ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Fri, Sep 22, 2006 at 02:39:38PM +0200, Johan Jansson wrote:
> On Tue, Sep 19, 2006 at 08:25:45PM +0200, Anders Logg wrote:
> > I've added support for evaluation of functionals to FFC and DOLFIN.
> > An example can be found in src/demo/pde/functional
> > 
> > A functional is defined as follows in FFC (here an energy norm):
> > 
> >   v = Function(element)
> >   M = (v*v + dot(grad(v), grad(v)))*dx
> > 
> > Then in DOLFIN, just assemble the functional directly:
> > 
> >   EnergyNorm::Functional M(v);
> >   real value = FEM::assemble(M, mesh);
> > 
> > Note that functionals may be defined for an arbitrary number of
> > arguments and that they don't need to be linear.
> > 
> > /Anders
> 
> Very nice. So the only missing piece for automated error estimation is
> to construct the residual (and the dual eventually)?
> 
>   Johan

And differentiation of functionals and nonlinear forms.

/Anders