[petsc-dev] How do you get RIchardson?

[Matthew Knepley]

On Fri, Sep 16, 2011 at 5:16 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> On Sat, Sep 17, 2011 at 00:09, Matthew Knepley <knepley at gmail.com> wrote:
>
>> So, you can accomodate your "Picard" iteration as Newton, but we cannot
>> accomodate the Picard iteration for
>> a system of nonlinear equations with that. So I made a class, named
>> Picard, to accomodate it. Are you really
>> still arguing about this?
>>
>
> This was a ridiculous argument when we first had it two years ago. I don't
> like using the name Picard for the implementation because it doesn't do what
> most PDE solver people refer to as Picard. Since both cases are strict
> subsets of what mathematicians call Picard, I don't think the name should be
> used for that thing in PETSc. Richardson is fine with me.
>

My definition is in no way a "strict" subset. Define your nonlinear operator
to have a solve, and it has what you want.

   Matt

-- 
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their experiments
lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20110916/54b86dc8/attachment.html>