Talk:Remez algorithm

Someone should modify the 'Procedure' section to flow more like: http://www.math.unipd.it/~alvise/CS_2008/APPROSSIMAZIONE_2009/MFILES/Remez.pdf 's guide to remez algorithm.. much better and more reliable --skimnc — Preceding unsigned comment added by Skimnc (talk • contribs) 07:50, 19 September 2011 (UTC)

Does the person who made the changes before me have a source to cite? I don't see here that the algorithm as stated here converges. --Pftupper 02:27, 4 November 2006 (UTC)
 * Yes, see Tony Ralston's NA books, or Forman Acton's 66.94.9.51 00:04, 5 November 2006 (UTC)
 * I'm still not convinced. Please cite more specifically, title, date, section, definition number, page. Penguian (talk) 03:14, 20 January 2008 (UTC)
 * I too agree that this algorithm is not correct. Here is a simple description of it: http://www.math.unipd.it/~alvise/CS_2008/APPROSSIMAZIONE_2009/MFILES/Remez.pdf —Preceding unsigned comment added by 131.130.17.148 (talk) 10:06, 18 January 2010 (UTC)

Can someone explain to me what in the section "Procedure" the E is and where it does come from? Does it stand for Error, and if so, for what kind of error? This is totally unspecified —Preceding unsigned comment added by 129.13.186.1 (talk) 15:43, 11 May 2009 (UTC)

Only the minmax
The first paragraph suggests that Remez algorithms can be used for more than arriving at just the minmax polynomial in the sense of Chebychev, such as optimal rational approximations. Then the second paragraph describes an algorithm restricted to the minmax. So I'm afraid this leaves a hole. My impression is that (from the titles of the papers) that this description is (more or less) correct for the original algorithm. I'm personnally interested in polynomial approximations. It looks like the modification in Remez algorithm is so small that authors tend to gloss over the difference. If nowadays Remez is used for these other types too, that must be acknowledged, of course. [An iteration who states that results (in floating point) must become equal is unsatisfactory IMHO.] —Preceding unsigned comment added by 80.100.243.19 (talk) 19:40, 8 October 2010 (UTC)

Lebesgue constant
In the section "On the choice of initialization", much important information about the Lebesgue constant is provided while in turn missing in Lebesgue constant: maybe move part of it? — Preceding unsigned comment added by 84.58.8.167 (talk) 03:46, 17 September 2012 (UTC)

Convergence
Does this algorithm converge? Under what conditions does it converge? The introduction and main body of this article does not seem to be pointing out these facts. Can someone help? Jay (User talk:Shantham11) 10:01, 4 December 2013 (UTC)