Martin Sandve Alnæs <[email protected]> writes:
>> Unless I'm completely missing the thread, the current definition treats
>> a contravariant vector as covariant (a covector).  This is definitely
>> not what was intended in the Unified DG paper, they just implicitly
>> stated the definition for covectors.
>>
>
> Yes, however we can't fix that without improving the type system, as you
> mentioned, which is a quite pervasive task that's not on anyones agenda.

Understood.  Personally, I prefer explicitness to overloading so would
prefer functions with different names for these things.  Then, if the
type system is improved later, all old correct code will continue to
work and incorrect code can be detected.

I was just being pedantic that the current behavior (applied to
contravariant vectors) is not the intended meaning in the paper, so the
rationale is not so much that the current behavior is from the paper as
that UFL does not distinguish vectors from covectors.

> So until then, the definition of jump in ufl is unambiguous and documented,
> although not quite up to par with the mathematical definition from the
> paper.

Attachment: pgpUbEIMtjHX4.pgp
Description: PGP signature

_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics

Reply via email to