I am not sure why this happened, other than to eliminate a 3rd party
dependence. The last I looked there was an attempt at doing the
derivatives symbolically with sympy, but I haven't messed with this, and
it may not be hooked up to the functionals).
Please note that this was always quite experimental as ffc is currently
incapable of dealing with derivative degrees of freedom -- the
transformations to the reference element are subtle.
Best,
Rob
On 07/27/2015 09:19 AM, Colin Cotter wrote:
Dear all,
Just checking that people are aware of the regression in FIAT, that
one cannot specify pointwise derivatives as nodes anymore.
I discovered this when trying to implement a Taylor basis element in
1D. It still makes use of ScientificPython.Derivatives but the import
of ScientificPython on line 87 is commented out.
all the best
--Colin
--
http://www.imperial.ac.uk/people/colin.cotter
www.cambridge.org/9781107663916 <http://www.cambridge.org/9781107663916>
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics
