« first day (1348 days earlier)      last day (3971 days later) » 

00:05
@robjohn Was the convergence of Chris's series determined?
@PedroTamaroff which series?
The starred one.
@PedroTamaroff you mean the one I computed to be $\frac{\pi^2}{8}$?
I mean the one whose $n$th term is $sin H_n/(nHn)$
@PedroTamaroff Oh, that's a different one.
00:08
If you click on the transcript, I suggested smething.
@PedroTamaroff If you write it as $$\sum\frac{\sin(H_n)}{H_n}(H_n-H_{n-1})$$ it looks convergent
@robjohn Ah, do some partial summation there. Works?
@PedroTamaroff it looks like it is tending toward a Riemann Sum for $\int\frac{\sin(x)}{x}\,\mathrm{d}x$
@robjohn Not entirely sure about that.
some details need to be worked, but I think it should work
@PedroTamaroff eh, the integral of the derivative is bounded, so the improper Riemann sum should be fine
00:34
@PedroTamaroff Hmm... now that I look at the details that need to be worked, it looks poor...
I think I can show that if this series converges, then $\sum\frac{\cos(H_n)}{n}$ converges...
and I don't think that one does.
00:56
I am going to sleep, guten nacht!
@WillHunting sleep well
@PedroTamaroff Now I can prove it does not converge.
@DanielFischer follow through the partial summation I mentioned a few lines up
You can show that if $\sum\frac{\sin(H_n)}{nH_n}$ converges, then $\sum\frac{\cos(H_n)}{n}$ converges. However, it is not too difficult to show that that diverges.
01:11
Is it on my end, or did that message in fact stop mid-sentence?
@DanielFischer Sorry, have an emergency, now I am on my phone, so I will continue later.
Ah, okay.
01:30
@robjohn OMG @robjohn I thought it was impossible.
@A̷n̷d̷y̷ what was impossible?
@A̷n̷d̷y̷ Ah, yes. It is quite doable with a little trig and geometry
and limits
@A̷n̷d̷y̷ That is why I wondered why the question had been deleted
01:42
Psh, $H_n$'s.
Connection between poly-time algorithms and polynomial-valued algorithms :>
0
Q: Necessary and sufficient condition for the existence of a polynomial time algorithm for problem $f: X \to Y$.

Enjoys Math$f : X \to Y$ is a deterministic polynomial-time algorithm for problem inputs $x \in X$ and problem outputs $f(x) = y \in Y \iff $there exists a polynomial $P_f \in \Bbb{Z}[x_1]$ such that $C\cdot P_f(|x|) = $ the running time on input $x, |x| = $ size of $x, C \in \Bbb{R}$. Conjecture Let $f: ...

02:03
I've added to it. I think it might be a thing
I spent about a half hour thinking about a problem that I misread. Hate it when that happens
@Karl Wat problem
You have two ideals $I,J$ of a Noetherian ring. If the extension of $I$ is contained in that of $J$ after localizing at every prime divisor of $J$, then $I\subseteq J$.
I read it as going through the prime divisors of $I$.
What is the Galois resolvent of the field extenstion $K[X_1, ... X_r] / K[S_1, ... S_r] $ ?
$S_1,...,S_r$ stand for elementary symmetric polynomials in $X_1, ... X_r$
seventeen
02:19
$3.14158^r$
@Mike Turns out the only difficulty in the problem is reading it correctly.
hello there
I would like to know how to create a abelian group table
wtf is with the "this site is safe" popup on SO links.
oh, nevermind
Um, I think P = NP guys, and here's a lemma in the proof: math.stackexchange.com/questions/751327/…
02:37
lol
02:56
@karl they want yu to know it's safe bro
howdy @Mike @Karl
hey @Ted
good talks today
Who talked to whom?
I've spent the day getting depressed grading differential geometry exams (average 54). UGH.
Oh, a conference was held at SCU today.
The BAD Math conference.
03:18
Hi @TedShifrin
I'm sorry about your poor students
I would probably be one of them unfortunately
03:35
Jeez, @Ted doesn't want to have a conversation
Can we use $AM\geq GM$ inequality to find the minimum of $\dfrac{x(1+y)+y(1+z)+z(1+x)}{\sqrt{xyz}}?$
@user127001 I was jeezing at him, not you. You're fine :)
@TedShifrin then you wouldn't want to grade my class' exam paper... average 25.. and that's 25 out of 80 have fun :)
anyway

http://math.stackexchange.com/questions/751410/prove-f-infty-a-infty-rightarrow-b-infty-is-a-bijection
Oh I thought you meant I was bothering him when I shouldn't be @Mike
03:46
@usukidoll, please help me!
@user127001 You should always brother Ted.
@Sush I wish I need help myself D:
@Mike @user127001: Ted was gone. Don't get your respective panties in a wad. Who spoke at the conference, @Mike?
any ideas...besides the sequence won't terminate... will go on forever and ever and ever
@usukidoll: I'm not reading all your argument, but the usual map is defined recursively on sets $A_n$ and $B_n$. There is no $A_\infty$, etc.
03:57
AY YAH!
so the infinity for A and B doesn't exist?!
wait a sec.. my prof wrote that on the blackboard yesterday
Right. Every element lives in some finite $A_n$ or $B_n$. Of course, now I need to rethink the whole thing, but ... Yeah.
D::::::::::::::: I've been given a dead assignment W T 3490802384048903249032
so it's impossible for $A_\infty$ to happen O_O
@Ted The best talks (IMO) were one of our own faculty's (Tamsen McGinley) and Satyan Devadoss's.
Gunnar Carlsson from Stanford also spoke, as well as a few oters.
maybe I should disprove the infinities
Others, not otters.
04:03
Cool, @Mike. People I don't know, except Carlsson.
Otters might have been cute :D
Well, Bay Area Discrete Mathematics. I imagine you wouldn't know most.
Yeah, I'm far from discrete. Probably far from discreet, too.
There was also free dinner.
Nice. :)
Anyway, it was fun. You have fun today too?
04:11
hmm what about proving $f_n: A_n \rightarrow B_n$ @TedShifrin since it's obvious now that $\infty$ doesn't even exist.
No. See above. But bedtime now ... Tennis comes early in the am.
Oh well. See ya.
@usukidoll: I don't understand the notation in your argument, but it's very late for me now.
The usual proof, which I love, is a picture of nested rings, @usukidoll. Pictures are good!
Night, all!
04:13
night
 
1 hour later…
05:24
Anyone here?
05:48
Hi
06:08
@TedShifrin Ah, I didn't see this one. Yes, this is fine. It was the previous estimate that I was responding to before.
@Chris'ssis It diverges.
06:25
@robjohn hey, you wouldn't happen to know if the general jordan decomposition of the tensor product (equivalently, kronecker product) of two jordan blocks is known would you?
A user asks a single question here and 4 hours later their account is deleted. Strange, isn't it?
i.e. $J_{n,\lambda}\otimes J_{m,\mu}=J_{k_1,\lambda\mu}\oplus\cdots J_{k_r,\lambda\mu}$ where the multiset $\{k_1,\cdots,k_r\}$ varies with $n$ and $m$ and potentially $\lambda$ and $\mu$
@MartinSleziak strange how?
I imagine users do that all the time
@anon This was among the first results when putting "kronecker product" and "Jordan blocks" into Google.
@anon It is not that easy to delete your account by yourself. (IIRC you need assistance of SE team.)
google is giving me an error message in I don't even know what language, but I presume it's saying that whatever page you linked is not for me to see
@MartinSleziak ah, so slightly strange. still a number of explanations.
however refreshing allows me to see the page
brilliant, google
@anon I believe this might be in Horn and Johson's book.
Johnson*
06:41
Weird, that account being deleted
@MartinSleziak thanks. I suppose I wasn't getting good results because I searched tensor product.
Neither meta nor main site is working at the moment. Is it just me, or do other people have the same problem?
same
same!
other SE sites seem to also be down (checked infosec and crypto beta), although overflow seems unaffected (not sure if they count as SE)
06:53
I was in the middle of posting a question on another SE site :-(
07:04
@Sawarnik, yes the same here!
Why is the chat working then?
Haha! Don't know!
different network
Chat has to go on in any situation!
@anon, thanks.
@Sush It would be better if it didn't. Will save everyone's time.
07:06
@anon How's it going man?
meh
@anon Why meh?
why not?
@anon Because if you're generic feeling is apathetic/below-average, maybe you should reevaluate things ;)
your* :)
07:53
@anon sorry, I don't
@MartinSleziak I believe it is easier if you have not done much (have little reputation).
@robjohn I was wondering if you could help me with the question I posted earlier
I'm in circles :X
@usukidoll which question?
2
Q: Prove $f_\infty: A_\infty \rightarrow B_\infty$ is a bijection

usukidollI am using the Cantor-Schroder-Beenstein Theorem to prove $f_\infty: A_\infty \rightarrow B_\infty$ is a bijection. The cases of $f_+: A_+ \rightarrow B_+$ and $f_-: A_- \rightarrow B_-$ being bijections are already done. The theorem states that for any sets $A$ and $B$, if there exist inject...

my head just hurts on this like I sort of get it but agdksajflsdak;fjsda all I can think about is that the sequence on $\infty$ doesn't end so it goes on forever
and since it's a bijection and it's a surjection and an injection
08:10
@usukidoll So this form starts with two injections?
two injections? yeah that's what the original question had and then by the injection def I could map it
and then I viewed my notes and try to figure out what the sequence is
so for injection and surjection the sequence for $A_-$, $B_-$, $A_+$, and $B_+$ stops at a certain point
but for the infinity... it keeps goin
so how do I draw that sequence and then create a surjection and injection map so that in the end the infinities is a bijection
08:28
this is super hard... how can the bijection infinity proof be two lines
?
that was one of the hints
and the other is to use the previous example as a guideline that's about it
ARgh
X___X!
09:29
@usukidoll So you see that applying $f$ to $A-g(B)$ and $g^{-1}$ to $g(B)-g\circ f(A)$ creates a bijection from $A-g\circ f(A)$ to $B-f\circ g(B)$, right?
sort of everything went fast in the lecture...
like this and that and wham prove this
so I had to view the notes a bit
What does $\sum_{4}^0$ mean? This is a question in concrete mathematics by Knuth et al. I think it does not mean anything unless we assign meaning to it.. it's just notation that can mean anything.
Correct?
ughhhh bangs head
@usukidoll so think of it this way...
Greetings
@robjohn That series actually converges.
09:31
apply f to a infinity g(b) and g^-1 to,,,,
gawd D:
@usukidoll $B-f(A)$ cannot be gotten by using $f$ on any element of $A$
@usukidoll what does $B-f(A)$ mean?
$B_- f(a)$
a function a is in $B_(-)$
the set of things in $B$ that cannot be gotten by applying $f$ to elements of $A$
09:33
?
???????????
@Chris'ssis it doesn't...
so b is unique
@robjohn It does ... :-)
we can't grab anything in B if we apply f to A elements?
arghhhhhh
this is so hard
@usukidoll $f(A)$ is the set of all elements in $B$ that are gotten by applying $f$ to elements of $A$, right?
09:36
ok wait so f(a) has elements that belong to B only if we apply the f to the elements of A
?>!
@usukidoll so $B-f(A)$ is the complement...
like oh ! like $ B \backslash f(A)$ ?
that is the same thing.
dang slash
$\backslash$
$B-f(A)=B\setminus f(A)$
09:37
by complement def all we have is B
so what if for $A_- f(B)?$ is that $A \backslash f(B)$
and all we have is A?
this is the theorem... cut-the-knot.org/WhatIs/Infinity/Bernstein.shtml so what happens for the infinity parts since the sequence will not terminate in A or B
Assume that
$$
\sum_{k=1}^\infty\frac{\sin(H_k)}{kH_k}\tag{1}
$$
converges. Then
$$
\begin{align}
&\sum_{k=1}^\infty\frac{\sin(H_k)}{kH_k}\\
&=\sum_{k=1}^\infty\frac{\sin(H_k)}{H_k}(H_k-H_{k-1})\\
&=\sum_{k=1}^\infty\left(\frac{\sin(H_k)}{H_k}-\frac{\sin(H_{k+1})}{H_{k+1}}\right)H_k\\
&=\sum_{k=1}^\infty\frac{\left(H_k+\frac1{k+1}\right)\sin(H_k)-H_k\left(\sin(H_k)\cos\left(\frac1{k+1}\right)+\cos(H_k)\sin\left(\frac1{k+1}\right)\right)}{H_{k+1}}\\
&=\sum_{k=1}^\infty\frac{H_k\sin(H_k)\left(1-\cos\left(\frac1{k+1}\right)\right)+\frac1{k+1}\sin(H_k)-H_k\cos(H_k)\sin\left(\frac1{k+1}\right)}{
@robjohn wow, let me read :-)
oh gawd d:
@Chris'ssis there is a part that I forgot to delete...
I am so confused x....x
09:42
@Chris'ssis the part around $(5)$ and $(6)$ is unneeded now...
I changed the later part of the argument.
@robjohn OK
@Chris'ssis Oh, no... strike that, they are needed. Sorry
@robjohn OK
@usukidoll you need to stop interrupting when someone is explaining things. It makes it hard to complete an explanation.
09:45
oh ... sorry
The mean square is being mean, lol
@usukidoll: so do you understand that $B-f(A)$ is the set of all elements of $B$ that cannot be gotten by applying $f$ to any elements of $A$?
by complement def... yeah
but for $A-f(B)$ that would mean that the sets of all elements of A can't be retrieved by applying $f$ to any elements of $B$ r
@usukidoll @usukidoll: okay, so the only way to cover those elements is by $g^{-1}$
the inverse of g
09:48
@usukidoll $f$ doesn't act on $B$
doesn't act?
I"m here XD.. just reading and piecing everything together
@usukidoll $f$ acts on $A$ and $g$ acts on $B$
oh so maybe A is its own mapping which is f and B has its own mapping g
@usukidoll so $A-f(B)$ doesn't make sense... $A-g(B)$ is probably what you mean
yeah
so there are A elements that aren't retrieved in g(B)
by complement def
09:51
so we use $g^{-1}$ to map from $g(B-f(A))=g(B)-g\circ f(A)$ to $B-f(A)$
we use the inverse of G to retrieve the elements of B that can't be gathered in f(A)
and then there's a composite g o f(A)
Hello, I have a small question, hopefully you can help
Question: Is following subset of $\mathbb{R}^{3}$ also a subspace. Explain why and why not.
$$U=\left \{ (x,y,z)\in \mathbb{R}^{3}|z^{2}=x^{2}+y^{2} \right \}$$

I proved that the subset is not empty as it contains $\vec{0}$ and it satisfies the condition since 0^{2}=0^{2}+0^{2}
I also proved that the subset is closed under scalar multiplication since $\alpha ^{2}z^{2}=\alpha ^{2}\left ( x^{2}+y^{2} \right )\therefore z^{2}=x^{2}+y^{2},\alpha \in \mathbb{R}$

However, I am not sure whether it is closed under vector addition or not.
@usukidoll yes, and similarly we use $f$ to map from $A-g(B)$ to $f(A-g(B))=f(A)-f\circ g(B)$ since $A-g(B)$ cannot be gotten by applying $g$ to any element of $B$
so maybe in the end the A and B infinities won't have any elements at all?
@usukidoll they might, but on those sets $f$ and $g$ are bijections.
09:56
B-f(A) we have elements in B only
A - g(B) we have elements in A only.. and there are elements... so we have to show that for the infinities the set f and set g is also a surjection and an injection
@usukidoll you are jumping away... also f and g are functions
right they are... sorry I'm stressing out over this faints
ughhh I'm not a pro.... I am slow at this :/
so we must somehow prove that the function f and g has an injection and surjection....... because a bijection - > injection and surjection
if it's just one or the other, it's not a bijection
@usukidoll so what we've just done is to use $f$ on $A-g(B)$ and $g^{-1}$ on $g(B)-g\circ f(A)$
10:01
g composite of f(A) ... applying the f(A) to g... and by complement def that's gone
and I'm left with g(B) for g inverser
This gives us a bijection on $A-g\circ f(A)$
WHAT!
whoa!
how did that happen so fast though?
@usukidoll this is not the end solution...
what happen to the surjection? injection ?
lost like crazy
@usukidoll $f$ is given injective so it is injective and it is surjective onto its image that is $f:A-g(B)\mapsto f(A)-f\circ g(B)$ is a bijection
10:07
$f: A \backslash g(B) \implies f(A) \backslash f \circ g(B)$
by complement definition
$f: A \rightarrow f(A) $ :O
@usukidoll similarly $g^{-1}:g(B)-g\circ f(A)\mapsto B-f(A)$ is a bijection
@usukidoll and the fact that $f$ is an injection
so for $g^{-1}: g(B) \backslash g \circ f(A) \implies B \backslash f(A)$
so by complement defintion.. $g^{-1}: g(B) \rightarrow B$
@usukidoll if $f$ is injective, then $f(A-B)=f(A)-f(B)$
in general, $f(A-B)\supset f(A)-f(B)$
f is injective ...
$f(A \backslash B ) = f(A) \backslash f(B)$
$f(A) = f(A)$
@usukidoll yes, and inverse functions are always injective.
So to recap...
$g^{-1}:g(B)-g\circ f(A)\mapsto B-f(A)$ and
$f:A-g(B)\mapsto f(A)-f\circ g(B)$ are bijections.
so looking at what sets are covered...
using these two piecewise defined functions, we have a bijection from $A-g\circ f(A)$ to $B-f\circ g(B)$
@usukidoll: do you understand so far?
10:21
..
Suppose we have $f: A \backslash g(B)$ and $g: B \backslash f(A)$
then by complement def we have $f: A $ and $g:B$. Anything in B can't be retrieved by applying $f$ to elements of $A$. $f$ acts on $A$ and $G$ acts on $b$. Now suppose we have $g^{-1}: g(B) \backslash g \circ f(A)$ then $g^{-1}: g(b)$ We can use the $g^{-1}$ map to retrieve the elements of B that can't be gathered in A
$g(B \backslash f(A)) =g(B) - g \circ f(A)$ Similarly, we can use $f$ to map from $A \backslash g(B)$ which produces $f(A \backslash g(B) ) = f(A) - f \circ g(B)$. the elements in A can't be retrieved by applyi
if I use the complement definitions of the $g^{-1}$ and on $f$ ... I could see it as my $g^{-1} : g(B) \rightarrow B$
$f: A \rightarrow f(A)$
what does $f:A$ and $g:B$ mean?
I applied the complement defintion x.x
but I have no idea what you mean by $f:A$ or $g:B$... it could mean that the domain of $f$ is $A$ and the domain of $g$ is $B$ but we are given that.
oh
so I shouldn't mess with $f: A \backslash g(B)$ and $g: B \backslash f(A)$
10:28
Well... $f:A-g(B)\mapsto f(A)-f\circ g(B)$ is a restriction of $f$ to the set $A-g(B)$ and the codomain is $f(A)-f\circ g(B)$
@usukidoll we make this restriction so that it is surjective
which is for a function $f : X \rightarrow Y$ with the property $(\forall y \in Y)(\exists x \in X)[f(x)=y]$
only we have $f: A \backslash g(B) \rightarrow f(A) \backslash f \circ g(B)$
@usukidoll only if $f$ is surjective
ack!
yeah
so oh wow what a mess to put in the definition
10:33
$f:A\mapsto B$ does not mean that $f(A)=B$, it means that $f(A)\subset B$
the range, $f(A)$, is s subset of the codomain, $B$, and they are equal when $f$ is surjective
this way $f:A\mapsto B$ leaves $B-f(A)$ possibly non-empty
and $g:B\mapsto A$ leaves $A-g(B)$ possibly non-empty
f(A) is contained in B ...
well yeah it won't be empty at all .. the sequence will be going on forever since it won't terminate
@usukidoll it might not terminate...
26 mins ago, by robjohn
$g^{-1}:g(B)-g\circ f(A)\mapsto B-f(A)$ and
$f:A-g(B)\mapsto f(A)-f\circ g(B)$ are bijections.
that is the first step...
D:
next step?? is??
@usukidoll well, in the cited statement above, we define a bijection from $A-g\circ f(A)$ to $B-f\circ g(B)$
10:46
since $A-g(B)\cup g(B)-g\circ f(A)=A-g\circ f(A)$ is a disjoint union
and then what happens next? surjective injective map and then it's proven that it's a bijection if we have those maps... yup the theorem does state that A and B are disjoint sets
@usukidoll then we work on making a bijection from $g\circ f(A)$ to $f\circ g(B)$ in the very same way
we peel layers off of the domain and codomain
so ok that means that we need a surjection map and an injection map for $g \circ f(A)$ to $f \circ g(B)$
11:03
maybe I should just subsitute letters.. since it's the same ... just with different letters... or maybe the order is switched or something
11:19
@robjohn Heya, do you remember answering a question about an integral that became $$ \frac{\pi }{2 \cos \alpha} $$ ? It was fairly recent (saw the question a few days ago tops)
11:39
@robjohn @DanielFischer In how many ways can three distinct numbers be chosen from the set $\{1,2,\ldots,2n\}$ such that the numbers are in increasing arithmetic progression?
I found it, thanks
@robjohn @DanielFischer good morning.
nerghhh
so much variables involved
unless $ g \circ f(A) \backslash g(B) $ for $g^{-1}$
$f \circ g (B) \backslash f(a)$ for $f$
@robjohn it was more of a guess based upon Euler-Maclaurin formula. It means the error term is moving within some bounds.
@robjohn I'll check that later on.
r9m
r9m
@Hawk fix the middle term, and count how many 3 terms APs can be formed ?
(not sure though that will work)
11:49
@r9m I am not fixing any middle term, I am choosing any 2 terms from the set of 2n terms, definitely I will get an AP.
But, the question is will it be unique?
@robjohn we may consider the term $$\frac{\sin(\log(n)+\gamma+ \mathcal{O(n^{-1})})}{n\log(n)}$$
@robjohn after manipulating the sine part, we can use Euler-Maclaurin formula ... I'll went into details when I'm back home
@r9m Do you find this strategy to be reasonable?
r9m
r9m
@Hawk fixing 2 terms (2n-1, 2n) and (2n-2,2n-1) results in the same AP (so repetitions occur in this way)
@r9m Is there any way for selective choosing?
r9m
r9m
i was trying to think if we fix say $k$ as the middle term of the AP, with common differenced = $r$, then the AP is $\{k-r,k,k+r\}$ .. so now all we have to do is vary $k$ and see how many $r$'s give an admissible AP
if $k \leq n$ ... and $k > n$ are the two cases to be considered
11:56
@r9m Suppose the term selected is the last term...what then?
r9m
r9m
0
@r9m ?
r9m
r9m
@Hawk for $1 < k \le n-1$ we have $k-1$ choices of $r$ to form 3 term APs and for $2n > k \ge n+1$ we have $2n-k$ choices for $r$ to form 3 term APs ..
why are we omitting the $n$th term?
r9m
r9m
oh write the n-th term case as well .. sorry missed that
12:06
@robjohn these integrals are convergent $$\int_2^{\infty} \frac{\sin(\log(x))}{x \log(x)}\cos(\gamma+\mathcal{O}(N^{-1}) \ dx \text{ and } \int_2^{\infty} \frac{\cos(\log(x))}{x \log(x)}\sin(\gamma+\mathcal{O}(N^{-1}) \ dx$$
Greetings @Chris'ssis :-)
@skullpatrol Greetings! How are you doing? :-)
@robjohn The last part is the reminder ....
(I need to leave now)
@Chris'ssis Fine thanks, how are you?
r9m
r9m
@Sush the term $\dfrac{\sum_{cyc} x(1+y)}{\sqrt{xyz}} \ge \dfrac{3\sqrt[3]{xyz(1+x)(1+y)(1+z)}}{\sqrt{xyz}} = \dfrac{3\sqrt[3]{(1+x)(1+y)(1+z)}}{\sqrt[6]{xyz}} \ge \dfrac{3\sqrt[3]{(2\sqrt x)(2 \sqrt y)(2 \sqrt z)}}{\sqrt[6]{xyz}} = 6$ ..
@skullpatrol Preparing to visit my grandfather. I was invited to eat some cookies.
12:09
yummy :-)
have fun
@skullpatrol Yeah :-)
Thanks
@robjohn because each one can be written as $$(\cos(\gamma)-\epsilon_N) \left(\int_N^{\infty} \frac{\sin(\log(x))}{x \log(x)}\right) \le\int_N^{\infty} \frac{\sin(\log(x))}{x \log(x)}\cos(\gamma+\mathcal{0}(N^{-1}) \ dx \le (\cos(\gamma)+\epsilon_N) \left(\int_N^{\infty} \frac{\sin(\log(x))}{x \log(x)}\right) $$
@robjohn and those integrals at extremities are straightforward.
$\text{SinIntegral}(\log(x))$
@r9m $\int\limits_0^{n+1} f(x)dx$ where $f(x)=\min\{|x-1|,|x-2|,\ldots,|x-n|\}$. I got $\dfrac{n+2}{4}$ but the answer seems different.
12:24
@Hawk Good afternoon. For the difference, you have $1 \leqslant d < n$, and the first term in the progression must be $\leqslant 2n-2d$, so altogether $$\sum_{d=1}^{n-1}(2n-2d) = 2n(n-1) - 2\sum_{d=1}^{n-1}d = 2n(n-1) - n(n-1) = n(n-1)$$ possibilities.
@DanielFischer Oh yes, that is really good approach. Thank you.
@DanielFischer $\int\limits_0^{n+1} f(x)dx$ where $f(x)=\min\{|x-1|,|x-2|,\ldots,|x-n|\}$. I got $\dfrac{n+2}{4}$.
Am I doing something wrong?
r9m
r9m
@Hawk the answer is n/4 ?
@r9m No, but how did you get that?
@r9m the answer is $\dfrac{n+3}{4}$
@Hawk On $[0,1]$, $f(x) = 1-x$, on $[n,n+1]$, it's $x-n$, so those two intervals contribute $\frac{1}{2}$ each, and the remaining $n-1$ intervals have a triangle of height $\frac{1}{2}$, so contribute $\frac{1}{4}$ each. That makes $$\int_0^{n+1} f(x)\,dx = 2\cdot \frac{1}{2} + (n-1)\cdot \frac{1}{4} = \frac{n+3}{4}.$$
r9m
r9m
@Hawk .. but I counted the triangles with base 1 and altitude $\frac12$ .. and there are two triangles in the extremities with base $\frac12$ and altitude $\frac12$ so put them together to form a single triangle with base 1 ..
12:33
@DanielFischer Oh, yes, I missed that part..thank you for the solution!
@Hawk A quick question for you, prove:
$$\displaystyle{\int_0^1 \frac{1+x^{30}}{1+x^{60}} = 1 + \frac{c}{31}}, \qquad \text{where } 0 < c < 1.$$
@ParthKohli Hi.
@Sawarnik Tried that problem?
I'm gonna give another one.
@ParthKohli Couldn't prove it :(
@Sawarnik It's got a beautiful solution.
@ParthKohli Please tell!
12:38
I'll state the question again:

Let $a,b,c \in \mathbb{R}$ such that $c(a+b+c)<0$. Then the quadratic $ax^2 + bx + c=0$ will always have:
1. Exactly one negative root.
2. Two complex roots.
3. Two real roots.
4. Two positive roots.
Let me give you some more time to solve it...
@ParthKohli the possible answers are option 1 and option 3
@Hawk Did you try my question?
@Sawarnik trying...but no progress yet
Ok :)
12:41
@Hawk Yeah, and 3 will always happen.
@ParthKohli I think I can prove it, just wait.
There's going to be a really unexpected theorem involved.
r9m
r9m
@Sawarnik $1 < \dfrac{1+x^{30}}{1+x^{60}} < 1+x^{30}$ for $x \in (0,1)$
@r9m Good!
Oh, really, I didn't see that coming!
@DanielFischer Let $n$ be a positive integer. Consider a square $S$ of side $2n$ units. Divide $S$ into $4n^2$ unit squares by drawing $2n-1$ horizontal and $2n-1$ vertical lines one unit apart. A circle of diameter $2n-1$ is drawn with its centre at the intersection of the two diagonals of the square $S$. How many of these unit squares contain a portion of the circumference of the circle?
12:47
@ParthKohli I think I can prove it when c is positive only :(
Ok, give a hint.
Here's the solution:

$c(a+b+c)<0 \Leftrightarrow f(0) \cdot f(1) < 0 \Leftrightarrow$ $f(0)$ and $f(1)$ have an opposite sign

$\Leftrightarrow$ a real root lies between $f(0)$ and $f(1)$ by Intermediate Value Theorem
Awesome solution!
So, for trivial reasons, both the roots are real - unique or not.
Wow!
Wonderful idea, and this poor guy is wandering with properties of quadratics.
Yeah, the solution stunned me.
12:51
I am still stunned! Ok, give the other one :)
Whenever you see expressions like $4a + 2b + c$ or $a + b + c$ or $c$, always keep in mind that we're talking about the values of the functions.
Yup, I have noted it in mind!
@ParthKohli The other q?
@Hawk $\int_0^{\pi/3}\frac{1}{1-\sin x} dx=?$
r9m
r9m
[rinpoche](http://www.youtube.com/watch?v=wKvlDGF1vEE)
@robjohn :) ;) .. awesome solution to the $\sin H_k$ series ..
This one doesn't have that much of a beautiful solution, but anyway...

Let $a,b,c$ be distinct real numbers. All of the following expressions are positive for real $x$:

$ax^2 + bx + c$
$bx^2 + cx + a$
$cx^2 + ax + b$

Let $\alpha = \dfrac{ab + bc + ca}{a^2 + b^2 + c^2}$. Which of the following will apply?

A. $\alpha > 1$
B. $\alpha > 1/4$
C. $\alpha < 1$
D. $\alpha < 4$

« first day (1348 days earlier)      last day (3971 days later) »