On Thu, Aug 4, 2011 at 3:21 PM, Roy Stogner <roystgnr@...> wrote:
> On Thu, 4 Aug 2011, John Peterson wrote:
>> Something about the mapping to the reference element changes as well.
>> Piola tranformation?
>> was all I could find on short notice.
> Oh, that's right. The vector directions depend on the element shape.
> So you need the Piola transformation to maintain H(div) conformity...
> and isn't it different (covariant transformation?) for H(curl)?
Well, deal.II defines separate enums for covariant, contravariant,
piola, nedelec, and raviart-thomas elements....
but some of those may be implemented the same way.