« first day (2870 days earlier)      last day (2447 days later) » 

00:24
why is it completely determined by s
and that makes it not a vector space?
do you have $u + v \in \{\phi_1\}$ if $u, v \in \{ \phi_1\}$?
or whatever thecorrect notation is
any two functions that you add together aren't in $ \phi $
so is it being closed under multiplication mean anything or is that useless
sure, there are lots of structures more primitive than vector spaces that it might be
maybe it's a semigroup
looks like one. check out this thing
maybe you can finagle it into a full-on group
groups are very rich
for a group, you need to have inverses, which necessarily means you have an identity also
inverse implies identity because by definition $xx^{-1} = x^{-1}x = e$ for an identity element $e$
01:09
@gf
@GFauxPas -1 is the inverse element of 1. So for every function I need an opposite one
 
1 hour later…
02:40
It is an abelian group, as $\Phi(s,x)\Phi(1/s,x)=1$.
 
1 hour later…
03:47
[Random]
Consider the following:
$A >_{n} B \equiv A > a > a > a > ... > B$
Let $\lim_{n\to \infty} >_n = >_{\infty}$
Then the following is true:
ooooops
uh... corrrection
$A_n >_n B \equiv A > a_1 > a_2 > ... > B$ where $a_i \in L$ and $L$ is linearly ordered
So n distinct elements in (B,A)
yeah
I am then wondering about some limiting case as follows, where the limit acts directly on the relation $>_n$ itself as follows:
$\lim_{n\to \infty} >_n = >_{\infty}$
and I want to demonstrate whether the following is a theorem:
$A + c >_{\infty} B + d$ if $d >_m c$ and $m$ is finite
So: If there's finitely many elements in (d,c) and infinitely many elements in (B,A) then are there infinitely many elements in (B+d,c+A)
03:56
well what I want to check is whether the ordering will not be changed if d and c are in a sense "infinitesimal" compared to A,B
except that I am trying to have a very minimal definition of infinitesimal by defining them using linear orderings
We knew that if in usual orderings if A > B and d > c then it is not in general true that A + c > B + d for example
I want to say that the following is true: If $\{a_i\}_{i=1}^n\in I_1$ and $\{b_i\}_{i=1}^m\in I_2$ then $\{a_i+b_j\}_{i,j=1}^{n,m}\in I_1+I_2$
Not sure that's good notation. Probably would have been better to use set builder
$I_1 + I_2$ is set union of the two sets?
no
adding endpoints
huh, interval arithmetic is a thing: en.wikipedia.org/wiki/Interval_arithmetic
04:02
I see
Hey guys
Hmm... if the set is linearly ordered, then since A and B are infinitely far apart, adding any "infinitesimal" to them will not change their ordering much (since you will need infinite many elements to be able to make A approach B for example, thus in general the endpoints of both sequences would preserve the order)
"The Boost collection of C++ libraries contains a template class for intervals. Its authors are aiming to have interval arithmetic in the standard C++ language."
This is absolutely horrifying
How bigger is the standard C++ going to become
I wonder what interval arithmetic is like when we start to sum cantor sets together...
in theory we can do something like this in a computer because while the cantor set is uncountable, there's a closed form algorithm to generate it at any precision, so we can get pretty close
04:28
Other things to be checked later: A game where the solution set is indeterminable but otherwise a winning strategy exists because of the symmetries in the solution set
I am starting to get interested in working with nondeterministic things and see if we can always exploit the symmetry of these things so that the outcome will be constrained regardless on what happens in the middle
 
2 hours later…
D_C
D_C
06:35
What does it mean to say that a meromorphic function has a pole at $\infty$?
In the above question can there be any other pole too except $\infty$?
@D_C we usually say that $f$ has some property at infinity if $f(\frac{1}{z})$ has this property at $0$
06:50
math experts: can I move partial x in this below to right side as shown in this screen shot?
07:23
your image is basically math cancer. The nonsense that they teach undergrads. But if you want proper notation, it just means $ \partial_x (\ln(F))=X(x)$ implying $ \ln(F)= \int X(x) dx$. God knows where y goes to.
Shouldn't that be $\partial_x (\ln (F)) = X(x,y) \implies \ln (F) = \int X(x,y) dx$
drats... can't edit my post anymore, but you know what I mean.
yeah otherwise someone has to justify why the $\partial_F (\ln F) \partial_y (F)$ bit of the chain rule disappears
08:26
Can objects in FdHilb be self-dual? :(
@JoBe Assuming that means the category of finite dimensional Hilbert spaces, then why not?
Because I'm not good at math and the Riesz rep theorem confuses me. My actual question is: Are objects selfdual?
Oh, and I also mean complex Hilbert spaces only.
What does it mean for an object to be self-dual? i know what it means for a category to self-dual
@TobiasKildetoft semidirect product is hard? :o
@MatheinBoulomenos an object there is a vector space
And actually, I am too unreflected to know if that is even my actual question, or if I want to ask something else.
An object is self-dual if it is isomorphic to its (say) left dual
08:33
@LeakyNun Yes, when first encountered, it takes some getting used to
@MatheinBoulomenos aha isn't the functor just $X \mapsto \operatorname{Hom}(X,\Bbb Z)$?
no nvm
@TobiasKildetoft :o
@MatheinBoulomenos is it $X \mapsto \operatorname{Hom}(X,\Bbb Q/\Bbb Z)$?
Semidirect products are easy, just take the extension corresponding to $0$ in $H^2(G,A)$
@LeakyNun what exactly are you asking?
@MatheinBoulomenos that only works if $A$ is abelian ^^
that functor shows that finite abelian groups are self-dual, yeah
@MatheinBoulomenos yay
that's interesting
Jun 5 at 21:43, by MatheinBoulomenos
more precisely, you can write down a functor $\mathbf{Ab}^{op} \to \mathbf{Ab}$ that sends colimits to limits and restricts to an equivalence between finite groups and their own opposite category
7.5 days!
and i only solved it because someone asked about dual
@TobiasKildetoft that's bs, it feels easy to me
I don't know how I felt when I first met it though
I have no memory of that
08:39
it feels easy to you $\ne$ it's easy to most people though
@LeakyNun you'll probably agree that this is easier than semidirect products: math.stackexchange.com/a/2653756/348926
@loch they are equal! :(
@MatheinBoulomenos deja vu: math.stackexchange.com/a/2818003/328173
@LeakyNun about your comment. This is a question about group theory. Cauchy's theorem is easier to prove than the fact that multiplicative groups of finite fields are cyclic
but it's less constructive
the proof of the existence of primitive roots is not constructive either
it's computionally expensive, that's why encryption by discrete logarithms works
08:47
I see
ok then
there are still some unsolved problems on primitie roots, e.g. is 2 a primitive root for infinitely primes? (Artin conjectured yes)
Well, Artin conjectured that every integer which is not a square and not -1 is a primitive root for infinitely primes
I remember reading that there are at most two primes numbers that are not a primitive root for infinitely many primes, but that we don't know one that is for sure
@MatheinBoulomenos is this related to global class field theory?
@LeakyNun I don't see how immediately, but the order of $n$ mod $p$ gives you the splitting behaviour of $p$ in $\Bbb Q(\zeta_n)$, so it's possible
I see
@MatheinBoulomenos is it true that $[\overline F : F] < \infty \implies [\overline F:F] \mid 2$?
08:55
yes
I've saw a relatively elementary proof of this once, maybe I can find it again
relatively proof?
is that the artin schreier theorem
@LeakyNun dropped a word, lol
@mercio yeah
note that the Artin-Schreier theorem and Artin-Schreier theory are not related
@LeakyNun with cebotarev's theorem you have natural densities for the set of primes where $n$ fails to be a primitive root for being a square, for being a cube, for being a 5th power etc... and when you take a finite number of those events, they are independant. The hard part is talking about infinitely many of those events at once.
I see
09:16
Finally completed this MathSciNet review.
@TobiasKildetoft how did you solve the issue with the two possible perspectives?
@MatheinBoulomenos By slightly changing the notation of the authors and with a small remark that the spaces under consideration were a natural thing to study from either perspective.
@TobiasKildetoft I see
The change was just to name the automorphism group $G$, rather than dragging around Aut(V) in a bunch of sibscripts
sounds like a useful notational change anyway
09:19
yeah, not sure why the authors kept that throughout.
@TobiasKildetoft would you mind glancing through what I wrote here so that I'm sure that I didn't write nonsense? (I'm not that familiar with Hopf algebras) math.stackexchange.com/a/2817354/348926
0
Q: How can we show that it is the local maximum?

EvindaI want to find the local maxima of $f(x,y,z)=-(x+y+z)$, when $$x^2+y=3, x+3y+2z=7.$$ I found that the local extremum is $(x,y,z)=\left(-\frac{1}{2},\frac{11}{4},-\frac{3}{8} \right)$. How can we show that $\left(-\frac{1}{2},\frac{11}{4},-\frac{3}{8} \right)$ is the local maximum?

Hello
Do you have an idea about my question?
@TobiasKildetoft I think with this perspective on the proof of Maschke's theorem, one should be able to prove the following generalization: If $R$ is a commutative ring and $A$ is a Hopf algebra over $R$ such that the counit $\varepsilon: A \to R$ has a $A$-linear section, then every $A$ module that is semisimple as an $R$-module is semisimple as an $A$-module
@MatheinBoulomenos It looks good
@TobiasKildetoft thanks
09:31
Hey @LeakyNun
Did you see my question?
@MatheinBoulomenos The generalization does sound plausible at least
(possibly you need to assume the Hopf algebra is finitely generated over $R$ or something)
Compute the hessian and see if it has det > 0
Do we not get as hessian the zero matrix? @Secret
@MatheinBoulomenos As I mentioned, the "averaging" works for symmetric algebras, and unless I am missing something, the counit of a Hopf algebra should be a symmetrizing trace
(hmm, possibly the Hopf algebra needs to be cocommutative for this)
ah, yeah cocommutative might be needed
I haven't checked the details
09:35
@MatheinBoulomenos do you have access to springer?
@LeakyNun What are you looking for?
If I'm on campus or I log into the university internet via VPN, then yeah
but I don't understand this
so could someone who can access it explain this to me, thanks
ah, I'm not sure if we have access to jstor at our uni
what are you looking for in this paper? maybe it's elsewhere
09:37
maybe I can send you a copy via email?
is that legal?
@Evinda @LeakyNun I don't think this question is a trivial one (at leas if one wants a systematic solution), because for this particular function, the Hessian and all its higher order derivatives are the zero matrix
thus derivative test is basically useless here and something else is needed to check
@LeakyNun I won't tell anyone that you did
@Secret Do you have an idea how else we could show this?
@Evinda done
@MatheinBoulomenos how is that illegal
I thought it's only illegal to mass distribute it
or sell it
I'm not sure about the legal stuff
I think it will be fine if you mail it
09:45
> (h) on an ad hoc basis and without commercial gain or in a manner that would substitute for direct access to the Content via services offered by JSTOR, sharing discrete portions of Content for purposes of collaboration, comment, or the scholarly exchange of ideas
oh, then it's completely fine
I think you should have my email
https://arxiv.org/abs/1708.04075
Other methods will involve balls in some way that I don't understand, so at your level, just parametrise it as Leaky did
@LeakyNun what question do you have about this article?
09:50
could you please go through the proof?
the idea is to use the primitive element theorem
Okay suppose $L/K$ is an extension such that every nonconstant polynomial in $K$ has a root in $L$
Then if $f$ is a polynomial over $K$, we can assume wlog that it's irreducible
Gilmer uses an argument on the roots, but you can also use formal derivatives to show that if $f$ is any polynomial, then $f(x^{p^e})$ is separable for some $e$
@LeakyNun Thanks a lot!!!
@MatheinBoulomenos $e = \deg f$?
@LeakyNun no
ok let's go there later
09:58
it's the inseparability degree of the spltting field of $f$
so in the first case, we assume that $K$ is perfect (so that we can take $e=0$) and we get that if $f$ is an irreducible polynomial over $K$, then if we take a splitting field of that (which is not a subfiled of $L$ a priori) is simple
So assume that $K(\gamma)$ is a splitting field of $f$
Then the minimal polynomial of $\gamma$ has a root in $L$
(by assumption)
let's call that root $\theta$, then $K(\gamma)$ is isomorphic to $K(\theta)$
oh wow
So $K(\theta) \subset L$ will also be a splitting field of $f$
heilige sch****e
this implies that $f$ splits into linear factors over $L$
@LeakyNun what's wrong?
das ist am kluegsten
10:02
Okay no for the case that $K$ is not perfect
$K_0$ is the maximal purely inseparable extension of $K$ inside $L$
i.e. all elements of $L$ that are purely inseparable over $K$
Now Gilmer shows that $K_0$ is perfect (I assume you know the equivalent characterizations for being perfect?)
we need to show that the Frobenius is surjective
So if we take an element in $K_0$, then the minimal polynomial of that over $K$ must be of the form $x^{p^n}-a$
I guess char=0 should really be char=1. it makes everything works.
the convention is wrong
because else $K_0$ would not be purely inseparable (I can go into more details on that if you want)
So now we can look at the polynomial $x^{p^{n+1}}-a$ over $K$. By assumption, this has a root in $L$
call $t$ a root
oh let's cal the element with the minimal polynomial $x^{p^{n}}-a$ $s$
then $t^{p^{n+1}}=a=s^{p^n}$, so $0=t^{p^{n+1}}-s^{p^n} =(t^{p}-s)^{p^{n}}$, thus $t^p=s$
which shows that the Frobenius $K_0 \to K_0$ is surjective, i.e. $K_0$ is perfect
Now if $K$ is not perfect, apply the theorem in the perfect case to $K_0$ and note that $K \subset K_0$
Is there anything unclear?
no
thank you very much
I suppose we can always use $K_0$ even if $K$ is perfect
yeah
but it will just be $K$ again if $K$ is perfect
sure
I had one thing to learn
I learnt it thanks to you
now I have three
like a hydra
1. why is it simple 2. why is that a field 3. why is it implied by Frobenius being surjective
1. If $K$ is perfect and $f \in K[X]$ an irreducible polynomial then the splitting field for $f$ is simple
10:16
1. primitive element theorem 2. transitivity of separbility 3. that's a nice exercise on separability, but I can also show you how to do it if you want
ah no, for 2 it should really say tower rule for separable degree
2. If $L/K$ is an extension where $K$ has char $p$ then the elements $x \in L$ such that $x^{p^n} = a$ for some $n \in \Bbb N$, $a \in K$ form a field
how to prove that?
hmm I guess it isn't hard
I don't see what goes wrong
ok nothing goes wrong
could you prove 3
10:19
Let $K$ be a field of characteristic $p$, then we prove that the following are equivalent:
1) Every algebraic extension of $K$ is separable
2) The Frobenius $K \to K$ is surjective
1) => 2)
Let $a \in K$, then consider a splitting field of $x^p-a$. If $\alpha$ is a root of that polynomial, then $(x-\alpha)^p=x^p-a$, so the splitting field can't be a separable extension unless $\alpha \in $K, but then $\alpha^p=a$, so the Frobenius is surjective
oh wow I just finished reading the primitive element theorem
it's comprehensible
2) => 1)
Let $L/K$ be an algebraic extension and $\alpha \in L$. Let $f$ be the minimal polynomial of $\alpha$ over $K$. We now that there is some minimal $n$ such that $f(x^{p^{n}})$ is separable and irreducible. We want to show that $n=0$. Let $f(x)=a_n x^n + \dots a_0$, then choose $b_i \in K$ such that $(b_i)^{p^n}=a_i$ (using the surjectivity of the Frobenius), then we get that if we consider $g(x)=b_nx^n+ \dots b_0$, $g^{p^n}=f(x^{p^n})$ which contradicts that $f(x^{p^n})$ is irreducible unless $n=0$
ok
thanks
35 mins ago, by MatheinBoulomenos
Gilmer uses an argument on the roots, but you can also use formal derivatives to show that if $f$ is any polynomial, then $f(x^{p^e})$ is separable for some $e$
why is this true?
multiple roots of $f$ are the same as common root of $f$ and $f'$ (by elementary properties of the derivatives)
10:35
yes
So if $f$ is irreducible, then either $f$ has only simple roots, or $f'=0$, since $\operatorname{gcd}(f,f')$ is a proper divisor of $f$
(that's how you show that in characteristic 0 everything is separable)
sorry I don't follow
multiple roots of $f$ are the same as common roots of $f$ and $f'$, which are the same as roots of $\gcd(f,f')$
$\operatorname{deg}(\gcd{(f,f')}) \leq \operatorname{deg}(f') \leq \operatorname{deg}(f) -1< \operatorname{deg}(f)$
But $\gcd{(f,f')}$ divides $f$
Ah the above was assuming that $f' \neq 0$
if $f' \neq 0$, then the above stuff on degrees makes sense
and then?
but then $\gcd{(f,f')}$ is a divisor of $f$ of strictly lower degree and $f$ is irreducible
10:41
oh!
ich bin ein Esel
so $\operatorname{gcd}(f,f')=1$, i.e. $f$ has no multiple roots
So the only way that $f$ has multiple roots is if $f'=0$
But if $f=a_nx^n + \dots + a_0$, then $f'=0$ implies that $a_i=0$ whenever $p$ does not divide $i$
this means that $f$ is a polynomial in $x^p$
so we can write $f(x)=h(x^p)$ for some $h \in K[x]$
genius
and then?
$h$ will also be irreducible, since $h(x)=h_1(x)h_2(x)$ implies that $f(x)=h(x^p)=h_1(x^p)h_2(x^p)$
then apply the same process to $h$
and then?
48 mins ago, by MatheinBoulomenos
Gilmer uses an argument on the roots, but you can also use formal derivatives to show that if $f$ is any polynomial, then $f(x^{p^e})$ is separable for some $e$
you're going down but the result is going up?
at some point, this has to terminate (i.e. we get a separable polynomial) because we decrease the degree everytime
ah lol, that was nonsense
I meant to say that every (irreducible) polynomial is of the form $f(x^{p^{e}})$ for some separable $f$
10:48
I see
20 mins ago, by MatheinBoulomenos
2) => 1)
Let $L/K$ be an algebraic extension and $\alpha \in L$. Let $f$ be the minimal polynomial of $\alpha$ over $K$. We now that there is some minimal $n$ such that $f(x^{p^{n}})$ is separable and irreducible. We want to show that $n=0$. Let $f(x)=a_n x^n + \dots a_0$, then choose $b_i \in K$ such that $(b_i)^{p^n}=a_i$ (using the surjectivity of the Frobenius), then we get that if we consider $g(x)=b_nx^n+ \dots b_0$, $g^{p^n}=f(x^{p^n})$ which contradicts that $f(x^{p^n})$ is irreducible unless $n=0$
then how does this work?
ah I did that mistake in the proof about the Frobenius as well
I think it's even easier if you don't mess that up
If $f$ is irreducible and not separable, then $f(x)=h(x^{p})$ for some $h \in K[x]$. But then if we write $h=a_nx^n + \dots + a_0$, then if we take $b_i \in K$ with $b_i^p=a_i$, we get that for $g=b_nx^n + \dots + b_0$. $(g(x))^p=h(x^p)=f(x)$
which is impossible as $f$ is irreducibe
we don't even need to repeat the process until we reach a separable polynomial, the first step is enough
@LeakyNun in my ANT1 oral exam, the first half was about subtle separability issues with extensions of $k(x)$ where $k$ is perfect and $x$ is transcendental
you get that the integral closure of $k[x]$ in a finite extension of $k(x)$ is still finitely generated as an $k[x]$-module
10:54
oh can you have a finite extension with infinitely many intermediate fields?
but this is nontrivial and you need that $k$ is perfect
@LeakyNun yes
like?
Do you know this nice fact about vector spaces over infinite fields? If $k$ is an infinite field and $V$ is a vector space over $k$, then $V$ is not the finite union of proper subspaces
You can show that a finite extension is simple iff it has only finitely many intermediate fields
the above lemma on LA gives an easy proof for the one direction
if $L/K$ has only finitely many intermediate fields, then $L$ is not the union of all proper subextensions (by the lemma), so choosing $\alpha \in L$ such that $\alpha$ is not contained in any proper subextension gives you that $L=K(\alpha)$
(you can use this + Galois theory to prove the primitive element thereom, too)
So it suffices (and is in fact equivalent, but the other direction is harder to show iirc) to give a finite extension that is not simple
For this you can consider $\Bbb F_p(x^p,y^p) \subset \Bbb F_p(x,y)$
11:10
hmm
11:44
Hello, I'm trying to figure out this limit:
$\lim_{t\rightarrow 0}f_t(x)=\lim_{t\rightarrow 0}\frac{\left | u(x)-tv(x) \right |^{p}-\left | u(x) \right |^{p}}{t}$
u and v are just some Lp functions
does anyone have an idea of how to find this? I tried l'Hospital but I'm not sure
11:59
so $u(x)$ is a number $a$, $v(x)$ is another number $b$, and you're looking at $\lim \frac{|u-tb|^p-|u|^p}t$ ?
yes
that looks like the definition of the derivative of a certain function of $t$
do you mean the function $f(t)=|a-tb|^p$
?
yeah
watches himself mess up his own notation at the first use of it when writing $u$ instead of $a$
lol
ok I'm writing it down
12:06
do you have $p>1$ ?
$\lim \frac{|a-tb|^p-|a|^p}t=b\lim\frac{|a-tb|^p-|a|^p}{bt}=b\left | a-tb \right |^{p-1}(-b)$
yes p>1
I think my answer is incorrect
I shouldn't have b^2
you shouldn't have a $t$ at all
I meant the derivative of a certain function of $t$, at $t=0$
ah oops I meant to say
b\left | a \right |^{p-1}(-b)
bp\left | a \right |^{p-1}(-b)
$b\left | a \right |^{p-1}(-b)$
third time is a charm
I also think you shouldn't have $b^2$
also the sign of the thing should depend on the sign of $(ab)$
not on the sign of $b$
but I don't know the sign of ab
backtracking, I'm looking for the derivative of $|a-tb|^p$ which is:
$p|a-tb|^{p-1}(-b)$
right?
12:17
wrong
ah so that's where I'm mistaken
what is it then?
it's that times the sign of $a$
why?
because the vertical bars are an absolute value
well, rather it's the sign of $(a-tb)$
yes and the derivative of the abs is 1 for x>0 and -1 for x<0
12:19
yes precisely
but you didn't include it
ah ok yes it's the sign of $(a-tb)$
but how did you infer that it would be just the sign of $a$?
because we want the derivative at $t=0$
ohh ok. So now at t=0 we have: p|a|^{p-1}(-b)sgn(a)
that looks good
you could even say $-bp|a|^p/a$
because sgn(a)=abs(a)/a
?
yeah
that's neat
thank you! :D
12:25
and if $a=0$ and $p>1$ it is $0$
why does p>1 here matter?
if it's not then it's not differentiable when $a=0$
ah yes I see
Thanks again @mercio! You've helped me more than once
I'm glad to hear that, though I don't remember lol
12:46
because it was several months ago, I recall asking a question about a particular ODE and a question in complex analysis
12:59
Problem: Let $V$ be a finite dimensional vector space, and let $W_1$ and $W_2$ be subspaces of $V^*$, the dual space. Show that $\mbox{Ann}(W_1) = \mbox{Ann}(W_2)$ implies $W_1 = W_2$...I could use a hint on how to solve this.
@MatheinBoulomenos I thanked you for this answer: " $\Lambda(V \oplus W) \cong \Lambda(V) \otimes \Lambda(W)$ (taking the graded tensor product), this holds for vector spaces, but it also carries over to bundles. Now if we take the projections $p_1: M \times N \to M$ and $p_2:M \times N \to N$, then we have $T(M \times N) \cong p_1^*TM \oplus p_2^* TN$, so we get $\Lambda(T(M \times N)) \cong \Lambda(p_1^* TM) \otimes \Lambda(p_2^* TN)$, so looking at the componenet in degree $m+n$,
we get $\Lambda^{n+m}(T(M \times N)) \cong \Lambda^{m}(p_1^*TM) \otimes_{\Bbb R} \Lambda^{n}(p_2^*TN)$"
I recognise Inn, Out, Aut but never seen Ann before
@user193319 try choosing a basis
@Secret It represents the annihilator subspace: $\mbox{Ann}(W_1) = \{v \in V \mid f(v) = 0 ~\forall f \in W_1\}$.
@loch A basis for $V$, $V^*$, or something else?
hmm...
13:10
well a basis for $V$ gives you a basis for $V^*$

anyway you can choose basis $e_1,\ldots,e_n,f_1,\ldots,f_m,g_1,\ldots,g_l$ where $e_i$ form a basis for $W_1\cap W_2$, $e_i,f_j$ form a basis for $W_1$, $e_i, g_k$ form a basis for $W_2$

and then you'll find that if $m,l \ne 0$ then you'll find linear forms which annihilate $W_1$ and not $W_2$ etc.
$v \in V$, $f_1(v)=0, f_2(v)=0$, $f_1(v)+f_2(v)=0+0=0 \implies (f_1+f_2)(v) = 0$
o wait...
$w_1 \in W_1,w_2 \in W_2$,$f \in V$, $f(w_1) = 0, f(w_2)=0$, $f(w_1+w_2)=f(w_1)+f(w_2)=0$
well then the first thing that pops out is that $dim(W_1)=dim(W_2)$ otherwise there is no way they can share the same set of $\text{Ann}$
as for $W_1=W_2$... two subspaces of the same dimension has to be identical because they are spanned by the same number of basis vectors?
I wonder how that will break down if $V^*$ is infinite...
13:39
@loch hi
@LeakyNun hey
13:54
For a group you need inverse elements
What if your elements are functions
Then could you say that your inverse elements are the inverse functions

« first day (2870 days earlier)      last day (2447 days later) »