« first day (1673 days earlier)      last day (3351 days later) » 

12:01 PM
Hey guys
Do u all have any textbooks for Ordinary Differential Equations as well
 
@BalarkaSen okay according to the question f is a continuous function on the interval [0,1]
Let's define the function
g(x)=f(x)-x let's take g(x)=0 then f(x)=x and x=0
similarly if we take x=1 and g(1)=0 then
f(x)=x otherwise f(x) is smaller than or equal to 1 and similarly if g(0) is not equal to 0 then f(0)-0 is greater then or equal to 0.in that case by the indeterminate value theorem there will be a value c $$c \to [0,1]$$ such that g(c) =0 and f(c)=c and therefore c is the fixed point
 
12:33 PM
@Sayan I don't understand your logic. "let's take g(x) = 0" why should I take that in the first place? You're asked to prove that $g(x) = 0$ for some $x$!
Your idea is more or less right, but you need to write up your thoughts carefully.
By the way, great stuff!
 
12:59 PM
hi @sayan
anyone here on the chat ?
 
wazzup?
 
1:18 PM
@SayanChattopadhyay We have told you to finish calculus first. =)
@infinitesimal I will be seeing my doctor next next Wed.
 
What will you say @ABeautifulMind
 
@infinitesimal Just say what has been happening and then get some meds for it.
 
You don't need medication @ABeautifulMind. Try cognitive therapy.
Perhaps even group therapy
 
1:39 PM
@infinitesimal I will think about therapy later on. I need to figure out exactly what kind and who to see.
 
Therapy should be first on your list and medication second @ABeautifulMind
Drugs are bad.
 
@infinitesimal Not necessarily. The therapists will probably tell me things I already know, you see.
 
@ABeautifulMind you will not know unless you try.
 
@infinitesimal Yes, which is why I am taking the meds first to stabilise myself, so that I can think more clearly and then decide which therapist to see.
 
A qualified therapist can also prescribe meds more suited to your condition if you let them @ABeautifulMind
 
2:03 PM
Is the isomorphism in the third isomorphism theorem (for abelian groups, say) natural?
nah, probably not
 
I sort of agree (in general) that therapy is a better option. More time-consuming in the short term though.
 
Compared to 10 years without asking for help?
 
Not getting help does waste a lot of time in the long run.
@Exterior I think there is a natural isomorphism. At least I remember being given one.
 
@BoniTea I'm a bit confused though, since the naturality I'm hoping for would involve three objects: if $A/B\cong (A/C)/(B/C)$ is there indeed naturality in all 3? In other words, do the set theoretic manipulations we use on representatives of equivalence classes indeed "not depend" on the sets?
 
2:37 PM
@BalarkaSen With what machineries? Certainly you can prove it with degree stuff - doesn't hatcher do so?
(It's much more natural, as Ted did, to prove the theorem with differential topology. Vector fields are really smooth objects; there's not really much reason to think about continuous vector fields.)
 
I don't see Hatcher doing it.
 
2.28
 
Anyway, the fact that if there is any nonzero vector field on S^n then there is a point with vector pointing directly inwards and a point with vector pointing directly outward does the trick, doesn't it?
 
Hatcher hatches eggs.
 
immediately after he introduces degree :p
 
2:40 PM
Miller mills eggs.
 
I don't understand what you mean
 
@MikeMiller What's not to understand? If there is a nonzero vector field on S^n, then there is a vector that points diametrically opposite to the center and a vector that points diametrically inwards to the center.
 
Please write down an actual proof, including all the details. I can't understand how you're trying to prove this.
I guess you wan tto invoke Borsuk-Ulam or something, but I'd really like to see "a vector field is a map [...] and if it doesn't do [...] then [...]"
 
Proof of what I mentioned above?
No, I have no intention of using Borsuk-Ulam.
 
Please write down the details instead of saying tidbits like that :P
 
2:46 PM
OK. Claim : If there is a continuous nonzero vector field $v(x)$ on $D^n$, then there is a point on $\partial D^n = S^n$ where $v$ points diametrically inwards and another point where $v$ points diametrically outwards.
 
There are loads of continuous nonzero vector fields on $D^n$.
 
Sure, but my claim holds.
Anyway, assuming this, there is no nonzero tangent vector field on S^n as each S^n can be realized as boundary of some disk and extended to a continuous vector field on D^n, right?
 
I can't see how this would provide a proof of the sphere case. You would have to be able to extend a vector field on the sphere to a nonzero one on $D^n$. How do you intwnd to do that?
 
Ack. That looks slightly nontrivial.
 
You should be able to fix your original idea - don't involve $D^n$ - but it won't work quite perfectly, because $S^n$ does have a nonvanishing vecto field for some $n$ :P
 
2:53 PM
@BalarkaSen So we finally meet.
 
@Parth Hi.
 
@BalarkaSen I'd be better with someone of my own league. So where has Sawarnik been?
 
Dunno, haven't talked to him in a while.
 
you have never met his highness?
 
@infinitesimal skullpatrol?
 
3:00 PM
yep
 
@MikeMiller So you mean to say that any nonzero vector field on S^n \subset R^{n+1} will have a vector that points directly outwards/inwards?
Well, of course that's true.
 
And why so...? I was telling you to stop saying of course and tell me why.
 
Yeah, well, let your vector field be $v : S^n \to \Bbb R^{n+1} - {0}$.
Compose with $\Bbb R^{n+1} - 0 \to \Bbb S^n$ given by $x \mapsto x/||x||$.
This gives a map $f : S^n \to S^n$.
Assume there is no vector that points directly outwards.
This means $v(x) \neq ax$ for any $x$.
But that means $f : S^n \to S^n$ has no fixed points.
But $f$ is homotopic to the constant map via $H(x, t) = f(tx)/||f(tx)||$, which has degree zero, and any degree zero map must have a fixed point.
Contradiction.
 
What is $f(tx)$? We're on a sphere.
 
Nonzero -- map onto the sphere.
 
3:07 PM
No, what is $f(tx)$? It's a map from the sphere. You can't scale.
 
Wait, wait, the homotopy should be $H(x, t) = f(tx)$.
I dunno what you mean by "what is".
 
You have written down something not defined.
 
$f$ is the composition of $v$ with $\Bbb R^{n+1} \setminus 0 \to S^n$. It's defined just above.
 
If x=(0,0,1), what is H(x,1/2)? rhe domain of f is the sphere.
 
It's f(0, 0, 1/2).
Ohh wait
Duh
 
3:11 PM
the domain of f is the sphere
 
I -- get -- it.
:P
That was stupid
$f(tx/||tx||)$, then, I guess?
 
Your proof can be fixed. Everything up until you decided the map was null-homotopic was fine. You need a constraint on $n$ (and I'm not sure you know what a vector field is). Ping when when you have a correct proof.
 
That doesn't work, no.
@MikeMiller Well, a vector field on a subspace of R^n is just a continuous function from the subspace to R^n.
I have read this definition in Munkres, at least.
 
@Balarka: You want a vector field on a surface to be tangent to the surface, typically, although this isn't what one requires in flux integral calculations in multivariable calculus.
 
Never heard of such a condition on the definitions.
 
3:26 PM
Well, you will :)
I haven't been following the context.
 
there was a question that i came up with last night here on chat which I didn't have the proper terminology to express, and which i suspect has a well-defined formulation
namely, how does take the old problem "for what point on earth does walking a distance $d$ south, then $d$ east, and then $d$ north return you to your starting point"
and state it in a way that can be generalized to a generic orientable surface?
 
Yeah, this is related to the hairy ball theorem. Your usual notion of NWSE is going to break down at the poles.
 
sure, there's a coordinate singularity
 
No, this is not a coordinate singularity.
It's because there can be no nowhere-vanishing vector field on the sphere.
Orientability isn't sufficient.
 
hmm. i was thinking of it as a singularity in the sense of longitude-latitude being a curvilinear coordinate system
 
3:31 PM
Nope, this is intrinsic to the topology of the sphere. It's not a fault of parametrizations.
 
oh, sure. i wasn't saying it wasn't
just that, in the definition of NWSE provided by that coordinate system, it's a singular point
 
Yeah, you have two such points with the standard latitude/longitude system.
But even if you use stereographic projection, you'll have one problem point. Guaranteed that you can't avoid that. Note that on a torus you can make a global prescription.
 
right. but while that's certainly relevant, the North Pole isn't the only solution to the returning explorer problem
one can also start on a line of latitude such that walking south a distance $d$ lands you on a line of latitude whose circumference around the earth is $d$, so that walking east a distance $d$ returns you to your starting point
 
Good point :)
 
heh. i didn't notice it until i saw it pointed out
 
3:35 PM
Hello @ted. You seem like a nice teddy to hug.
 
LOL, Jasper.
 
and more generally any circle of circumferance $d/n$ will give rise to a set of solutions
 
Good day. About to disappear for an advising appointment and then class. Oh, btw, @Semiclassical, today we discuss work, kinetic energy, and the fundamental theorem of line integrals (which is exactly what you were talking about yesterday).
 
Morning, @Ted. Enjoy class.
 
3:37 PM
Good night, @Mike.
Tests were discouraging.
 
High scores 81, 82.
 
Yikes. What class is this?
 
Afternoon @ted @mike
 
My multivariable math class. Test on change of variables, beginning of differential forms, contraction mapping, inverse/implicit function theorems, manifolds. It's hard.
 
3:38 PM
morning Jasper
It sounds nice, @Ted
 
If they had been calm and taken their time, half the class should have been able to score 90 or so, but it rarely works that way.
 
@TedShifrin Anxiety is very hard to deal with.
 
it really is. i thankfully never had trouble with test anxiety, but writing? ughhhh
 
I have been through so much nontest anxiety that I will never have test anxiety, LOL.
 
Baaah I don't know how to do this.
 
3:40 PM
@TedShifrin: talk to you later, then. i'd love to see a generic statement, at least for those winding solutions.
though i guess it might just amount to: if i start with some point on a surface, and start walking 'east', do i ever return to my starting point?
 
4:25 PM
YOHOO guess what I found on the floor in the street :D
 
Hi @BalarkaSen
@BalarkaSen saw my answer ?
 
@TedShifrin send me the test :3
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ If he does that, you can sell the answers back to future classes ;-)
2
 
@robjohn How do you think I make money ? :P
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ Just print them.
 
4:38 PM
@ABeautifulMind As a metter of fact, I found a genuine 50€ bill on the road today :O
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ it's one step above those who collect and sell phone numbers, email addresses, etc
@ɧɿρρԹʅȝՇԵՐՎԾՌ I was wondering where I dropped that...
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ since you don't link your comments, I don't know to which comment your frown was in reference...
 
The last one :D
 
hi @ɧɿρρԹʅȝՇԵՐՎԾՌ @semi :D
 
4:40 PM
@Sawarnik Hello
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ better :-)
 
Have you seen Ramanewbie? @Hippa?
 
@Sawarnik Not since yesterday
 
Which language have you changed your name to? :D @Hippa
 
4:41 PM
@Sawarnik WAT ????
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ he's trying to access the current directory
 
@Sawarnik Where did you hear that ?
@Theorem Mind if I prove you ? :P
@Sawarnik Urm... that's wrong... in several ways
 
@robjohn Hi robjohn :)
@ɧɿρρԹʅȝՇԵՐՎԾՌ Surely if you want to :)
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ Uhh
 
@Theorem how are things?
 
4:45 PM
I see a Hindi and a Greek letter in it.
 
I see a Hindi and a Greek.
 
@Sawarnik they are just characters that look like the Roman characters they are supposed to be
 
@robjohn I know :D
 
@robjohn I have been working on something but i have no one to tell me if i am right or wrong .
 
@Theorem Is it posted on main?
 
4:46 PM
@robjohn How you been ?
 
@Theorem pretty good, for the most part
 
@robjohn yes i posted but nobody responded math.stackexchange.com/questions/1163470/…
@robjohn How is the weather ? Is it friendly ?
 
@Theorem I gave a +1.
@Hippa Are you on fb? :D
 
@Sawarnik Will it help me in any way to get some attention from mathematicians
 
don't know.
 
4:50 PM
@Sawarnik Nope
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ why is that?
 
I don't see why it would be useful for me :-)
 
@robjohn I would be grateful if you could look into it when u have time .
@robjohn And do let me know if the question even makes sense .
@ɧɿρρԹʅȝՇԵՐՎԾՌ did u vote as well ?
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ OK , thanks
 
5:02 PM
@SayanChattopadhyay I did.
It's okay, but you need to improve your proof-writing.
 
Yes a bit.....
 
@BalarkaSen Did you see the proofs book he is using? I never looked at it.
 
Yes, it's a good book.
I referred it to him.
 
What is the title and author again?
 
Hammack, Book of Proof.
 
5:04 PM
@BalarkaSen any book which contains the full set theory...I mean axiom of choice and all
 
No need to do axiom of choice now, @Sayan. Study naive set theory first.
 
@SayanChattopadhyay You should study axiomatic set theory much later on.
 
If you ever do algebra, you'll encounter Zorn's lemma anyway, which is axiom of choice in disguise.
 
@SayanChattopadhyay I know you are very eager to know many things, but trust me, one step at a time, we will guide you. Finish Hammock and Apostol first.
2
 
I think I have very less set theory to finish
 
5:06 PM
It's a lot, @Sayan. Are you doing all of the exercises?
 
@SayanChattopadhyay Balarka already says you need to improve your proof writing.
 
Yes I am all of them not even leaving a single question
 
Then it's lots, @Sayan. It'll take a few months to finish all of them.
How much have you done?
 
Till direct proof
 
So you did I.3.?
If you did, do this exercise :
 
5:09 PM
Let's see
 
You have n letters and n envelopes.
 
The letters have been labelled 1, 2, 3, ..., n each. The envelopes are also labelled 1, 2, 3, ..., n each.
 
How many ways can you put a letter in each envelope such that the $i$-th letter doesn't go inside the $i$-th envelope for any $1 \leq i \leq n$?
Again, don't google.
 
5:13 PM
I am not even sure how you can google that.
 
Right :P
There's a standard name for this number, though, which you can easily find by googling.
But I haven't mentioned it anywhere, so have fun!
 
Thanks for the q
 
No problem. Take your time.
 
Haha, I don't know the name. Forgotten all combinatorics.
 
Well, I got to go. Ping me when you find a proof, @Sayan.
 
5:16 PM
@ABeautifulMind what does the question mean by how many ways
 
@SayanChattopadhyay For example, putting letter 1 in envelope 1 and letter 2 in envelope 2 is different from letter 1 in envelope 2 and letter 2 in envelop 1. These are 2 different ways.
 
Oh thanks
So @ABeautifulMind the question is like the letter with number 1 should not go into the envelope 1 but can go to other envelopes
 
@SayanChattopadhyay Yes.
 
Great......I am nearing the answer let's see how it goes
 
@Hippa: Did you solve that problem with the contraction mapping principle?
 
5:24 PM
@ABeautifulMind can you do me a favour
 
@SayanChattopadhyay What is it?
 
Pls tell @BalarkaSen that I won't have internet for the next few days till 20th of march so I will give him the proof for the question by then
 
@SayanChattopadhyay That message you just typed will get to him.
 
Oh..but even in case.....
 
wow, to be free of the internet for 2 weeks ... such wondrous peace !
 
5:26 PM
@SayanChattopadhyay Don't worry about it, it's OK. If you worry so much you will go mad like me.
 
It is kind of I will concentrate full on biology......@TedShifrin
 
good idea, @Sayan. You need to do well overall in school.
 
@TedShifrin Not really, I'm not to sure how to get back to $x-f'(x)$ since $\phi'=1-f'(x)$
 
I got 10 A's for my O level and 4 A's for my A level exams @sayan.
 
I am loner @ABeautifulMind I live life in craziness
 
5:28 PM
@Hippa: You're misinterpreting linear maps. $I$ is the derivative of $h(x)=x$.
 
Oh I thought it was the identity. And what is $h$ ?
 
@SayanChattopadhyay Don't call yourself crazy until you become like me.
 
@SayanChattopadhyay It takes a lot of self-control to stay away from the internet for 2 weeks :D
 
it is the identity, @Hippa.
And the derivative of the identity map on $\Bbb R^n$ is the identity map.
 
Oh yeah q_q
@TedShifrin But why 1/3 then ? wouldn't any $q<1$ have worked ? (I get $|\phi(x)-\phi(y)|<|x-y|/3$)
 
5:30 PM
To get a fixed point, you need to make sure the map maps a closed set $X$ to itself
 
@TedShifrin Thanks for the italics, lol.
 
oh, stop complaining, Jasper :P
Do they still use Dieudonné's Éléments d'Analyse in France, @Hippa?
 
I only have the first 4 volumes. But I love those books.
They will go to California with me.
 
5:33 PM
Different versions have different volumes.
 
Well, I believe he wrote a few more after what I bought in the 80's.
 
My library has a six book set.
 
googles
 
It's not sold anywhere anymore I think.
The first volume is still sold I think.
 
@TedShifrin Ok so $|\phi(x\in B)-\phi(0)|\le |x|/3\le1$
 
5:34 PM
It's called Foundations of Modern Analysis.
 
@TedShifrin---I suppose the right questions to ask re: the returning explorer is: Suppose I start walking in a certain direction on a specified surface. 1) are there paths on this surface which eventually close? 2) for those that do close, what circumference of such paths are possible?
 
And $|y-a|\le2$
tadaam ?
 
You need |x| on your vectors, @Hippa!!
 
@BalarkaSen my obvious guess was $n!-\sum\limits_{i=1}^{n}\binom{n}{i}(n-i)!$ but then I noticed I'm counting some of the permutations a lot of times :(
 
for a sphere, all such paths eventually close, and the largest possible circumference is that of the equator
 
5:36 PM
@ted I think that Dieudonne's books are outdated like Bourbaki.
 
@TedShifrin and now ?
 
You think a lot of things, Jasper.
 
i imagine that the curvature of the surface is a key part of the story, but i dunno how to make any of that precise :/
 
Jasper, Dieudonné's books were always very terse, but have excellent exercises. People like Jack Lee and I come along and make stuff easier to read for modern students, but that doesn't make the older stuff bad.
@Hippa: I'd like to see a thorough exposition, but I believe you have it.
Yes, @Semiclassical, clearly curvature is highly relevant.
 
right.
 
5:38 PM
I think they sell Goursat's Course of Analysis as well, 3 volumes.
 
the case of a torus is interesting. if I walk so that my straight line corresponds to one of the loops of the torus, then it closes after one rotation
 
There was a Dover edition but I don't know if it is in print.
 
So many ways to approach an integral. Sigh.
 
And Godement has this 4 volume treatise on analysis, 2 of which have been translated.
 
more generally, i can pick the direction such that i wind around the two loops an integer number of times each (distinct!)
 
5:39 PM
@Semiclassical: If you take a point and a tangent vector at that point, there's always a great circle in that direction. But there are others, too.
 
@TedShifrin Professor i have small question on what i have been working on .
2
Q: Euler-Lagrange Equation and "Eigen Value "

TheoremThe Eigen value $\lambda(t)$ which is characterised by the Rayleigh quotient (where $t$ is a scalar variable): $$R(u,\Omega_t)= \frac{\int_{\Omega_t} |\nabla u|^2 dy }{\int_{\Omega_t} u^2 dy}$$ $\lambda(\Omega_t) = \mathbf{min} \{R(v, \Omega_t) : v \in H^{1,2}(\Omega_t)\}$ The minimiser $u*$ o...

@TedShifrin Would you mind looking into it when you have time ?
 
guys, i am interested enough in math that related in computer programming, if you guys know what theory or book should i learn may i know it? or some source of math that related to computer programs.
 
Amann has this 3 volume analysis work which has been translated.
 
right. i guess part of the issue is that one needs to select a choice of coordinates before one has a well-defined notion of direction
i.e. what does it mean to walk east
 
Jasper, I have 4 volumes, and I think there may be at least 9.
 
5:41 PM
@TedShifrin OK. I know some versions put a few volumes into one book, so the number of volumes/books is confusing.
 
but for the torus, one also has space-filling orbits when things aren't commensurate
 
@Theorem: I would have to spend an hour to read that. I don't have the time.
 
@TedShifrin Ok. Let $y\in B$ such that $|y|\le1$ and $\phi(x)=y+x-f(x)$. $\forall x,z\in B,|\phi(x)-\phi(z)|\le|x-z|\max D\phi$. $D\phi=1-f(x)$ hence $|\phi(x)-\phi(z)|\le|x-z|/3$. What is more, $\forall x\in B,|\phi(x)-\phi(0)|=|\phi(x)-(y-a)|\le|x|/3\le1$. Since $|y-a|\le2,|\phi(x)|\le3$. From the [whatever you call it] contraction theorem, $\phi$ has a unique fixed-point, hence there exists a uniqe $x$ such that $f(x)=y$.
 
i would like to post this question to the main site, i'm just not sure what kind of question one should precisely ask
 
@robjohn I just finalized this one with the elementary knowledge of the real analysis $$\int_{-\infty}^{\infty} \frac{e^{a x}}{1-e^x} \ dx=\frac{\pi}{\tan(a \pi)}, \space 0<a<1$$ Of course, it can be nicely done by contour integration too (actually, this might be the recommended way). I'll add it to my book too.
 
5:44 PM
@TedShifrin Shall i make an attempt to explain in short ?
 
@Hippa: For the mean value inequality, you need an operator norm on the $D\phi(c)$.
 
Oh yeah, right
 
@TedShifrin If $u^* \in H^{1,2}(\Omega)$ is a minimum of the $$\lambda(x, \Omega_t) = \frac{\int_{\Omega_t} {\mid \nabla u(x) \mid }^2 \, dx + \alpha \oint_{\partial \Omega_t} u^2\, ds} { \left( \int_{\Omega_t} u^q \, dx\right )^{2/q}} $$ , i am trying to find the necessary conditions (equations) that $u^*$ will solve .
 
@TedShifrin It's a nice method though !
 
That's the approach to prove the inverse function theorem, @Hippa, when you eventually get there.
 
5:49 PM
MIT accepts Chinese on top of French, German or Russian for foreign language requirement. I am not sure there is any other place that allows that.
 
hi mr @Pedro
 
By the way @ted do you know Chinese as well?
 
Ni Hao ;)
 
@Theorem: I imagine that sort of thing is well-known to the right sort of analysts/applied mathematicians who work on that stuff, but I'm not one.
Not a word, Jasper.
 
I thought you knew everything. =)
 
5:50 PM
No, I'm getting stupider by the day.
 
@TedShifrin Ok .
 
@Theorem O_o
gets meme ideas
 
@TedShifrin I hope the illness does not progress.
 
What Ted doesn't know is that it's not illness, I'm just slowly sapping all his knowledge to take as my own.
 
I corrupted his knowledge, so you are taking in all the wrong theorems.
 
5:52 PM
I thought you were referring to your illness, jasper.
 
I was thinking about your Alzheimer's.
 
LOL, I hope I don't have it for quite a while yet.
My mom has it, not I keeping fingers crossed
 
I will pray for you.
Prayer need not be religious you know.
 
LOL ... I have heart disease and have had cancer ... that's enough.
 
@TedShifrin Top : 'Everything you've learnt, you've forgotten ?' - Bottom : 'Alzheimer'
:/
 
5:54 PM
That's not funny, Hippa.
 
@ɧɿρρԹʅȝՇԵՐՎԾՌ smacks
 
not at all joke-worthy, @Hippa.
 
I smacked first, so I win.
 
There are millions of people and even more millions of their family members (including me) suffering from it.
 
Hi @JasperLoy @TedShifrin
 
5:55 PM
afk 10-20 minutes
 
However, it is true that I have forgotten a lot of mathematics that I knew years ago. It is also true that I still know a lot more than most people :P
 
Hi Bart.
 
hi mr eyeglasses
 
I still remember the quadratic formula, you know.
 
I had to use that in diff geo yesterday, Jasper, so that's good.
 
5:56 PM
I can even prove it with eyes closed.
 
So the courses for next semester are out, and a professor I really dislike is teaching a course I need to take
 
With eyes closed you may scribble on top of what you've already written.
Now you know how some of my students have felt, mr eyeglasses.
 
Two professors I really dislike got their full professorships early, they were very bad teachers.
The kind I dislike a lot is those that use other people's lecture notes without reading them up beforehand.
Or those that give lecture notes that are full of errors.
 

« first day (1673 days earlier)      last day (3351 days later) »