syfi team mailing list archive

[Robert Cimrman <cimrman3@xxxxxxxxxx>]

Anders Logg wrote:
> So there are two steps: First number all entities based on a
> lexicographical ordering as in the UFC manual, then number everything
> else based on the global index.
> 
> UFC is only concerned with the first step.
> 
> We also need to define something we can call "alignment", which is how
> two cells meet at a common facet. For two triangles, the alignment is
> 0 or 1. For two tetrahedra, the alignment is 0, 1, 2, 3, 4 or 5.
> This can also be based on a lexicographical ordering of matching
> vertices.

Hi,

I have installed SyFi with the aim to integrate it with my
NumPy/SciPy-based Python FE code (just released at
http://ui505p06-mbs.ntc.zcu.cz/sfe), so I am silently watching this
list. Now I have hijacked your discussion because I have implemented
this alignment (called orientation) in C, based on lexicographical
naming, too, albeit for edges only (0 or 1).
In 3D I have not needed more than P2 elements (edge nodes only), so I
avoided the thing, as it was getting complicated (tetra- and hexa-edra
together, wedge elements, pyramides...). Are you going to consider also
those mixed element type meshes? Do you allow degenerated elements? I
would like to have this finally in my code. (I silently assume you need
the alignment to be able to assign global numbers to the extra element
nodes, that are not present in the underlying mesh. If this was not the
case, ignore the message, please.)

best regards,
Robert Cimrman