dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Jan 05, 2009 at 03:26:14PM -0500, Shawn Walker wrote:
> ok, then this higher order mesh data should definitely have it's own  
> function space.  Is it possible to have a FunctionSpace variable (for the 
> mesh) within the usual FunctionSpace variable for the variational 
> problem? It's either that or have two separate FunctionSpace variables 
> floating around and pass them to the bilinear form everytime.
>
> - Shawn

I imagine something like this:

A FunctionSpace V has a variable coordinates (a Function).

That Function in turn has a FunctionSpace W (as all other Functions).

V does not have to be the same as V, but it can (for isoparametric
elements).

-- 
Anders

Attachment: signature.asc
Description: Digital signature