What does this MR do?

• Rewrite local operator so support multiple DG schemes & upwinding algorithms, and combinations thereof
• Add experimental config keys dg.experimental which alter these settings at runtime:
  • method, upwinding, and weights
• Change default upwinding scheme to none

Is there something that needs to be double checked?

Disabling the upwind scheme...

• Does not (significantly) change 1D ODE benchmark residuals for homogeneous media
• Improves 1D ODE benchmark residuals for layered media
• Reduces deviation to Muphi results in 1D transient benchmarks for homogeneous media
• Removes timeout issues of reference tests

Can this MR be accepted?

• Add new keys to parameters.xml
• Add CMake option -DEXPERIMENTAL_DG_FEATURES
• Add section on these features to README.md
• ODE tests pass
• Test vs Muphi passes

Related issues

Closes #43 (closed)