« first day (1675 days earlier)      last day (3345 days later) » 

7:02 PM
ugh, i need to stop putting off typing up these notes
 
OK, time to go get rid of some of the stuff from my house and spend the weekend organizing for what gets carted off ... Back later. Have a good weekend, all.
 
@semi What field you working on?
 
condensed matter physics, at least in principle. right now it's mostly just math stuff, which i'm fine with
 
Oh wait, are you in the math or physics dept?
 
physics
 
7:05 PM
OMG, I did not know.
 
@Semiclassical What do you do in condensed matter phys ?
 
i guess the thread running through the work i've done with my advisor is: how do we interpret some of the calculations we do in various systems as period integrals on a riemann surface, and what technology can we import from pure math in order to tackle stuff
 
Sounds more math-y than physics-y
 
mk mk. No 'real' physics so far then/
 
7:07 PM
well, here's the three cases we've juggled
one: the statistical mechanics of an ion channel, which leads to a (potentially non-hermitean) schrodinger-type equation whose spectrum describes stuff like the pressure and such
which we analyzed by semiclassical methods, which (for us) meant finding good ways to calculate certain integrals
 
semiclassical methods
 
Bohr-Sommerfeld quantization, mostly.
two: the splitting of energy levels in magnetic molecules with large spins. (one gets a kind of semiclassical description in terms of the size of the spin of the molecule.)
 
semiclassical description
 
spin coherent states, to be precise
the third model, which i'm trying to write up something on now, is the SSH model of polyacetelene
which is known to be a simple example of a 1D topological insulator
so, the underlying motivation may be more mathematial than physical; but we're trying to apply it to genuine physical systems
 
ADG
Hello
did you saw this?
someone trying to use Euler's series/Taylor Series etc. to prove $\cos0=1$ but he doesn't realize that these techniques involve circular reasoning
 
7:18 PM
on the one hand I know of an interesting product of closed manifolds. on the other hand, i'm sort of morally opposed to this question.
 
ADG
anyone on my msg?
 
how is this not indeterminate form? goo.gl/mqK1Oi
 
Can anyone think of a nice application of the fact deformation retracts are closed under pushouts?
The formal statement is that if $$\begin{matrix}A &\xrightarrow{f}& B \\ \ \ \downarrow i& & \ \ \downarrow j \\ C &\xrightarrow{g}& D \end{matrix}$$ is a pushout square with $A$ a deformation retract of $C$, then $B$ is a deformation retract of $D$.
 
When taking limits, you look at points nearby where you're going, not the point itself, @JoeStavitsky. For $\Delta x \neq 0$, that's perfectly well-defined - it's just zero.
 
can you give me an example of an element P in $ \mathbb{R} \left[ X \right]$ such that for all $ U \in \mathbb{R} \left[ X \right] $ $ U^{"}-XU^{'}\neq P $
help
 
7:30 PM
@MikeMiller, right, I understand intuitively, but how is that algebra valid? I mean, you're dividing by 0, no? I get that the limit says it's not zero, but it doesn't look like the kind of thing that would be correct on an exam :/
 
No, you're never dividing by zero, @JoeStavitsky.
You are only evaluating this at points where $\Delta x \neq 0$ - and then taking the limit as it goes to zero.
And at points where $\Delta x \neq 0$, it's just zero over some nonzero number.
 
a-HA!
enlightenment!
 
ADG
If you use something(=1) to prove something else(=2) the result must not be included as an assumption in something(1)??
@MikeMiller have your say?
is it true?
 
can you give me an example of an element P in $ \mathbb{R} \left[ X \right]$ such that for all $ U \in \mathbb{R} \left[ X \right] $ $ U^{"}-XU^{'}\neq P $
help
 
@user161440 think about U''-XU'
what is something you can say about that polynomial?
 
7:37 PM
?
 
what is a feature of polynomials that is sometimes discussed? it came up in your latest question, for instance.
 
has degree $n
?
do you mean the dgree
 
degrees, yes. what degree can $U''-\color{Red}{X}U'$ have?
more importantly, what degree can't it have?
 
if U has degree n
 
don't think too hard about U's degree
just look at that red X there in U''-XU' - what implications does it have for the degree?
 
7:41 PM
degree 1
X has degree 1
 
what does X's presence mean for U''-XU'?
 
think.
 
then the degree is grater than or equal to 1
?
 
ADG
?
 
7:46 PM
@anon
 
@user161440 yes! (or else U''-XU' is 0 if U' is 0). so what's an example of something U''-XU' can't be?
 
we already know U''-XU' can be 0.
 
if the degree is 0?
 
pay attention: we've concluded U''-XU' is either 0 or has degree >= 1. what is something that is neither 0 nor has degree >=1?
@user161440 if you mean nonzero constant, then yes.
for instance, U''-XU' cannot be 1.
 
7:50 PM
@anon -infinty
-infinity?
 
ADG
?
 
@user161440 why are you throwing a random word at me? yes, I know -inf is conventionally the degree of the 0 polynomial, but why are you bringing that up? I don't see the relevance. Did you read what I just wrote above?
 
@anon You gave in
 
because you say somethin than that is not zero and not greater than 1?
 
@PaulPlummer constants have degree 0, so I thought the user had finally figured it out
@user161440 we're talking about polynomials and their degrees remember?
you can't carry on a mathematical conversation or argument if you randomly forget what we're talking about!
5 mins ago, by anon
pay attention: we've concluded U''-XU' is either 0 or has degree >= 1. what is something that is neither 0 nor has degree >=1?
 
7:53 PM
@anon sorry
 
i know this one!
 
@anon if I consider the the function
f(U)=U''-XU' from R[X] to istsel
to itself
can I determine f(R[X])
can I determine f(R[X])
 
8:15 PM
If $U = a_0 + a_1X +\cdots + a_nX^n$, then $U' = a_1 + 2a_2X + \cdots + na_nX^{n-1}$
so $XU' = a_1X + 2a_2X^2 + \cdots + na_nX^{n-1}$.
 
then ?
f(R(X)) is going to be a set
!
can I find it ?
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ I'm front
 
I think you would have to try to find it before you can determine whether or not you can find it. Also what do you mean by find it, $f(R[X])=f(R[X])$
 
8:36 PM
@user161440 if P is in f(R[X]) then we can recover its preimage U. thus it makes sense to ask when we can recover U given U''-XU'. what information can you recover about U given U''-XU'?
 
@anon please can you give me the defnition of $\limsup_{x\rightarrow-\infty} F(x)$ using $\varepsilon$ thank you ?
 
@Vrouvrou well, $\limsup\limits_{x\to-\infty}=\lim\limits_{X\to-\infty}\,\sup\limits_{x<X}$
so if you can translate lim and sup individually into an epsilon definition then you can do it with limsup as well
 
@DanielFischer Could you take a look at my question? cs.stackexchange.com/questions/40145/execution-time-of-function
 
...
 
8:54 PM
I know that f(R[X]) is going to be a set of polynomial do they have to satisfy a certain condition ? in other word if P satisfy the proprety....then P is in f(R[])
 
@anon but i can't because i don't know if the limite is equal to $\pm\infty$ or a constant ?
@anon can you see my question:math.stackexchange.com/questions/1178050/…
 
9:08 PM
@anon you see it?
 
9:21 PM
Hello
may I ask a question here ?
seems nobody is here
since it says don't ask to ask I'm just gonna ask
I'm having a bit of a problem, I have a linear system and I'm asked to know when there is one solution or none or infinite
I've put it in a matrix and got this as the final line
0 0 4-a^2 | 2-a
what I think is right is if 4 - a^2 = 2 -a then there is a single solution
4 - a^2 = 0 but 2 - a doesn't equal 0 then there are no solutions
if 4 - a^2 = 0 and 2 - a = 0 then there are many solutions
is that right ?
 
Almost. Your first bit is not quite right. It should just be "if 4-a^2 is nonzero, then there is a single solution." (Why?)
 
aha so that was my mistake
because it's saying that n * z3 = something
so z3 is specific ?
 
Right. And if n is not zero, you can just divide.
 
so single solution when a not = 2, - 2
no solution when a = -2
and many solutions when a = 2
that should be the answer.
thanks mike
I have another question, sorry to bother you I'm solving an exam but I don't have the answers for it to verify my solutions.
 
9:42 PM
No problem with the above. And sorry, I've got to go; hopefully someone else can help.
 
if (A-I)^2 = 0 prove that A is invertible
(A-I)(A-I)=0 => A-I = 0 => A=I => det(A)=1
ok, take care thanks for the help.
 
9:57 PM
@Ali.B For matrices, $M^2 = 0$ does not imply $M = 0$. Expand $(A-I)^2$ and rearrange so that you get $A(\dotsc) = I$ to see that $(A-I)^2 = 0$ implies the invertibility of $A$.
 
ok
A^2 - 2A + I = 0 => A(A-2) = I => A is invertible is this right ?
 
yes
 
@robjohn you're far too silent these days.
 
@Ali.B Almost. You have a sign error in there.
 
-I thanks
A(2-A) = I
 
10:05 PM
yup
 
1
Q: Execution time of function

evindaThe following pseudocodes are given. quicksort(A,p,r) if p<r then q<-partition(A,p,r) quicksort(A,p,q-1) quicksort(A,q+1,r) partition(A,p,r){ x<-A[r] i<-p-1 for j<-p to r-1 if A[j]<=x then i<-i+1 swap(A[i],A[j]) swap(A[i+1],A[r]) return i+1 Wh...

Could you take a look at my question?
 
10:28 PM
@ɧɿρρԹʅȝՇԵՐՎԾՌ ii is eaten already
 
11:00 PM
@Chris'ssis Sorry, working on something, but I've had a lot of household things going on the last week.
 
@DanielFischer I want to solve the recurrence relation $T(n)=7T\left( \frac{n}{3} \right)+n^2$.
I have tried the following:
$a=7 \geq 1, b=3>1, f(n)=n^2$
$n^{\log_b a}=n^{\log_3 7} \approx n^{1.77}$
So $f(n)=n^2 = \Omega(n^{\log_b a+ \epsilon})$, for $\epsilon=0.23>0$
$a f \left( \frac{n}{b}\right)=7 \left( \frac{n}{3}\right)^2=\frac{7}{9}n^2\leq cf(n)$ for any $c \in [\frac{7}{9},1)$.
So from the case 3 of Matser theorem, we have that $f(n)=\Theta(f(n))=\Theta(n^2)$.
@DanielFischer Could I jusify it like that or should I avoid saying that $n^{\log_3 7} \approx n^{1.77}$ ?
 
@evinda I'd avoid $n^{\log_3 7} \approx n^{1.77}$. $\log_3 7$ is exact and short.
 
@DanielFischer So should I write the whole number wolframalpha.com/input/?i=log_3%287%29 ?
 
@evinda No. Just write $\log_3 7$. It's irrational, so you can't write an exact decimal representation anyway. You can, if you want, add a parenthetical remark $\log_3 7 \approx 1.77$ (or if you prefer you can use four or five or two significant digits, all you need is that it's $< 2$, so anything showing that is good enough).
 
A ok.. @DanielFischer So you mean that I could write it as follows, right?
$n^{\log_b a}=n^{\log_3 7}<n^2 (\log_3 7 \approx 1.77)$

So $f(n)=n^2 = \Omega(n^{\log_b a+ \epsilon})$, for $\epsilon=0.23>0$

$a f \left( \frac{n}{b}\right)=7 \left( \frac{n}{3}\right)^2=\frac{7}{9}n^2\leq cf(n)$ for any $c \in [\frac{7}{9},1)$.
So from the case 3 of Matser theorem, we have that $f(n)=\Theta(f(n))=\Theta(n^2)$.
 
11:15 PM
@evinda Put more space between the $n^2$ and the parenthetical remark, maybe a newline. And you have a typo in the last line, Matser instead of Master.
 
Oh yes, I am sorry @DanielFischer Thank you :)
 
@robjohn OK
I'd like to reach the performance of doing no math for 3 days (at least).
People that hate math would be very surprised to find out how addictive math can be. :-)
 
11:39 PM
Addiction requires an addictive personality.
 
@infinitesimal I dropped smoking some years ago without any problem (I didn't smoke so seriously, that's true).
 
@Chris'ssis I didn't know you smoke.
 
@ABeautifulMind but that wasn't a too serious smoking. I didn't smoke for long.
 
That's what I mean. @Chris'ssis A serious addict.
 
@infinitesimal Instead I might say I'm addicted to sugar ... (it's hard to stay away from sweets)
 
11:44 PM
One who has, as they say, a "death wish" and will not stop have a certain personality type to begin with.
Would you stop eating sweets if your doctor told you that you are in danger of getting severe diabetes @Chris'ssis?
 
@infinitesimal Definitely. The problem is not the addiction, but the quantity in general. As long as you eat less, all seems fine (well, there might be some exceptions).
 
Then you're not an addict.
 
@infinitesimal The true Will produces great changes ... but I wouldn't like to fall an extremity like no sugar eating at all.
 
Indeed.
 
In the end it's a matter of equilibrium.
 
11:52 PM
Then plan ahead and start cutting back slowly now.
You know too much of anything is no good :-)
 
True.
 
you're fine, his link is smiling
 

« first day (1675 days earlier)      last day (3345 days later) »