Re: [math-fun] A New Way To Solve Linear Equations

15 Aug 2012

      * Joerg Arndt <arndt@jjj.de> [Aug 15. 2012 08:44]:
...
* Ray Tayek <rtayek@ca.rr.com> [Aug 15. 2012 08:03]:
...
https://rjlipton.wordpress.com/2012/08/09/a-new-way-to-solve-linear-equation...
[...]
...
Impossible but true: a new approach to linear systems
The non-mentioning of "black box" methods make me wonder.
Richard Brent could certainly give a more pertinent comment.
Should have read the comments beforehand:

John Sidles
August 9, 2012 12:09 pm
  As I understand (imperfectly no doubt) the above remarks, the
  process outlined bears a considerable similarity to (what are
  called) Krylov subspace methods. Perhaps some person more
  knowledgeable than me will comment upon this! :)

cf. http://en.wikipedia.org/wiki/Krylov_subspace