[dune-functions] Problems writing vector-valued vtk files

[Christian Engwer]

On Thu, Mar 19, 2015 at 10:12:40AM +0100, Carsten Gräser wrote:
> Am 18.03.2015 um 22:14 schrieb Christian Engwer:
> > This would become easier when introducing tensors ;-)
> Perhaps. However I still think that the user will have
> to decide in the nontrivial cases.

well, when properly implementing tensors (don't get me wrong, I'm not
sure we want to implement this...), the type of the tensors involved
in a multiplication would imply whether this is a contraction or a
tensor product.

Christian

> Carsten
> 
> > Christian
> > 
> > On Wed, Mar 18, 2015 at 04:18:30PM +0100, Carsten Gräser wrote:
> >> Am 18.03.2015 um 16:13 schrieb Oliver Sander:
> >>>>> One could argue that it is called ...scalar... because it uses scalar bases,
> >>>>> even to create vector-valued functions.  I'm not perfectly sure that this is
> >>>>> a good long-term point of view.
> >>>> In the long-term we should get rid of 'scalar'. The only thing
> >>>> to specify is how to multiply coefficients and values of the local
> >>>> basis. Maybe we can factor this out in a first step such that one
> >>>> can povide this as a lambda.
> >>>
> >>> I've been thinking along the same lines.  Getting rid of the 'scalar' would be nice;
> >>> the users shouldn't have to worry about scalar and nonscalar bases too much.
> >> I'm afraid he has to. Because if coefficients and basis functions are
> >> vector-valued, there's no longer a canonical range type for the
> >> function. It get's even worse, for matrix valued coefficents.
> >>
> >> Best,
> >> Carsten
> >>
> >>>
> >>>>
> >>>> For the moment I've implemented support for vector/matrix-valued
> >>>> coefficients only.
> >>>
> >>> Works-for-me, thanks!
> >>>
> >>>>
> >>>> Best,
> >>>> Carsten
> >>>>
> >>>> BTW:
> >>>> Generic multiplication of various matrix/vector/scalar
> >>>> combinations is something we've put a significant amount of
> >>>> work into in dune-fufem.
> 
> 
> _______________________________________________
> dune-functions mailing list
> dune-functions at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune-functions
> 

-- 
Prof. Dr. Christian Engwer 
Institut für Numerische und Angewandte Mathematik
Fachbereich Mathematik und Informatik der Universität Münster
Einsteinstrasse 62
48149 Münster

E-Mail	christian.engwer at uni-muenster.de
Telefon	+49 251 83-35067
FAX		+49 251 83-32729