[Dune] Re: Bemerkung Gitter und Gitterhierarchie.

[Robert Kloefkorn]

Hi Oli, hi others,

> The problem I see
> is that in the finite element community there is a pretty clear
> conception of what a grid is, and this conception does not
> include hierarchic structure.
I only know a view people from the finite element community, but I think
that they don't see it like this way. (for example A. Schmidt and K.G
Siebert to name some of them)

> In fact you hardly ever need that
> hierarchic structure anyways ???

You need the hierarchic structure every time. For example during
Adaptation Data prolongation and restriction can be done much easier and
more efficient, when using the HierarchicIterator instead of level wise
adjusting the data.

In one of your comments to the paper you mentioned, that one should
adjust the definitions in the grid paper to what is available "on the
market". I don't see is that way.

In our paper we want to mathematically describe what a Dune Grid is.
We might find in the literature some definitions and constructions that
are useful and then we should perhaps use them. But normally it should
be possible to write the paper without any definition of what a simplex
or whatever is from other papers. That's my opinion to this point.
The paper is only good for presenting an abstract framework to what we
already implemented.

> Here is my view on the issue:
> The dune-grid interface implements a 'grid hierarchy', which
> is a set of (level) 'grids'.  'Grids' have no hierarchical structure
> at all.

Then my first question would be: Why don't we have a hierarchy of Grid
classes implemented in Dune instead of one Grid class? Why are we
presenting a abstract definition that later is not implemented in the
code. I think that does not make sense to anyone.

Cheers,

Robert


-- 

  Robert Klöfkorn           <robertk at mathematik.uni-freiburg.de>

  Mathematisches Institut              Tel: +49 (0) 761 203 5631
  Abt. für Angewandte Mathematik       Fax: +49 (0) 761 203 5632
  Universität Freiburg
  Hermann-Herder-Str. 10
  79104 Freiburg

  http://www.mathematik.uni-freiburg.de/IAM/homepages/robertk