM
And is that it :o?
 
Since $(6=2\cdot 3) \not\mid (21=3\cdot 7), 6 \not\mid (2\cdot 5)$, but $6 \mid 210$. I reply to myself q.
 
I think that's logically sufficient, yes. But probably one should actually work out what x,y,u,v are.
 
So I have to find x,y,u,v
 
So you'd need 147x+105y=2121-> 7x+5y=101, if I'm looking at that right.
Simplest way I know to solve that is to note that 7(3)+5(-4)=1 by inspection
and then 100=5*20, so 7(3)+5(20-4) should equal 101.
and it does.
similarly, you've got 303u+606v=2121 -> u+2v=7
which is pretty easy to eyeball
 
how can I prove this equation holds if we divide that square of side 1 into seven right triangles?
 
12:41 AM
huh.
well, all those right triangles are similar.
And in particular you know that the ratio of the two legs is x/1.
So that should be enough to start working out what the internal lengths are.
 
user338510
Hello my friends
 
I tried with the pythagorean theorem
but I get some crazy equations
 
yeah, that's part of it too: EH should be $\sqrt{x^2+1}$, for instance
but even so, that does not seem easy.
But, for instance, you should have CE = x/sqrt(x^2+1) and DE=1/sqrt(x^2+1)
knowing DE should then be enough to figure out DF and EF, and so forth
 
user338510
@Twink I defined side $EC$ as $x$, and started Pythagorean Theorem from there, but that it too lengthy of a process.
 
but is $x$ is defined as $AH$
 
12:47 AM
Maybe the smart way to approach it is in terms of areas?
the area of the AEH triangle is x/2.
 
I also tried with areas but didn't get anything
 
the sum of all the areas is 1
 
user338510
Sorry, $EC$ is another variable called $y$
 
yeah.
if I were going to do a supplementary variable here, I'd take the smallest length. that happens to be IG here
 
12:48 AM
and the area of each triangle is the product of two sides divided by 2
 
well, you know that the sides of each triangle are in the ratio x : 1 : sqrt(1+x^2)
 
user338510
Maybe note that AHB and IBC are similar triangles.
 
user338510
From there, solve for their area by defining a new variable for IB
 
yeah, this seems bloody tedious no matter how you slice it.
 
user338510
Oh never mind @Twink I found out how you do it. Use that ABH has an area of (1/2)x and BHDC is 1-(1/2)x. Also, you can define BHDC as a trapezoid, with area of $(1/2)(1-x+1)\cdot 1$. Set those areas equal to one another, to solve for $x$.
 
user338510
12:59 AM
Problem is, that does not work.
 
you mean $1-\frac{1}{2}x=\frac{1}{2} (2-x)$? I get $0=0$.
 
user338510
Yea I just noticed
 
No offense but whenever I see your username, the first thing that passes in my mind is I don't know :D @idk
 
user338510
same
 
1:24 AM
One interesting thing to note is that, if you move triangle ABH to the right so that AE overlaps CD, then you get a parallelogram
I'm not convinced that helps much, but it's cute.
I think one could then slide triangle ICE up as to form a square...but again, not sure why one should bother.
like that
 
I don't even know if that's a square
Hlello
 
you're right, it's not.
it's got the same area as the original square, but it's a rectangle
 
Hm. So sides of length $\sqrt{1+x^2}$ and…
 
original area was one, so it'd better be 1/sqrt(1+x^2)
 
$\dfrac{\sqrt{1-x^2}}{1-x^4}$.
 
1:31 AM
x^4?
 
Yeah, that's not better. $1/\sqrt{1+x^2}$ it is.
 
actually, it's not quite the same
 
(Multiplying by the conjugate on the top and bottom)
 
should it be a square root in the denominator?
 
1:32 AM
@AkivaWeinberger in here
 
I rationalized the denominator
 
uh
you'd rationalize the denominator by multiplying by sqrt(1+x^2)
 
Oh
Right, duh, sorry
 
no worries
I'm not convinced it actually makes the problem easier in any case. it just looks sorta cute.
i want to say one should be using the fact that GH/IH = FG/GH = EF/FG = DE/EF = CD/DE = EH/AD = sqrt(1+x^2)
 
made it back alive
 
1:37 AM
but ugh that looks awful.
 
$1\sqrt{1+x^2}$ = IC = IG + GE + EC = EC(1+FE+FE^2)
EC = $x/\sqrt{1+x^2}$
1 = x(1+FE+FE^2)
ED = 1/\sqrt{1+x^2}
FE = ED/\sqrt{1+x^2}
FE = $1/(1+x^2)$
 
CD=1,
DE = (1+x^2)^(-1/2),
EF = DE/sqrt(1+x^2) = (1+x^2)^-1,
FG = EF/sqrt(1+x^2)=(1+x^2)^(-3/2)
GH = FG/sqrt(1+x^2) = (1+x^2)^(-2)
IH = GH/sqrt(1+x^2) = (1+x^2)^(-5/2)
 
$1=x\left(1+\dfrac1{1+x^2}+\dfrac1{(1+x^2)^2}\right)$
 
$(1+x^2)^2=x((1+x^2)^2+(1+x^2)+1)$
 
1:43 AM
and that's consistent with what I just wrote, I think
$1+2x^2+x^4 = x+2x^3+x^5+x+x^3+x$
$x^5-x^4+3x^3-2x^2+3x-1=0$
huzzah
 
Arright what do we win
 
satisfaction
 
Yours looked more systematic, I imagine you were about to add BI = $x/\sqrt{1+x^2}$ to it
 
well, the way I'd finish would be to multiply some of those by x to get HF and FD. (I didn't actually need to do IH in retrospect)
and then 1=AH+HF+FD = x+ x(GH)+x(EF)
 
Mm
$x\approx0.378916$
 
1:48 AM
nice
 
So remind me about symmetric matrices
 
hmm. If $x=\tan \theta$, then $1/(1+x^2)=\cos^2\theta$
 
Why do they always have (real) eigenvectors?
An orthonormal basis of them, in fact
 
so the above can be written as $1=\tan\theta[1+\cos^2\theta+\cos^4\theta]$
 
Interesting
 
1:51 AM
tbh I don't remember. the fact that (real) symmetric matrices have a nice eigendecomposition is a fact I use a lot more than I derive
I want to say it's just the spectral theorem tho
 
So I managed to prove it to myself earlier today (after a long time thinking about it) in terms of the fundamental theorem of algebra
 
makes sense.
 
Like, $\det(\lambda I-A)$ must have some complex roots, if it's real then $Av=\lambda v$ and $A\bar v=\lambda\bar v$ and add them and we have a real eigenvector
If it's not real then $\lambda(v^\top\bar v)=v^\top A\bar v=\bar\lambda(v^\top\bar v)$ but that means $v^\top\bar v=0$ so $v=0$ ('cause $v^\top\bar v=\sum|v_i|^2$)
I'm just asking here because I felt like there was probably some proof that didn't involve complex numbers
But maybe not
 
There's a bunch of proofs of the spectral theorem, I think
 
In any case, my Riemannian geometry book used it a few times today
and I felt like I couldn't continue until I proved it, so that took some time
 
1:59 AM
I'll note that the proof in the wiki page for the spectral theorem is for Hermitian matrices
 
And the subspaces with different lambdas are orthogonal 'cause $\lambda_ww^\top v=w^\top Av=\lambda_vw^\top v$ so if $\lambda_w\ne\lambda_v$ then $w^\top v=w\cdot v=0$
 
right
 
So like in Gauss's theorem, where the intrinsic curvature of a surface immersed in $\Bbb R^3$ is the product of the two principle curvatures,
those principle curvatures are the eigenvalues of a certain self-as joint map
so they're orthogonal, and their product is the determinant of that map
(The Gauss spherical mapping, which maps a point to its unit normal vector${}\in S^2$)
I don't know why those are the maximum and minimum curvatures of curves obtained by intersecting the surface $S$ with planes normal to $S$
Specifically why maximum and minimum
So, I'm guessing, the maximum and minimum values of $A(x)$ with $\|x\|=1$ are found at the eigenvectors? @Semiclassical
 
@BalarkaSen X released an album, "?"
 
@AkivaWeinberger the proof of the spectral theorem via Lagrange multipliers makes this clear I think
 
2:12 AM
I have completely forgotten Lagrange multipliers
It's somewhere in Ted's book
Probably the chapter right after the last set of exercises I started (and abandoned)
 
Lol
 
So in Riemannian geometry, at least so far, the way it looks in this book, you just have things randomly turn out to be tensorial by happy coincidence
Like, the right terms happen to cancel out
 
If you look at the frame bundle the curvature and torsion appear more naturally, admittedly.
 
There's this one thing where, say you have a surface or manifold $M$ immersed in $\Bbb R^n$. Take a point $p$ on $M$ and two vectors $x,y$ at $p$ tangent to $M$
 
Do Carmo definitely takes a bit of a "we defined it this way and worked out consequences" kind of approach
 
2:18 AM
So the way you define this $B(x,y)$ is, extend $y$ to a vector field $Y$ on $M$,
 
@BalarkaSen Oh god there's one song where he screams full blast. Whoa. Visionary.
 
or even just extend it to a field on a curve on $M$ with $x$ as a tangent vector, the field tangent to $M$ everywhere of course
 
You're already at immersions?
 
and take the directional derivative of it at $p$ in the direction $x$,
and then project the result onto the orthogonal complement of the tangent space at $p$.
Or even do that with $x$ instead of $y$, 'cause $B(x,y)=B(y,x)$ it turns out.
So there's so much room for choice there, but it turns out that it's well-defined, by magic. (But only if you do project it onto the orthogonal complement of the tangent space)
(The other component is not well-defined the way I've just described it)
 
This is to be expected from the classical surface theory
 
2:21 AM
Hm, actually
So if we orient it so that the tangent plane is the xy-plane
and think of the surface as the graph of z=f(x,y), at least locally,
is this just the second partial derivative? In those two directions?
(Well, times a unit vector in the direction of the z-axis)
I guess yeah, if you choose the field $Y$ and the curve through $x$ correctly
(Now I have notation troubles; $x$ there is a vector in the horizontal plane, not the axis)
@0celo7 In any case, yeah, I'm at immersions
I've been reading it pretty quickly
I just finished the chapter on Jacobi fields earlier today
(disappointed that it wasn't about the algebraic type of fields)
 
no one cares about those
the algebraic ones, that is
@BalarkaSen Oh god he's screaming again!
 
So hold on, say I have a function $f(x,y)$, and I have a vector $v=(a,b)$, what's $\frac{\partial^2f}{\partial v^2}$?
 
what?
 
$a^2f_{xx}+2abf_{xy}+b^2f_{yy}$, right?
@0celo7 Like, the second derivative in the direction of $v$
 
Hess f (v,v)?
 
2:30 AM
If that's what it's called then yeah
 
@AkivaWeinberger yes
@BalarkaSen Now he's singing in Spanish
 
Mhm. So do Carmo's $B(x,y)$ (which he writes as $\overline\nabla_{\overline X}\overline Y-\nabla_XY=(\overline\nabla_{\overline X}\overline Y)^N$) is essentially just the derivative in the direction of $x$, and then the derivative in the direction of $y$, when you orient the surface such that the tangent plane is horizontal
(Here the immersion is $f:M\to\overline M$, $~x$ and $y$ are tangent at $p$ to $M$, $~\overline\nabla$ is the connection on the surrounding space $\overline M$ while $\nabla$ is the connection on $M$ which is identified with $f(M)$,
and $v=v^T+v^N$ is the decomposition into the tangential and normal components where $v^T\in T_pM$ and $v^N\in(T_pM)^\bot$.)
But when he just writes the formula, you don't know that you're supposed to have that intuition which is why it's annoying
@EricSilva So yeah, what you said
 
It's pretty intuitive if you've learned the classical surface theory
 
Suppose that you are given positive integers a, b, c. Prove that if there are any integers
x, y satisfying ax + by = c then there are infinitely many solutions to this equation.
Attempt:(edited)
If a and b are coprime you can always find integers p, q such that

ap + bq = 1 using Euclid's algorithm.

Then multiply by c to get

ax' + by' = c ......(1) where x' = cp, y' = cq

and finally let x = x'+bt y = (y'-at) and substituting into the
equation

ax + by = c we get

a(x'+bt) +b(y'-at) = c
Can you check if it's right?
 
2:40 AM
There's a much easier way to do that
Add and subtract ab
 
I think do Carmo probably would expect it going into his Riemannian book considering he wrote a classical surface one
 
@BalarkaSen Now he's whispering, but someone else is screaming.
 
c = ax + by = (ax + ab) + (by - ab)
= a(x+b) + b(y-a)
So then just repeat that as many times as you want
If we add and subtract Kab we get something similar
c = a(x+Kb) + b(y-Ka) @DarkVampiricAbstractArtist
 
Do I keep solving a(x+b) + b(y-a)?
 
What do you mean "solving"?
 
2:44 AM
Oh is K some integer of ab?
for*
 
K is any integer
The point is that if x'=x+b and y'=y-a then ax'+by'=c
As for your proof, you assumed a and b are coprime, but that might not be true
 
I used it from the lecture notes, but yeah you're right.
 
(If x'=x+Kb and y'=y-Ka then ax'+by'=c, so that's infinitely many 'cause you can choose infinitely many K)
 
This technique is called "add zero in a clever way", I'm pretty sure
'cause by writing c = ax+Kab + by-Kab, you're essentially just adding Kab - Kab, or zero
 
2:51 AM
"do nothing, but do it in a clever way"
most of algebra is just that :P
 
I finally have a use for $\mathfrak R$
it's the best letter
 
$\frak K.$
$\frak My~god,~what~is~this$
 
$\mathfrak L$
oh, dam son
 
$\frak\text{Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn}$
$\frak\text{Die deutsche Sprache ist eine westgermanische Sprache.}$
 
3:04 AM
@AkivaWeinberger $\frak{dei\,\, Mutti}$
 
Ugh... the spacing is all wrong!
FOR THE LOVE OF GOD, KERN!
 
> Noun
Mutti f (genitive Muttis or Mutti, plural Muttis)
1. mom, mummy
Proper noun
Mutti f (genitive Mutti)
1. A nickname for Angela Merkel.
 
Wot
It means mom
Angie's nickname is Angie
 
We all call our advisor "Papa"
 
@XanderHenderson #metoo
 
3:11 AM
I saw a terrible t-shirt the other day
 
@XanderHenderson keming
 
on the front, it said #metoo
on the back, it said "and I liked it"
 
@AkivaWeinberger that is why sans serif fonts are the debil
THE LITERAL DEBIL!
 
$\rm keming$
$\rm kerning$
Hm
 
3:13 AM
$\text{rm}$ is still something of a problem
 
Let p be a prime and suppose that n = m^2 for some m ∈ N. Prove that if p | n then p^2 | n. My attempt: since n=m^2=m*m, therefore p divides n, p=nc, for some c integer, p^2=n(nc^2), nc^2 in Integer, therefore divides p^2 divides n=m^2
 
but it isn't the end of the world
 
Is this right or wrong?
 
$\frak kerning$ $\frak keming$
What is that "k"
 
it is a fraktured $\frak k$
or is it a fractalled $k_{k_{k_{k_k^k}^{k_k^k}}^{k_{k_k^k}^{k_k^k}}}^{k_{k_{k_k^k}^{k_k^k}}^{k_{k_k^k}‌​^{k_k^k}}}$?
 
3:14 AM
@DarkVampiricAbstractArtist "P divides n" means "pc=n"
@XanderHenderson oh lord
 
Feb 28 at 20:43, by Balarka Sen
$\gamma^X_p[0, t] \cup \gamma^Y_{\gamma^X_p(t)}[0, t] \cup \gamma^X_{\gamma^Y_{\gamma^X_p(t)}(t)}[0, -t] \cup \gamma^Y_{\gamma^X_{\gamma^Y_{\gamma^X_p(t)}(t)}(-t)}[0, -t]$
 
anyone here good at combo?
 
Depends?
 
Suppose $x$ is a binary string of length $n - k$, where $n, k$ are integers and $n > k$. An 'extension' of $x$ is a string of length $n$ which is formed from $x$ by inserting exactly $k$ total new 0s and 1s at any position in the string $x$. How many distinct such 'extensions' are there from $x$? Does it depend on more than the length of the string $x$ (i.e., the value of $k$)?
Based on numerical evidence the answer is 'it does not depend on the actual string x'.And the number of distinct extensions is $\sum_0^k \binom{n}{i}$
 
3:15 AM
hrm... I was hoping that they would start to overlap
oh, well
 
Please use \$blah\$, not \$\$blah\$\$
 
apologies
 
n=pc, for some integer c, n^2=p(pc^2), then pc^2 ∈ Z, therefore p^2 divides n=m^2
 
I don't see how you got that last step
It looks like you showed that p divides n^2, not that p^2 divides n
 
n^2=p^2 x c^2, c^2 ∈ Z, therefore p^2 divides n.
 
3:22 AM
I don't understand what you're doing
You just gave me n^2 as p^2 times something, that means p^2 divides n^2, not n
You need to use the fact that p is prime
It's not true if p is not required to be prime
 
anyone?
 
I'm thinking
 
akiva do you agree it is trivial for 0^{n-k}
 
Trivial for what?
 
that the formula holds
 
3:27 AM
What do you mean "for 0^{n-k}"
 
i.e., if x = 0^{n-k} then there are exactly \sum_0^k nCi extensions
 
aesthetics question:
 
Oh like the string 0 repeated n-k times
 
exactly
yes, sorry the notation is slightly overloaded
 
$“0”^{n-k}$ I guess
 
3:28 AM
it really means the concatenation of the vector (0, ..., 0) \in F_2^{n-k} if you like
 
Right I see
 
$$ \vec{E} = \left\{
\begin{array}{cc} -\dfrac{1}{4\pi \epsilon}\dfrac{Q}{r^2}\hat{r}, & a<r<b \\ 0, &\text{else}
\end{array} \right. \qquad
\vec{D} = \left\{
\begin{array}{cc} -\dfrac{1}{4\pi \epsilon_0}\dfrac{Q}{r^2}\hat{r}, & a<r<b \\ 0, &\text{else}
\end{array} \right.$$
 
Is there a way to simplify the sum?
 
unfortunately, not very easily. there is not closed form expression for the first k binomial coefficients
 
Does that look weird? I'd be fine with those two piecewise functions on their own, but I'm not sure how good it looks with them in parallel
 
3:29 AM
It doesn't even fit on my screen (mobile)
 
hrm
I can put boxes around them, so maybe that'll work best e.g.
$$ \boxed{\vec{E} = \left\{
\begin{array}{cc} -\dfrac{1}{4\pi \epsilon}\dfrac{Q}{r^2}\hat{r}, & a<r<b \\ 0, &\text{else}
\end{array} \right.}$$
 
but do you agree that the sum is correct when $x = 0^{n-k}$, @AkivaWeinberger
 
Great, and the argument I use is that all the k-extensions of x, are formed via inserting i many 1s and k-i many 0s.
for i between 0 and k.
each term of the sum corresponds to a choice of i.
and for uniqueness, you only have to argue that each term is unique. i.e., there are exactly n choose i many distinct strings formed by inserting i many 1s into x.
Anyways, it is clear that it holds for $0^{n-k}$.
 
But it gets reversed for $[1]^{n-k}$
 
3:32 AM
sure you now insert i many 0s.
 
Like the result is the same but the argument is backwards
 
but the questions is what to do for something mixed.
so it seems like somehow one needs to reduce to 0^{n-k}
 
So 0000 when you extend it by two
 
if the formula is correct one should be able to reduce to $0^{n-k}$.
 
That's how many?
 
3:34 AM
hm?
what do you mean by that
 
0000 extended by two
 
what is "it"
 
How many extensions of 0000 when n=6 and k=2
1+6+15=22?
 
thanks. it should in theory be 6 + 1 + 6C2
idk what 6C2 is
15 apparently
 
So now take 0101 when n=6 and k=2
 
3:35 AM
so there should be exactly 22 unique extensions
ok.
should still in theory be 22 unique extensions
 
Hopefully
 
did you want to verify that ?
 
Might be annoying to do it all out
 
i mean it will almost certainly be true. I've tested this formula by simulating a bunch of random strings for various n, k and its true
I did it for like 100,000 strings of various n and k lol
 
3:36 AM
yeah :) so it is probably true
 
0101 is like unusually symmetric though
 
yeah, same with 1010
 
Adding 01 to the start is the same as adding it to the end or the middle
or adding 10 various places
 
lemme see if I can find my code
that produces all the unique guys
 
Let p be a prime and suppose that n = m^2 for some m ∈ N. Prove that if p | n then p^2 | n. My attempt again: since n=m^2, then n^1/2 = m, therefore if p divides m, m=pc, for some c integer, then n^1/2 = pc (since m=n^1/2), then squaring it n=p(pc^2), pc^2 ∈ Z , therefore divides p^2 divides n=m^2
 
3:45 AM
If p divides m. But you didn't prove that p divides m
Did you prove, in class, that if p|ab, then either p|a or p|b (where p is prime)?
 
@AkivaWeinberger
"100100"
"101001"
"101010"
"101000"
"101100"
"110100"
"010100"
"100010"
"010010"
"100110"
"110010"
"100101"
"010101"
"101011"
"101101"
"110101"
"110110"
"010110"
"101110"
"001010"
"011010"
"111010"
there you go lol
 
Never have only c|ax+by
 
Do you know about the uniqueness of prime factorization?
 
2,3,5? When any integer is greater than 1?
is prime
 
3:55 AM
You're not making sense
OK, here's a starting point: assume p doesn't divide m. Since the only factors of p are 1 and p, and p isn't a factor of m, the greatest common factor of p and m must be 1. In other words, they're coprime.
Since they're coprime, there exist a and b such that ap+bm=1.
Remember that m^2 is a multiple of p. See if you can find a contradiction.
(Once you do find a contradiction: that would mean that our assumption that p doesn't divide m is false; that is, p does divide m. Then m=pc for some c, and m^2=p^2*c^2, and thus p^2 divides m^2=n as you wanted to prove.)
(So you just need to find that contradiction.)
 
Alrighty then, I guess I'll use proof by contradiction.
 
4:11 AM
> Ramsey theory and Fermat’s Last Theorem
 
4:27 AM
You're a fan of discrete maths?
There's a youtube video (I think from Numberphile) that talks about the party problem.
 
4:54 AM
anyone able to help: mathb.in/23477 (COMBINATORICS)
 
Isa
Why in the solution of first order PDE the argument of the function must be positive?, for instance the PDE $u_t+tu_x=0$ has solution $u(x,t)=\phi(\frac{t^2}{2}-x)$, why is necessary to have $\frac{t^2}{2}-x>0$ ?
 
5:28 AM
Because it's strictly positive and that it's a continuous function.
@Isa
 
+1 Adding to the confusion is the fact that students are very likely to encounter different definitions for common objects in different textbooks: in one class real numbers are complete "by definition," in the next they satisfy the last upper bound property "by definition," and so on. The student is left with the impression that there is a platonic concept of a "real number" with a hodgepodge of properties, some of which require proof and some of which don't, and no obvious difference between the two. — user7530 Aug 17 '15 at 0:38
 
Isa
@DarkVampiricAbstractArtist Is $\phi$ strictly positive and continuous?
 
Definitions are motivated, I think, not proved. — Akiva Weinberger Aug 17 '15 at 0:31
@EricSilva Just found this old comment of mine and am thinking of do Carmo
 
@DrewBrady take the vertex as the set of all binary strings a1a2...an for length n and do the same for b.
I think it's a hypercube, so you have to assign 0 and 1s
well its in {0,1} anyways.
 
huh what vertex?
 
5:48 AM
@AkivaWeinberger lol
 
11 hours ago, by Kasmir Khaan
@ACuriousMind thank you for moving his ( secret's) messages to another dimension :D he was making the chat experience hell for us for a long time :/
11 hours ago, by Semiclassical
for my part I don't actually mind the messages. I find them irrelevant to my own interests, usually, but not offensive. (by contrast, I find the random questions/appeals for help to be a bit exhausting at times)
for my part I don't actually mind the messages. I find them irrelevant to my own interests, usually, but not offensive. (by contrast, I find the random questions/appeals for help to be a bit exhausting at times)
(In case this is not clear: Kasmir Khaan is a nonweird while Semi is a weird)
 
Zee
What’s going on
 
just me rambling
carry on the maths, nothing much to worry about
 
Zee
What do you think of dynamical systems theory ?
 
I am not sure if I am familar to it, you mean the things like studies of systems of PDE and ODE?
 
Zee
5:59 AM
I don’t know much about it too frankly
I think it’s strictly about systems of ODEs
 

« first day (2783 days earlier)      last day (2228 days later) »