« first day (2423 days earlier)      last day (2604 days later) » 

12:01 AM
This is one of those things I consider an artifact of our language's evolution, and which I think we could comfortably discard.
 
(I'm back) No, it's "who." You always go by the dependent clause @MeowMix
 
@BalarkaSen are you here ?
@AkivaWeinberger do you want to discuss something related to fibrations ?
 
I don't know much about them
 
@Fargle Changing pronouns just because it's object vs. subject is kind of absurd. Too bad people don't pick up conlangs :(
 
@MeowMix In other languages there are many, many more cases to consider than just nominative and accusative. German has 3(.5), Latin has 8. I like our system, I think it lends itself to somewhat unambiguous speech without a lot of overhead, but I dunno.
 
12:05 AM
Conjugation is a weird thing too.
Is there a point to it?
In english it's pretty useless because you have to use a pronoun anyway, although in some languages like Spanish, you can omit the pronoun
 
Here's an interesting feature of language: the difference between "and" and "but." Logically, they're the same (they're both $\land$) and yet we have two different words for it
I wonder if any language doesn't differentiate between the two
 
I mean the connotations differ
And that difference is a pretty useful one to have
 
@AkivaWeinberger I'm not sure. I can't think of any examples that don't differentiate. But it is very useful to have that difference.
Jinx.
 
internal "drat"
 
@Fargle They're (almost?) never interchangeable, but they still both mean $\land$ from a purely logical perspective, right?
 
12:08 AM
Yeah
"Funny but scary" = "Funny and scary" but the latter sounds weird...
 
@AkivaWeinberger Yes. "This but that" = "This, and (contrarily) also that"
 
Hah my name has been called 3 times! I'm FREE
 
I once heard something about the notion of "common ground"
It's the set of things I know and you know, and I know you know, and you know I know, and I know you know I know, and… etc
So the idea is that the reason we have conversations, the reason we use language, is to expand our common ground.
And we need all of our new sentences to fit into that common ground, or else it doesn't work.
 
@AkivaWeinberger There's a logical puzzle around this concept.
 
@Akiva I bet you don't know that I know what you know I know.
 
12:16 AM
If I said, like, "The government's evil mind-reading cameras know your every move," it would sound strange if we didn't already establish that such cameras exist
So, maybe the usefulness of "and" and "but" is that they let us know more easily how to fit everything into our common ground.
And, since the notion of a common ground is a linguistic universal, it would make sense if every language had an "and/but" difference.
 
If not a specific word, then a modifier for "and" at least.
 
Tomorrow I'm going to actually eat breakfast
Because there are pancakes :]
 
W00t
makes a justin->just in pun
 
12:33 AM
@AkivaWeinberger This is the main obstacle with reliable communications over internet. The packet could get lost so you have to acknowledge it, but the acknowledgement could get lost... lol
 
I bought this awesome desk lamp lumens.com/…
now I do math in style
 
1:14 AM
Does $(AB)\vec{x} = A(B\vec{x})$?
 
@MickLH "Maybe 1000 acknowledgements we should trust each other" "But what if I send you the 1000th acknowledgement and don't know if you received it?"
@MeowMix Yeah
 
OK, then this problem is easy as shit.
I really hope the supervisor lets me skip algebra 2 and pre-calc at least :(
 
A vector is just an $n\times1$ matrix, and matrix multiplication is associative
 
good point
Oh shit, this isn't guaranteed to be invertible, so I'm wrong.
Hi @Aaron
 
Hi backatcha!
 
1:16 AM
i have beautiful topology example that will change the world
 
and, what would that be?
 
do you know about topology?
 
A lil' bit
 
@MeowMix What was the problem, may I ask
 
well, think about $[0,1]$
 
1:18 AM
don't spoil it for me, but :P
 
thinks hard about $[0,1]$
 
is there a proper closed connected subset that contains $0$ and $1$?
 
No lol
 
(or a proper connected subset at all?)
ok
 
1:19 AM
Connected subsets of the real line are intervals
 
^
 
now what if a connected space had this property with every two of its points
 
(Fun fact: this needs proof @MeowMix)
 
such a space is a mess
 
@ForeverMozart O.O
 
1:19 AM
but exists
 
Suppose $A$ is $m \times n$ matrix and $\mathbf{x} \in \Bbb R^n$ satisfying $(A^TA)\mathbf{x} = 0$. Prove that $A\mathbf{x} = 0$
 
I take it this is not Hausdorff @ForeverMozart
 
(Don't spoil!)
 
no, there is one in the plane. I constructed some but that is not my main contribution
 
@MeowMix There's a hint in the textbook, isn't there?
 
1:20 AM
Yes.
 
@ForeverMozart O.O
 
It says find $||A\mathbf{x}||$ But I'm not sure what to do.
 
I constructed one in $\mathbb R ^3$ so that every compactification has the OPPOSITE property between every two points in the original space
 
@ForeverMozart Not path-connected, though, I take it
 
correct
 
1:22 AM
@ForeverMozart What do you mean by the opposite property
 
it cannot be path connected
for any two points in the original space there is a proper closed connected subset of the compactification containing them
that is the opposite
so it is a strange space
okay maybe not change the world, but interesting to topologists anyway
 
@Akiva how am I supposed to find that>
 
@AkivaWeinberger This is pretty much the usual solution. The "what if" is answered by "well then we're screwed" and we just hope that the $p^{1000}$ probability is small enough that it never happens lol. (concretely, we usually have $p \approx 0.01$ so it's not that crazy of a leap of faith)
 
is that interesting to anyone :(
I spent a full week constructing the example
 
@ForeverMozart I don't know what it means, how do I apply it to algorithms? :P
 
1:32 AM
it's too infinite I'm afraid
 
Don't quit that easily! Fear is part of digging through unknowns!
And there's always a way to apply it concretely
What can I represent with these sets, and once I do that then what unexpected thing does your idea show the existence of?
 
well I'll try. I may be able to prove a duality theorem between these spaces and locally connected cut-point spaces (like [0,1]).
they are kindof opposite
oops I meant (0,1)
every point is a cut point
 
@MeowMix Hint: $x\cdot x=$
 
$||x||^2$
?
 
@AkivaWeinberger do you understand?
 
1:39 AM
Hey, how do you get two atoms with 6 protons to go on a date?
With CARBON DATING.
 
oh no
 
@MeowMix A mind is a terrible thing-
 
"What a shame it is to lose one's mind"
@MeowMix How does that apply to the hint?
@ForeverMozart Yeah
 
it is very interesting to see which properties of a space can be reflected in a compactification... to me anyway
that's my speciality
many results in this area and I have a new one
 
Well $A\mathbf{x} \cdot A\mathbf{x} = ||A\mathbf{x}||^2$
 
1:51 AM
yes
 
@MeowMix is this equal to $\mathbf{A}^\text{T} (\mathbf{A}\mathbf{x})$
 
Yes I noticed that earlier
 
@MeowMix There's a formula in that chapter that involves dot products and matrices
 
OH
Ted's favorite?
 
@MeowMix Can you invert $\mathbf{A}^\text{T}$?
I'm just wondering if $\mathbf{A} \cdot \mathbf{x} = \mathbf{A}^{-\text{T}} \cdot 0$ is broken
 
1:56 AM
@MickLH $A$ might be singular
@MeowMix Yup!
 
I wish I was smart enough to realize that lol
@Akiva I'm worried (and she probably will) that the supervisor will say no
 
@MeowMix You clearly hadn't heard Ted talk about his favorite formula enough
 
What's the formula?
 
@Semiclassical $Ax\cdot y=x\cdot A^\top y$
 
2:12 AM
If I have $\mathbf{v}$, $\mathbf{Av}$, and $\mathbf{Bv}$, can I compute $\mathbf{ABv}$?
I don't really know how to even attack this problem...
even assuming the matrices commute
 
by the way @AkivaWeinberger, the answer to math.stackexchange.com/questions/2158686/… is yes but I'm trying to reconstruct the proof
 
2:35 AM
Quick, kind of broad question in logic
What's the closest to a constructive Lowenheim Skolem theorem
 
3:19 AM
Let $P_n(\Bbb{C})$ be the vector space of polynomials up to degree $n$. Then:
Claim: A number $x$ is transcendental if $\forall p \in P_n(\Bbb{C}), x\not\in \textrm{ker} (p)$
 
considering $p$ is not a linear map, you wouldn't write $\ker p$
rather write $p^{-1}(0)$
and no, that's false
not relevant
 
O wait, nvm. There are algebraic numbers that lies in $p^{-1}(0)$
 
for example, $\forall p\in P_{\color{Red}2}(\Bbb C)\setminus \{0\},\sqrt[\color{Red}3]{5}\not\in p^{-1}(0)$, but $\sqrt[3]{5}$ is not transcendental
 
yes indeed
Hmm... a transcendental number is defined to be a number that is not a root of any nonzero polynomial with rational coefficients. Let me check the definition of a root again...
hmm... a root of a polynomial is a zero of a polynomial, that is some y such that p(y)=0
hmmm... how about:
Claim: A number $x$ is transcendental if $\forall n \in \Bbb{N}\& \forall p \in P_n(\Bbb{C})/ \{0\}, x\not\in p^{-1}(0)$
 
write $P_n(\Bbb C)\setminus \{0\}$ instead and you're good
or just let $\Bbb C[T]$ be the space of all polynomials and say $\forall p\in\Bbb C[T]\setminus\{0\}, x\not\in p^{-1}(0)$
no need to split that into a separate statement for each degree
 
3:31 AM
I see, that one is indeed more concise
Now, checking the cardinality of $\Bbb C[T]$...

Given a $p \in \Bbb C[T]\setminus\{0\}$, the coefficients of $p$ is given by surjective maps $m: \Bbb{N}\to \Bbb{Q}$. This means $|\Bbb C[T] \setminus \{0\}|=|\Bbb{N}^{\Bbb{Q}}|=\aleph_0^{\aleph_0}=2^{\aleph_0}=\mathfrak{c}$
 
$=\aleph_1$
confident smirk
 
Ah yes, CH only said whether $\mathfrak{c}=\aleph_1$
But $2^{\aleph_{\alpha}}=\aleph_{\alpha+1}$ is always defined by cantors theorem
 
Lol I'm just gonna roll with GCH in fact
 
Hmm, so this space is as large as the reals... Now by definition of algebraic numbers, the cardinality of the set of all $p^{-1}(0)$ is countable
need to think about how that translates to the cardinality of the subset of $\Bbb{C}[T]$ that satisfy $p(x)=0$ for all $x \in \Bbb{C}$...
[To be proved] Let $\Bbb{A}$ be the set of algebraic numbers. Then if $\exists p \in \Bbb{C}[T], p(\Bbb{A})=0$ then $p=0$
Proof:
Pick a $n \in \Bbb{N}$ and $a_n \in \Bbb{Q}$. Then
$$p(x)=\sum_{i=0}^n a_nx^n$$

Now by assumption $\forall k \in \Bbb{A}$

$$p(k)=\sum_{i=0}^n a_n k^n=0$$

The zero polynomial can be written as $0(x)=\sum_{i=0}^n 0 x^n$. Therefore

$$p(k)=\sum_{i=0}^n a_n k^n=\sum_{i=0}^n 0 x^n=\sum_{i=0}^n 0 k^n$$

Since $x^i$ and $x^j$ are linearly independent for $i \neq j$, we can compare coefficients and get

$a_n=0$

Now since $n$ is arbitrary, the above applies to all $n\ in \Bbb{N}$ hence $\Bbb{C}[T]$. Therefore $p=0$ as required
Actually I think I am missing something sufficient... I need to think about this more
Suppose there is $q$ such that $q(\Bbb{A})=0$. We need to show that $p=q$ otherwise this will not work
Digging MSE to see whether there are other sources that said anything about whether real numbers can be nilpotent
Nope, if there is at least one nonzero nilpotent element in the reals, then the reals will ceased to be a field (since nilpotents have no multiplicative inverses)
Ok so that means $e^e \neq 0$ which means we can still proceed along this line of thought laid out above
 
4:10 AM
$\Bbb A$ is infinite and a polynomial can only have finitely many zeros
 
but the combination of all $p\in \Bbb{C}[T] \setminus \{0\}$ with $\Bbb{A}$ has to be at least countable right?
 
21 mins ago, by Secret
[To be proved] Let $\Bbb{A}$ be the set of algebraic numbers. Then if $\exists p \in \Bbb{C}[T], p(\Bbb{A})=0$ then $p=0$
you're only talking about one hypothetical $p$ in that statement
 
ah yes. In that case p must be the zero polynomial as only that can have infinite many zeros
 
hi guys
random question
 
4:13 AM
0.0
 
@TheGreatDuck shoot
 
What do you think is an optimal number of questions/answers one should have on their account in order to have a nice variety whilst not appearing like they abuse the site?
after all, there is a point at which too many questions just wastes kilobytes on the server...
 
That will need to be a lot, otherwise SE will go down in a few days from memory overload
 
@TheGreatDuck Not applicable
 
@MickLH are you phd?
 
4:16 AM
The worth of the data outweighs the cost of data storage
2
@ForeverMozart I cut out of uni way before that
 
but you like to research?
 
well yeah but you also don't want to appear bad to other users, right? too many questions and you appear dependent on the site...
 
Depends on the kinds of questions, really.
 
@ForeverMozart Yeah I love it, I work on some type of research every day
 
@Semiclassical no i just mean the total number of questions. Any questions. Assume a variety.
what's the max?
100? 200? 500?
 
4:19 AM
And just having a lot of answers isn't necessarily a sign of a quality user. There are some who get a lot of rep just by doing a lot of crap answers.
 
no...
 
@Semiclassical some of my most trivial answers got the most votes :)
 
I think 100+ is pushing it, but then I'm not anywhere near that.
 
i said what is the maximum a user is allowed to have before being patronized.
and I meant how many questions should one keep on their account at a time.
 
I don't really care about my points anymore. I only think about a question if it looks interesting
 
4:20 AM
before cleaning them out for another batch.
 
Eh. I don't really care about the max number so much as the frequency of creation.
 
Quality
over quantity
 
im just thinking that one should have a select number of questions tailored to them as a user and get rid of questions after a while that are random
 
I mean, if you build up a reserve of questions without answers then that's probably a sign you're not posing great questions.
 
4:21 AM
@skillpatrol assume any quality. I literally mean that even if they are perfect question what is the number that stack exchange allows?
 
Ah, you mean is there a built-in system limit?
I'd be surprised if there was.
 
@Semiclassical of your questions are way too challenging... I posted some that are actually open problems just to see what people here could do... crickets
 
Well, sometimes they're challenging because they're not well-posed :)
 
@Semiclassical i mean more in the sense of community consensus.
 
If it's a matter of community consensus then I'd be even more surprised if there was a specific upper limit, especially one that's independent of quality.
 
4:23 AM
like... even if every question is answered and well received, at what point should one start deleting old questions to make room for new ones and to also make the user's profile more easily navigable?
 
Wait, delete answered questions?
 
isn't the purpose of asking questions also to give a decent appraisal of what the user is currently interested in/strong in?
 
No, you don't do that period.
 
Have you asked on meta? @TheGreatDuck
 
@Semiclassical or somehow remove them from your account.
 
4:25 AM
Deleting answered questions is a slap in the face of whoever put in the effort to answer the question. Moreover, it goes against one of the main purposes of this site: to collect good questions/answers.
 
@skillpatrol no. I just wanted a quick opinion. Thought it might make a good conversation starter. :)
@Semiclassical that's why I said delete or somehow remove them from the account of the poster. Isn't there a way to deassociate them?
like to "clean up" someone's account?
 
Well, you can community-wiki them.
 
yeah, what's the optimal number someone should have on their account at a time before making them community wiki?
 
Actually, no.
Community wiki is only for answers.
 
4:26 AM
i've heard mods can do that though
for questions
 
Guys there's definitely not a limit on how many posts you should make
 
Based on the proof that arctic and I have discussed (and amended afterwards). Given $S \subset \Bbb{A}$ and $S$ infinite, if $\exists p \in \Bbb{C}[T]$ then $p(\Bbb{A})=0$.

That means, for $S \subset \Bbb{A}$ finite, the number of $p \in \Bbb{C}[T]$ such that $p(S)=0$ is going to be the set of all elements $(\{p\},\{S\})$ that maps to zero.
 
They would trivially enforce it if it were even slightly an issue
 
@MickLH we're talking about community consensus
on how many should be seen on one's profile at a time
 
Maybe. But if it's not an option for regular users then the entire point is silly.
 
4:27 AM
like... to narrow the scope of one's profile.
 
@TheGreatDuck From stack exchange's point of view, it's "the more the better"
 
If those questions haven't received useful answers, then removing them is fine.
 
Our thoughts fuel their search engine ratings, answered or not
 
If they have, though, then you're doing a disservice to the community by removing them.
 
@Semiclassical assume the questions are perfect with perfect answers. I'm only speaking in terms of let's say "hiding all of the random set theory questions someone posted because they want to be known for real analysis"
and I only mean removing them from your account
i am not at all proposing the deleting of old answers.
 
4:30 AM
$\{S\}$ is easy, $|\{S\}|=|2^{\Bbb{A}}|=\aleph_1$
$\{p\}$ however is not so easy
 
There's a bit of discussion re: community wiki here, including some statements on it for questions: stackoverflow.blog/2011/08/19/the-future-of-community-wiki
None of which, I should note, takes the questions associated with a given account as a priority.
If someone wants to be known for a certain subject of questions/answers, then what they should do is keep answering those kinds of questions/answers so that that's the top tag on their page.
As @MickLH notes, the SE pov is that more is better than less.
If there are issues of quality, that's one thing. But if the questions that have been asked previously are good, then if you want to get associated with a different subject you'd better ask more questions in that area.
The idea of there being a stigma associated with having too many quality questions is, to my mind, absurd.
 
oh
i always assumed the purpose of the user profile is to cherrypick your strongest qualities in mathematics and show them off.
 
In addition, $p$ actually exists in the form of equivalence classes since the solution set of $p$ and $\lambda p$ are the same. Ok this is taking too long, I am going to go back to chemistry for a bit
 
If you want to highlight your skills, do it in your bio.
 
how do I make my research more popular?
 
4:37 AM
oh i dont care really
XD
it was just a random thought
i honestly find the idea just as weird as you
 
@ForeverMozart What is your research again?
 
i just thought that was the community's viewpoint
probably why I don't really ask questions anymore. I don't want to get thumped by a mod for clogging up the site with too much data.
 
strange connected properties in topology
it seems to be out of fashion
 
Topology as in point-set or as in algebraic?
 
non-algebraic
 
4:39 AM
Yeah, that's a tough sell.
 
why is it unpopular?
 
Because of the lack of algebra?
 
ok, then why is algebraic topology so hot?
 
ah wait a second, by the fundemental theorem of algebra, a degree n polynomial always have n roots. So that means the set of p that satisfy $p(S)=0$ is going to be all of $\Bbb{C}[T]$. Have to think about how to take equivalence classes here though, later
 
Because algebraic topology has a ton of applications, for one.
 
4:41 AM
#1^
 
I mean, that's not the entirety of it.
There are questions that people ask in algebraic topology that I don't have any sense of what a possible application would be.
 
I think as long as you work with separable metric spaces there will be applications
 
But, for instance, topology gets mentioned a lot in condensed matter physics precisely because the math is useful for describing physical systems.
(e.g. topological insulators, chern classes)
 
I just find it too limited
only so much to do with manifolds
 
4:45 AM
What's too limited?
 
@Semiclassical I'm considering asking a question but I do not know if it will be well received.
 
On MetaStackexchange?
 
I've been trying to think of whether or not there is an elementary function (or some integral) with two inputs and is commutative but is not a simple composition of addition and multiplication. Would that be a good question to ask or would that be a waste of everyone's time?
 
Could you be more specific please?
 
i'm sorry. What do you mean in particular needs to be "more specific"?
:p
 
4:53 AM
but how to make my research sexy
 
@ForeverMozart somehow throw in the multiplication of three variables a b and r. Pretend like you had no idea it would happen that way and you didn't see. Other than that... research shouldn't be "sexy". It should be beneficial to mankind and truly worthwhile.. ;)
 
it is interesting mathematically, but maybe not beneficial to mankind
 
weeel....
i dont know what to say then
is it research as in for a school project or research as in "literally a new thing in math"
if it's the latter, then it is beneficial to man, most likely.
you just don't know how yet.
 
it's a very unstable process
my current rate is 1 new discovery every 4 months or so
 
5:08 AM
@TheGreatDuck $max(a,b)$ and $min(a,b)$ are both commutative.
 
@Fargle hrmm....
is it also associative?
 
Hey guys!
 
@TheGreatDuck Both are, yes.
Look into tropical algebra/tropical geometry for more details. You can actually treat max (or min) like "addition" over the reals plus $-\infty$ (or $\infty$), and addition like "multiplication", and get a ring-like structure.
 
5:24 AM
Hey there!
 
How's it going @skill?
(For some reason I always read your name as "skull patrol", no idea why)
 
Not too bad. How are you pal?
My old user name was skullpatrol.
 
Oh maybe that's why
And I'm doing well, resting before the storm
 
Exams?
 
Nope, I've finished those, I'm in spring break now and next week classes are gonna start again
 
5:31 AM
hey guys
can anyone help me with this algorithm analysis
def nested(p)
   "assume p > 1"
   q = 1
   while q <= p:
       i = 1
       while i < q:
           print(i)
           i = i * 3
       q = q + 1
how would you find the number of iterations for the inner loop?
 
Write a recurrence relation
 
it appears to me that the loop runs forever
oh no its just that your answer will depend on $p$
 
i wrote a solution using summation
$\sum_{q=2}^p b= n(n + 1)$
I dont know why its not translating here
 
I got a different solution
 
what is yours? I feel like mine is wrong
 
5:47 AM
Well... isn't it homework?
 
no, its a past assignment
but the solutions don't come out for a week and we have a test coming up soon
 
I'm gonna give you the benefit of the doubt, you're only cheating yourself if you're lying!
I got $\mathcal{O}(\log p!)$
 
i'm winning bigly
 
hmm
why would you do p!?
how did you come up with that
 
I solved the inner loop and then solved the outer loop in terms of a sum of the inner loop solutions, then used some identities to get a closed form that happened to be that
 
5:52 AM
but it only concerns the inner loop
not the outer loop
 
In the inner loop $i$ grows like $3^n$ after $n$ iterations, so you should expect $n$ iterations where $3^n = q$
 
the number of iterations for only the nested loop
 
@Fargle google product integral. i presumed only commutative/associativity were necessary to develop a third integral based around a nontrivial operation. I think there is another property preventing it from existing at the infinitely dense level. Perhaps the idea of another operator that can "scale down" the accumulation rate is needed. I'll reflect on these thoughts.
 
@Theo if you take log base 3 on both sides, you'll obtain $n = \log_3(q)$ as a close guess of how many iterations given the value of $q$
If you truly only are concerned with the inner loop then that's your solution, but I'm skeptical that it's a miscommunication
I'd think they meant to ask about the total number of print commands issued for an evaluation at a given choice of $p$.
 
hmm okay
 
5:59 AM
The way I got from that to the factorial was also simple: The outer loop invokes the inner loop for each value from $1$ to $p$, so you can just evaluate the inner loop estimate at each of those values and total them up
 
thank you for taking the time, I think I need to re read some of my notes to get a better grasp
 

« first day (2423 days earlier)      last day (2604 days later) »