Q0

[gdmcbain]

I see that ElementQuad0 isn't there. I want this for projecting (quadrilateral) Taylor–Hood #241 results for output to a finite volume code. I'll try adding it.

[gdmcbain]

I see that ElementTriP0 isn't used anywhere. I wonder what a reasonable demonstration or test would be?

[gdmcbain]

Perhaps I'll modify the forced convection ex25 from MeshTri to MeshQuad and then project to ElementQuad0 for plotting with MeshQuad.plot.

[gdmcbain]

Qualms: Is it wrong to derive ElementQuad0 from ElementH1? Should the second value returned by lbasis, the gradient of the basis, be undefined rather then zero?

Is what's sought here more like a discontinuous Galerkin element? Or are P0, Q0, &c., conventionally accepted aberrations?

[gdmcbain]

Is the rationale that the abscissæ X passed to lbasis are always quadrature points and so always interior (is that true?)? And there the gradient is indeed well defined and zero.

(I think I have seen quadrature rules with points on facets or vertices, but maybe only for special tricks which mightn't concern us here.)