I'm working in the metric space $l^{\infty}$ trying to show that a certain subset is not complete. So I construct a sequence of sequences $(x_n)=(x_1,x_2,\ldots)$ where for each $x_n \in (x_n)$ $$x_n=(1,1/2,1/3,\ldots,1/n,0,0,0,0,0,\ldots)$$ The confusion creeps in when I look at an example in Kreyszig-"Introductory Functional Analysis with Applications" where he defines an arbitrary sequence of sequences as$$(x_m)=(\zeta_1^{(m)},\zeta_{2}^{(m)},\zeta_3^{(m)},\ldots)$$ As he wants to prove that it's Cauchy he goes on to say that $$d(x_m,x_n)=\sup_{j \in \mathbb{N}}|x_j^{(n)}-x_j^{(m)}|$$ Now I do not understand where he's going with this notation, as it seems that he's defining the $m$'th element of his sequence as the sequence taking the $m$'th element of $\zeta_1$ as the first element, and the $m$'th element of $\zeta_2$ as the 2nd element and so on and so forth.

This seems to me to be a rather unnatural way of going about it, as I'd like to just define the sequence such that the $m$'th element is simply $\zeta_m$. In fact that's how I need to define my original sequence to get the result I need.

So my question is: Is my simplistic way of defining a sequence in a sequence space correct? If so how should I notate it so I can express the metric properly when dealing with $x_m$ and $x_n$? Or am I just completely misunderstanding Kreyszig's notation and he is expressing what I want to?

Specifically for my sequence I want $(1,0,0,0,\ldots)$ to be the first element of the sequence, $(1,1/2,0,0,0,\ldots)$ to be the second element and $(1,1/2,1/3,\ldots,1/n,0,0,0,\ldots$ to be the n'th element of the sequence, and so on. What notation should I use to describe this sequence?


I believe you are misunderstanding Kreyzig's notation. When he uses the index m, he means the m'th sequence of the sequence of sequences.

... as the sequence taking the $m$'th element of $\zeta_1$ as the first element

Here is the root of your misunderstanding, $\zeta_1$ is nothing here (unless you define it), $\zeta_1^{(m)}$ is the first element of the $m$'th sequence.

I'm not sure if you understood the wanted metric, he is comparing for these two sequences $x_m$ and $x_n$, the difference between each coordinate.

| cite | improve this answer | |
  • $\begingroup$ Ok I'll edit my question to define what I want from my sequence and if you can show me how to notate it I'd be much obliged. $\endgroup$ – K.Power Feb 23 '16 at 17:43
  • 1
    $\begingroup$ Oh you want only a good looking way to write it? I mean, you can always write in words (as you described), I believe there is no need to use too much mathematical notations if in words is simpler. Another way I see it could be is to define the canonical vector $(e_m)= (0,0,...,1,0,0,.. )$ and then your sequences would be $(x_m)= \sum_{n=1}^m e_n \frac{1}{n}$. And this should be super friendly for checking non-Cauchyness $\endgroup$ – pancho Feb 23 '16 at 18:06
  • 1
    $\begingroup$ Thanks so much, that's what I was looking for! I do understand Kreyszig's notation now too so thank you. $\endgroup$ – K.Power Feb 23 '16 at 19:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.