David Wells

I'll probably think of more later, but I can't remember if we banned quadrature rules with negative weights: if not that should raise a warning.

true, but its just a warning.

That is part of the change in this PR, though its not in the CMake build system right now. Presently examples are always set up (i.e., we unconditionally generate Makefiles...

Of course, if pieces are ready to merge (like this), then life is easier for everyone if we put them in once they are done.

@brennansprinkle Thanks for following up on this. I am sorry that this has been languishing in the mean time. I will put aside some time later this week to read...

We talked about this some more on slack. Since things on slack tend to get lost I'll summarize my thoughts on the discussion here. Fundamentally, we have read-write access to...

I'd have to see the scheme sketched out to be sure but the third one seems like it shouldn't be that hard to implement with a class inheriting from `IBFEMethod`...

> Hmmm, I guess actually maybe it is OK already, because each solver will have its own EquationSystems object. We cannot have multiple `EquationSystems` objects attached to a single `Mesh`...

Things may work by coincidence but it's likely to cause problems - for example consider the snippet ```cpp void EquationSystems::init () { const unsigned int n_sys = this->n_systems(); libmesh_assert_not_equal_to (n_sys,...

Before we start writing code I'd like to determine *how* we are going to get rid of the names in a compatible manner. A lot of code currently depends on...