Talk:Successive parabolic interpolation

Convergence Rate
I don't think that saying

"...an extremum can be found in approximately O(1.324) time using function values alone..."

makes sense as that would designate an algorithm of constant running time, where the author of this article does not even say what the running time may be dependent on.

Since this is an interpolation technique and not a difference function (i.e. a sequence of numbers converging to a fixed point), we probably should be talking about either algebraic or exponential convergence rates, eh? --Somewikian (talk) 11:47, 4 February 2009 (UTC)


 * Let me correct myself. This is about a converging sequence of successive numbers, i.e. we should be talking about convergence rates in its usual sense, i.e. see what they have at Rate_of_convergence and alter this article to fit that notation. --Somewikian (talk) 11:51, 4 February 2009 (UTC)
 * Yes, it was sloppy to use the big-O notation. I have attempted a clarification. Thanks, Btyner (talk) 14:53, 7 February 2009 (UTC)

Reference
Is there a better primary reference? The given reference is from 2002, surely there's an earlier, more authoritative one. 130.216.54.198 (talk) 01:29, 1 July 2009 (UTC)
 * I'd say Heath is considered pretty authoritative in his field, but you are more than welcome to add another. Btyner (talk) 02:10, 1 July 2009 (UTC)