Talk:Euler's continued fraction formula

Starting a new page
Hi, all!

Just a quick note to outline my intentions for this new article. I intend to present the formula the way Euler did (but in English, not Latin :), then state it in a more modern formulation, then indicate how it fits in with the articles convergence problem and fundamental inequalities, and cap it off with a few examples involving familiar power series (logarithm, exponential, maybe an arctan formula). DavidCBryant 23:19, 10 January 2007 (UTC)

Connection between complex logarithms and Pi
The continued fraction derived from the Mercator series for logz provides a cute continued fraction expansion for &pi; when z = i. The algebra is pretty simple, but since it involves a value on the boundary of the circle of convergence for the series, I really should check around for some stuff (like Abel's test for convergence on the boundary) and reference that before sticking the example into this article. DavidCBryant 20:20, 12 January 2007 (UTC)

Irrationality of e
David, first of all I think it's a very nice page that you've made! Many of us either never learn about continued fractions (what with all those series and infinite products), or at any rate are not all that familiar with it, beyond perhaps periodicity of continued fractions for quadratic irrationalities, bad approximability of $$\phi,$$ etc. I was actually looking for Euler's proof for irrationality of e: apparently, his was based on continued fractions, the now standard infinite series proof with factorials is due to Fourier is from a later period, 1815 or so. I can pretty much guess what it (Euler's proof) is from the continued fraction expansion of $$e^z$$ on the page, but if you have a copy of Introductio, as it appears, do you think you could check for it and include it (the proof)? It would seem like a nice application of the expansions. I suspect Lambert's proof of irrationality of $$\pi$$ was also based on the expansion that you gave. Arcfrk 06:12, 13 March 2007 (UTC)


 * Hi, Arcfrk. Thanks for the kind words. Sorry to be so slow to respond.
 * I don't have a copy of Euler's book. Perhaps I should have written the footnote differently – what I have is a book (Wall) that quotes a little bit of Euler's stuff.
 * I did do a bit of poking around, and I think Euler's proof that e is irrational was based on the regular continued fraction expansion for e, which is

e = 2 + \cfrac{1}{1 + \cfrac{1}{2 + \cfrac{1}{1 + \cfrac{1}{1 + \cfrac{1}{4 + \cfrac{1}{\ddots}}}}}} $$
 * or, in a more compact notation, e = [2; 1, 2, 1, 1, 4, 1, 1, 6, 1, 1, 8, ...]. According to Szüsz and Rockitt, Euler put this expansion of e in his Introductio, but it's not clear that he (Euler) could actually prove it. Szüsz and Rockitt do give a proof based on the continued fraction expansion

\frac{e^{2/m}+1}{e^{2/m}-1} = m + \cfrac{1}{3m + \cfrac{1}{5m + \cfrac{1}{7m + \cfrac{1}{\ddots}}}}. $$
 * They attribute this proof to Oskar Perron. When m = 2 a correspondence can be drawn between the convergents of this fraction and the 2nd, 5th, 8th, 11th, etc convergents of the regular continued fraction for e &hellip; it's kind of a pretty demonstration, so when I have the time I'll try to attach it to the article about e somehow. DavidCBryant 18:36, 4 April 2007 (UTC)

Error in "Notes"
The 4th note:

^ a b This series converges for |z| = 1, except when z = ±1, by Abel's test (applied to the series for log(1 − z))

makes no sense since it is self-contradictory! It should read:

^ a b This series converges for |z| < 1 by Abel's test (applied to the series for log(1 − z)).

However, I am unable to edit the Notes myself. Is there a "master editor" who can do so? —Preceding unsigned comment added by Glenn L (talk • contribs) 19:23, 11 August 2008 (UTC)


 * I fixed the note, and anyone who can edit the page, can edit the notes. They have to be edited from where they are in the aticle.Eli355 (talk) 12:54, 8 June 2018 (UTC)

From a deleted article
The following proposed identity was in an article that got deleted. Somewhere along the way it was proposed that this identity be merged into this article:

\log\frac{1+w}{1-w} = \cfrac{2w}{1 - \cfrac{w^2}{3 - \cfrac{4w^2}{5 - \cfrac{9w^2}{7 - \cfrac{16w^2}{\ddots}}}}}\, $$ Michael Hardy (talk) 18:26, 1 February 2009 (UTC)

Continued fraction for pi
Currently, the continued fraction for pi is derived by using the continued fraction for the inverse hyperbolic tangent (since $$ \tanh^{-1} x = \frac1{2} \ln \left(\frac{1 + x}{1 - x}\right)$$) and then evaluating this when $$x = i$$. I propose that the continued fraction for pi be moved to under the continued fraction for the inverse tangent, and evaluating this when x = 1. This is better since it does not involve complex numbers. The continued fraction for the natural logarithm should be for $$ \ln (1 + x).$$ Eli355 (talk) 13:10, 8 June 2018 (UTC)
 * This has been done. Eli355 (talk) 20:08, 2 July 2018 (UTC)

Needlessly confusing approach
The section Euler's formula begins as folllows:

"If ri are complex numbers and x is defined by

x = 1 + \sum_{i=1}^\infty r_1r_2\cdots r_i = 1 + \sum_{i=1}^\infty \left( \prod_{j=1}^i r_j \right)\,, $$ then this equality can be proved by induction^

x = \cfrac{1}{1 - \cfrac{r_1}{1 + r_1 - \cfrac{r_2}{1 + r_2 - \cfrac{r_3}{1 + r_3 - \ddots}}}}\, $$.

"Here equality is to be understood as equivalence, in the sense that the $n$'th convergent of each continued fraction is equal to the $n$'th partial sum of the series shown above. So if the series shown is convergent – or uniformly convergent, when the ri's are functions of some complex variable z – then the continued fractions also converge, or converge uniformly."


 * I find this approach to be needlessly apt to confuse readers.


 * Suggestion: Just define the nth convergent by a nicely displayed finite continued fraction, and state that it is equal to the nth partial sum of the summation.


 * Then you could display the equality of the limits of both sides of the equation as n → ∞.


 * (Which avoids the need to redefine the word "equality" — never a good idea.)