[dune-pdelab] dune-pdelab | Feature: Divergence and curl as DGF (#15)

Jö Fahlke jorrit at jorrit.de
Thu Jun 25 15:47:46 CEST 2015


Am Thu, 25. Jun 2015, 14:06:12 +0200 schrieb Christian Engwer:
> Date: Thu, 25 Jun 2015 14:06:12 +0200
> From: Christian Engwer <christian.engwer at uni-muenster.de>
> To: dune-pdelab at dune-project.org
> Subject: Re: [dune-pdelab] dune-pdelab | Feature: Divergence and curl as
>  DGF (#15)
> 
> Perhaps we should quickly vote on this problem:
> 
> With 3.0 we plan to integrate dune-functions with PDELab. This means
> that the DiscreteGridFunction interface asa we know it currently will
> not supported anymore.
> At the same time we want new DGF implementations for certain
> applications. In order to avoid confusion and unnecessary deprecatioan
> cycles, we should make clear to the user that some features are
> present but not actually a stable interface.
> 
> a) we don't care and introduce these new DGF implementations
> b) we introduce a cmake switch similar to the
>    experimental-grid-extensions
> c) we introduce a namespace experimental
> 
> Current votes:
> 
> Steffen: (c)
> Christian: (c)
> 
> If you care; please add you to the list.

I do care, but I cannot give an opinion since nobody cares to explain the
original issue to me.  And I did ask [1].  What is wrong with the existing
curl DiscreteGridfunctions?

[1] http://conan2.iwr.uni-heidelberg.de/git/pdelab/dune-pdelab/merge_requests/15#note_1068

And I also don't understand the point about unnecessary deprecations.  When
the revolution (i.e. dune-function) comes we will probably deprecate _all_ old
discrete gridfunctions, whether curl or div or just plain old regular ones.
Until then we use the existing ones, and where those are not enough we can
continue to create special ones in the style of the existing ones and
alongside the existing ones (not an option above).  We will have to deprecate
them, yes, but they will very likely be deprecated along with all the other
old ones, so I don't see any additional confusion there.

The picture is a little different if you plan to create the special
gridfunctions already in the style of dune-functions.  _Then_ they are
experimental, and can go into a namespace experimental (c).  Or can be
controlled by a buildsystem-switch (b), but I don't quite see what we would
gain from that over (c), besides additional work.

Regards,
Jö.

-- 
Jorrit (Jö) Fahlke, Institute for Computational und Applied Mathematics,
University of Münster, Orleans-Ring 10, D-48149 Münster
Tel: +49 251 83 35146 Fax: +49 251 83 32729

Der Horizont vieler Menschen ist ein Kreis mit Radius Null - und das
nennen sie ihren Standpunkt.
-- A. Einstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 811 bytes
Desc: Digital signature
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20150625/cc8d2002/attachment.sig>


More information about the dune-pdelab mailing list