Hum, anyone knows of an already implemented (or quick to implement) algorithm in MMA to invert a toeplitz matrix? I mean, one that isn't O(n^3) and that doesn't require to build it
@acl I did say once before, he acts like the sort of guy who frequently gets his teeth kicked in at bars...
@Rojo Well, if you're asking how to implement Levinson-Durbin functionally... it looks ugly. I'd stick to the procedural implementation. Are your Toeplitz matrices SPD? Banded?
@Rojo I don't particularly trust the "superfast" methods; they only really pay off for large dimensions, and there's the instability you mentioned.
The good thing is, that my daughter is too little to be on facebook and in a few years we are running out of oil anyway.. No oil, no energy, no internet, no facebook.. Unfortunately no Mathematica too. The we can tell stories from the past when we were good at something ;-)
yesterday I experienced one of the joys of introducing mathematica to people for the first time
the point in time at which they realize that if they select a large amount of code, press a random key and then enter by mistake, their code is gone for ever!
"not, there is no other undo; that's all. when did you last save?"
yes. it's one of the main reasons I mainly code in workbench now
(although I always end up starting in the frontend, growing my code beyond what I find comfortable in the fronted, moving to workbench and wasting time renaming symbols like [Alpha] to alpha etc)
oh well nothing is perfect
as NoEscape so eloquently put it, the frontend is lacking :)
This is a spinoff of this question
Defining
$$f_0(x) = x$$
$$f_n(x) = \log(f_{(n-1)} (x)) \space (\forall n>0)$$
and
$$a_0 = 1$$
$$a_{n+1} = (n+1)^{a_n} \space (\forall n>0)$$
How to calculate
$$\lim_{n \to \infty } f_n(a_n) $$
(an "experiment" here)
This has come up a few times now: migrating old questions causes problems.
The most popular solution to this to date has been to reset votes on migrated posts. This is do-able, but frankly it feels like treating a symptom, and doing so in a way that penalizes folks who answer what are by all ap...
So now nobody can migrate old questions, period. However, we have a process in mind for allowing organized migration of older questions to a smaller site. Keep an eye out for that in the next few days... also on MSO. I'll drop by here with a link once it goes up.
@rm-rf Could we apply for migrating all Mma tagged questions here at once and decide by ourselves what to do with them? When SO gets to zillions of C++ questions, out little nest there will be migrated to magnetic tape
@belisarius No, that would be chaotic... 2k+ questions is a lot. Most of them are fine where they are... we should only focus on those few really good ones that are more about the quirks of the interface/software and language than general programming issues.
@belisarius There is no alternate here... we can't move fast. Things aren't going to be migrated en masse, no matter what. I'm already facing a very stiff resistance for the handful of questions that we asked for. Better to pick your battles... and your questions.
This is a spinoff of this question
Defining
$$f_0(x) = x$$
$$f_n(x) = \log(f_{(n-1)} (x)) \space (\forall n>0)$$
and
$$a_0 = 1$$
$$a_{n+1} = (n+1)^{a_n} \space (\forall n>0)$$
How to calculate
$$\lim_{n \to \infty } f_n(a_n) $$
(an "experiment" here)
