« first day (2591 days earlier)      last day (2444 days later) » 

3:00 PM
Hi, I was doing a quick programming thing and came across something: When I calculate area of a circle with radius r. Then why is a circle's area with radius r+1 and a circumference of a circle with radius r+0.5 the same?
 
$\pi (r+1)^2 = 2\pi\left(r+\frac{1}{2}\right)$
which implies $(r+1)^2 = 2\left(r + \frac{1}{2}\right)$
 
@Annabelle Did you test this with other values of r? Like, say, r = 10?
 
etc just solve for $r$ if you find a solution for all $r$ you're done
 
Seems to happen with every value?
 
otherwise it won't work
 
3:03 PM
mornin guys
 
hi @dodsy
 
@Annabelle I'd guess you have a bug in your code. The area of a circle with radius 10+1 is $121\pi$ and the circumference of a circle with radius 10+0.5 is $21\pi$.
 
can anyone tell me if the above I wrote is correct?
 
Hmm @El'endiaStarman I'll check that!
 
@Semiclassical first classes tomorrow! I saw my first physics assignment due on the 22nd and it seems fairly difficult.
 
3:08 PM
kk
my student's first HW is pretty brutal looking
just in terms of length
 
If both the roots of $x2-ax+a=0$ are greater than 2 then determine the interval in which $a$ lies.
I used the following two conditions:
$\alpha\beta>2$ and $D\ge0$
Thus I obtained $a>2$ and $a \in (-infty,0)\cup(4,infty)$ and their intersection gives: $(4,\infty)$ but answer shocked me. It says: "No such a exists"
 
oh, on that note. remember how my uni bureaucracy had me in a lurch?
 
yeah how's that going?
 
well uh
everything's fixed now :|
 
Uni bureaucracy ftw
 
3:09 PM
Yay!
 
Recently thinking about a statistical mechanics stuff:
 
basically the next day the tuition people got back to me and said my refund had been approved. apparently a director intervened to get it expedited
 
even teaching the first year class?
 
in The h Bar, 28 mins ago, by Secret
[The Cube]
 
that's awesome dude
 
3:10 PM
and then the front office realized they needed a different TA for the quantum course
so I landed that too
 
omg
awesome dude
everything is solved!
 
so in one day I went from in a lurch to "oh, everything's fine."
 
2 mins ago, by Abcd
If both the roots of $x2-ax+a=0$ are greater than 2 then determine the interval in which $a$ lies.
I used the following two conditions:
$\alpha\beta>2$ and $D\ge0$
Thus I obtained $a>2$ and $a \in (-infty,0)\cup(4,infty)$ and their intersection gives: $(4,\infty)$ but answer shocked me. It says: "No such a exists"
$a$* (before exists)
$x^2$*
 
Oh sorry I made a mistake in the question I was asking: @El'endiaStarman
 
@Abcd it often helps to think about quadratics in terms of 'completing the square'
 
3:12 PM
The actual thing is, why is a circumference of the circle with r+0.5
The same as the difference between the area of circles with r and r+1?
 
which in this case would be $(x-a/2)^2+a-(a/2)^2=0$
 
If both the roots of $x^2-ax+a=0$ are greater than 2 then determine the interval in which $a$ lies.
I used the following two conditions:
$\alpha\beta>2$ and $D\ge0$
Thus I obtained $a>2$ and $a \in (-\infty,0)\cup(4,\infty)$ and their intersection gives: $(4,\infty)$ but answer shocked me. It says: "No such a exists"
@Semiclassical okay, and that is false $\implies$ no such a exists
@Semiclassical Can't $a$ be imaginary?
 
well, if $a$ can be complex then talking about an interval isn't really sensible
 
What did your adviser have to say about the whole ordeal? @Semiclassical :^)
 
so presumably it means that there's no interval on the real axis.
he was just glad to see it resolved
he wasn't too worried about it, though.
 
3:14 PM
I see.
 
which I sorta agree with and sorta don't
I mean, it did get resolved
but that wouldn't have happened had I not investigated on my own
 
Okay @Semiclassical. By the way, I am the old Abcd (using my new account now.). Hope you recognised.
 
so
shrug
@Abcd I figured.
@Abcd I think my completing the square suggestion was probably unnecessary, but I think I can use it like this.
 
@Semiclassical What's the other method then>?
 
Let $r,s$ be the roots. Then your quadratic must be of the form $(x-r)(x-s)$.
 
3:17 PM
yes
 
If you multiply that out, you get $x^2-(r+s)x+rs=x^2-ax+a=0$
so you have $r+s=a$ and $rs=a$.
Now, if $r,s$ are both supposed to be bigger than 2, what can you say about $a=r+s$?
 
@Annabelle Oh, that makes much more sense. Well, $\pi (r+1)^2 - \pi r^2 = \pi r^2 + 2 \pi r + \pi - \pi r^2 = 2 \pi r + \pi$, and $2 \pi (r + 0.5) = 2 \pi r + \pi$.
 
@Semiclassical greater than 4
 
you can say more than that.
not much more, but that's not the right bound
 
@El'endiaStarman Yeah I got it now too, they both reduce to 2r+1 in the end
well
if you cancel out the pi on both sides
 
3:20 PM
i mean, if both were equal to 2, that'd be $2+2=4$...
 
@Semiclassical yes, edited. Then?
 
right.
 
@Semiclassical that's it?
 
no.
you also have to have $rs=a$.
 
@Semiclassical rs>4
 
3:22 PM
but that's a problem. suppose r=s=3, so r+s=6.
then rs=3^2=9 is bigger than 6.
(that's an example, not a proof)
 
@Semiclassical is it always true? Is there a theorem that says so too?
 
hrm.
there's a theorem like that, but I'm not sure it's the right one in this context.
moreover I don't think we should have to cite any external theorem.
 
@Semiclassical I forgot the full form of hrm, what is it? :p
@Semiclassical Okay, understood.
 
basically, we want to argue that $r,s>2$ implies $r+s\neq rs$
 
@Semiclassical But we have $>$ sign not $=$ sign in the inequalities. .
 
3:26 PM
well, sure.
 
What's the full form of hrm? Google says, "Human resource management"
 
i think that it's probably true that $r+s<rs$ if $r,s>2$.
uh
hrm is just a vocalization
like 'hmm' or 'ummm'
 
Okay.
 
hrmmmm
 
@Semiclassical Okay then we would get 4<4 which is false.
so no such a exists.
 
3:27 PM
stop.
that's not valid reasoning
 
then?
 
to have r+s=4, you'd need r=s=2
which are both equal to 2, not greater.
 
Oh yes.
 
The statement we want to prove---but haven't yet---is this
If $r$ and $s$ both exceed $2$, then $rs$ exceeds $r+s$ and therefore there's no $a$ that can equal both.
I think that's true.
for instance, if r=4 and s=3 then rs=12 exceeds r+s=7
 
@Semiclassical Please recheck.
 
3:30 PM
deeerp
fixed
totally meant to write what you had
 
@Semiclassical I knowww
 
anyways
 
@Semiclassical This reasoning can be wrong if $a<0$
 
sure, but that's not allowed. both roots are greater than 2, so their sum $r+s$ is at least 4
so $a$ is at least 4
 
ok
 
3:34 PM
I feel like this is too complicated, though
 
Semiclassical, do you know the possible types of pathways for excited rotational energy levels to relax, particularly those that directly relax from any excited state with correct symmetry to the ground state?
 
and I think I have a simpler argument, at least in terms of intuition
@Secret nope
 
ok nvm
 
let $y=x^2-ax+a$ be our quadratic equation
some quick facts about such a quadratic. first off, it opens vertically upwards
because it's $+x^2$ not $-x^2$
second, every quadratic has an axis of symmetry
lastly, observe that when $x=0$ or $x=a$ then $y=a$
given that, try to make a sketch of what such a quadratic would look like. you'll run into a problem.
 
3:57 PM
@Semiclassical previous method was better.
 
yeah...
plus, I realize now there's a simple way to get to the end of the previous method
if $a=r+s=rs$, then we can solve the last equation to get $r$ in terms of $s$
oh, even better. I can manipulate that last equation to $rs-r-s+1=1$
but I can factor the left-hand side to get $(r-1)(s-1)=1$
but if $r,s$ are bigger than 2, then r-1,s-1 are bigger than 1. see the problem?
 
The equation $ax^2+bx+c=0$ has real and positive roots. Prove that the roots of the equation $ax^2+a(3b-2c)x+(2b-c)(b-c)+ac=0$ are real and positive.
Now we have,
$b^2-4ac>0$.
$(-b/a)>0$
$c/a>0$
I need to prove :
$a^2(9b^2+9c^2-18bc)-4a^2(2b^2-3bc+c^2+ac)>0$

I managed to prove all the terms positive (you may ask me if you suspect) but am stuck on the last term $-4a^3c$. How do i prove that $-4a^3c$, too, is positive?
 
ew.
not going to lie, that problem just seems dickish
that's not a judgment on how hard it actually is, just on how it looks.
 
@Semiclassical Only first 2 line are the statements of the problem.
 
I know.
 
I think the trick is that the first three terms in the second equation can actually be factored.
 
@Semiclassical But when my method is working, why should I retreat at the last step?
 
well, for one, you seem to be running into a problem that can't easily be surmounted
moreover, I'm pretty sure that in your problem that $-a^3c$ is never positive.
If $c/a>0$, then either $c,a$ are both positive or both negative.
but then $ac$ must also be positive, so you've got $-(a^2)(ac)$. a^2 is positive, and so is ac, so that's negative.
 
@Semiclassical Have I made a mistake in expansion?
 
4:11 PM
@Semiclassical hmm = yes or no? (Sorry)
 
yeah, you have. $(3b-2c)^2$ isn't $9b^2+9c^2-18bc$
hmm = I'm thinking
 
@Semiclassical Okay, I had copied the question wrongly in the notebook, but that's not going to affect the result that I had previously obtained.
 
@Balarka: This one's for you.
Hi Semiclassic and Abcd.
 
@TedShifrin Hello! (from new account). I hope you remember me :p
 
@BalarkaSen This is the Cartan formula for $\mathscr L_X$, @Danu. I use this approach to do variational stuff with moving frames all the time.
You have to decide if $\gamma$ is moving, though.
I can't tell you're any different, @Abcd.
 
4:17 PM
@TedShifrin Profile picture changed (better than the dirty brown of the old account) and reputation decreased.
 
@Abcd right.
other error: you should have only $4a$ not $4a^2$
I'm not convinced that entirely resolves it, though.
 
The equation $ax^2+bx+c=0$ has real and positive roots. Prove that the roots of the equation $a^2x^2+a(3b-2c)x+(2b-c)(b-c)+ac=0$ are real and positive.
Now we have,
$b^2-4ac>0$.
$(-b/a)>0$
$c/a>0$
I need to prove :
$a^2(9b^2+9c^2-18bc)-4a^2(2b^2-3bc+c^2+ac)>0$

I managed to prove all the terms positive (you may ask me if you suspect) but am stuck on the last term $-4a^3c$. How do i prove that $-4a^3c$, too, is positive?
@Semiclassical correction in question^ ($a^2$)
 
the last term in your last inequality should be 4a not 4a^2
oh, was it supposed to be a^2 x^2 originally?
 
@Semiclassical No, it should be $4a^2$
@Semiclassical yes
 
in that case I'll withdraw that complaint.
in that case I'm firmly to 'i dunno' territory
I do think this approach seems harder than it needs to be, but I'm not certain on that
I guess here's what I notice: $a^2x^2+a(3b-2c)x+(2b-c)(b-c)$ factorizes nicely.
Do you see how?
What?
 
4:24 PM
@Semiclassical No
 
it may be more apparent if I introduce the following labels: $p=2b-c$, $q=b-c$
in which case the polynomial becomes $a^2x^2+a(p+q)x+pq$.
 
@Semiclassical an ac in the end too
 
no. I'm only talking about the first three terms.
 
ok
 
and my claim is that $a^2 x^2+apx + aqx +pq$ can be factorized
 
4:27 PM
I am eager to know why my method is wrong :(
 
Well, I think the issue is this. It's easy enough to argue that $a^2(3b-2c)^2$ is positive.
What's hard is showing that it's bigger than $4a^2(2b^2-3bc+c^2+ac)$.
 
@Semiclassical I expanded that too and proved all of them were positive. And was left only with 4a^3c
 
How?
My main bone of contention is that you seem to be trying to prove that $-4a^2(2b^2-3bc+c^2+ac)$ is positive.
And I very much doubt that's always true.
 
@Semiclassical $5a^2b^2 $ is +ve
5a^2c^2 is +ve
$-18a^2bc$ is +ve @Semiclassical
And so on...please ask me of any individual term. I can't type all of them :( @Semiclassical
 
Slow down. You're writing stuff that makes no sense.
You have $-4a^2(2b^2-3bc+c^2+ac)$, yes?
 
4:34 PM
@Semiclassical We have more stuff too. $a(3b-2c)^2$. I expanded both.
 
ah. hadn't taken that into account
lemme see (checking myself)
when I expand all that out, it simplifies down to $$a^2 (3 b - 2 c)^2 - 4 a^2 ((b - c) (2 b - c) + a c)=a^2 b^2 - 4 a^3 c$$
 
@TedShifrin It seems to me that $S^{\omega+1}$ is clearly homotopy equivalent to $S^\infty$. I'm pretty sure you can nuke that Whitehead's theorem argument.
 
(I'm using mathematica to do the algebra so as to avoid errors)
 
On the other hand Grumpy Parsnip posted a comment below that makes it obvious.
 
in which case I can factorize that as $a^2(b^2-4ac)$, which is obviously positive under the conditions of the problem.
 
4:37 PM
@Semiclassical And Bob's your uncle
 
not quite. what that shows is that the roots are both real.
 
@Semiclassical My mistake was: copying the question wrong. (3b-3c) instead of (3b-2c)
 
but you also want to show they're positive, don't you?
gotcha
 
@Semiclassical yes
 
right. on that I'm less sure.
if both roots are real and positive, then $-b/a$ should be positive.
so if both roots of the second equation are real and positive, then we'd better have $-a(3b-2c)/a^2=-3b/a+2c/a>0$ as well
but, ah, you already had $-b/a>0$ and $c/a>0$.
so I think that clinches it
 
4:41 PM
@Semiclassical true. $Q.E.D.$
Thanks.
 
Now, lemme point out how I'd have approached it :)
 
Okay
 
the second quadratic equation can be written thusly:
$$a^2x^2+a(3b-2c)x+(2b-c)(b-c)+ac=(ax+2b-c)(ax+b-c)+ac=0$$
now, since $c/a>0$, we've got $ac>0$ as well
 
@Semiclassical My method was better.
 
Yo
 
4:44 PM
now, suppose $ac$ weren't there. then we'd have the quadratic $(ax+2b-c)(ax+b-c)=0$, which has real roots and opens upwards
...aw, crud. there's a hole in where my argument would be going.
I still feel there should be a good way to take advantage of that factorization
but coulda-woulda-shoulda
 
@Daminark Oy
 
hey anyone know if $A\in {\rm M}_n(\mathbb Z_2)$ and $u,v\in\mathbb Z_2^n$ then is $M\cdot(u+v)\stackrel{?}=M\cdot u+M\cdot v$
 
matrix multiplication always acts linearly on vectors.
 
I was told to prove this using M as an automorphism on $\mathbb Z_2^n$
does that makes sense?
 
Well, if $M$ is an automorphism, it should preserve the group operation.
 
4:52 PM
I am doubtful because $\bar 1+\bar 1=\bar 0$
 
what's $\overline{1}$ supposed to be here?
 
$\mathbb Z_2 =\{\bar 0, \bar 1\}$
equivalence classes
 
ah.
I'm not seeing your objection in any case.
Also, presumably you meant $A\cdot(u+v)$ earlier.
 
yes, exactly
 
I guess I'm unclear what you're after. Are you trying to prove that $u\mapsto A\cdot u$ is an automorphism, or are you trying to obtain $A\cdot(u+v)=A\cdot u+A\cdot v$ as an implication of it being an automorphism?
 
4:55 PM
the latter one
 
okay. But I'm forgetting---isn't an automorphism in this setting supposed to preserve the group operation?
That's certainly part of the definition of iso- or homomorphism
 
yes exactly
that's why I am having problem how to show that
 
Then I'm again confused. $A\cdot(u+v)=A\cdot u+A\cdot v$ is precisely the statement that $u\mapsto A\cdot u$ preserves the group operation
 
number of real roots of $ax^2 + 2b|x|+c=0$(where a,c,b are distinct positve real numbers) varies from 0 to 4 right? But answer given is 0. Is answer given wrong? Please verify.
 
(which in this case is addition)
@Abcd no
a is positive, x^2 is nonnegative, b is positive, |x| is nonnegative, c is positive
 
4:59 PM
@Semiclassical so?
 
well, then ax^2 and 2b|x| are both nonnegative
 
@Semiclassical I get it you define $A\in {\rm M}_n(\mathbb Z_2)$ and $\phi_A : \mathbb Z_2^n\to\mathbb Z_2^n$ s.t. $x\mapsto A\cdot x$
 
so you've got nonnegative + nonnegative + positive. can that every be zero?
 

« first day (2591 days earlier)      last day (2444 days later) »