dolfin team mailing list archive

["Johan Hoffman" <jhoffman@xxxxxxxxxx>]

> On Thu, Aug 21, 2008 at 02:24:15PM +0200, Johan Hoffman wrote:
>> We are about to start a msc project on adaptive geometry modeling in
dolfin. The idea is to introduce surfaces in dolfin, representing the
"exact" geometry, separate from the mesh representing the "approximate"
geometry, to be able to adaptively refine the mesh geometry
>> approximation.
>> This is a continuation of our previous work on parallel adaptivity in
dolfin (currently discussed in another thread), and also here
>> distributed
>> parallel computing issues will be considered from the start.
>> It comes down to introducing a class Geometry, where surfaces are
represented in a suitable way, typically close to a standard CAD
representation (e.g. NURBS), with the basic functionality being a
projection of an arbitrary point onto a surface, and the ability to
import
>> standard CAD surfaces.
>> This should be fairly orthogonal to other ongoing work in dolfin as far as
>> I am aware, but we are happy for any input to consider.
>> /Johan
>
> Sounds good. Do we need a DOLFIN XML format for geometries? Or are there
standard formats we should use (or convert from)?
>
> --
> Anders

I think it may be natural to have a DOLFIN XML, being as simple as
possible, using a representation (NURBS?) that we would like to use
internally in dolfin. And we would like to convert from standard CAD
formats, and related formats such as cgns (www.cgns.org). But this is not
set yet and will be part of the project, and is open for a discussion
here.

/Johan

 _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev
>