On Mon, Mar 24, 2014 at 11:41 AM, Stefano Zampini <[email protected]
> wrote:
> Matt,
>
> I wish to test PCBDDC with maxwell equations in H(curl). Can you tell me
> how far are the actual implementation of PetscFe and PetscSpace to
> implement lowest order nedelec elements and linear functionals evaluation?
>
The easiest thing to do is look at the implementation in FIAT first and see
what they did. I am attaching
the relevant chapter from the FEniCS book. I would use Nedelec of the 2nd
kind since then you can
use plain old polynomial P spaces. You just have to code up the dual space,
which is just a bunch of
quadrature rules over edges and faces.
The big pain will be that it has vector degrees of freedom, which I have
not tested yet. You need to transform
the input vectors before doing the integral on the reference element. I can
help if you get stuck.
Thanks,
Matt
> best,
> --
> Stefano
>
--
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
