Is it possible to implement a dG method on a surface mesh? (not HDG). This requires computing the jump in the co-normal derivative of a Lagrange finite element function across a mesh edge. Not only that, but one needs the product of the jump in the Trial function * jump in the Test function.
It is clear how to do this on a flat domain from the examples. And I see the HDG on a surface example, but that does not have a product of jumps of Test and Trial. Is there an example of this somewhere?
-Shawn