`HyperbolicSystem`: Simplify nodal source interface

[tamiko]

Let's get rid of some unused interface for nodal sources:

• We do not need a high-order and low-order split for nodal sources: the only candidate for PDEs that I can see might need one are equations with well balancing, and we have exactly one such equation (the shallow_water module) where we don't need it.
• Let's remove the time parameter t as well - if this is ever needed then we should make it part of the precomputation loop. Maybe for entertainment I will simply add such sources with a function parser to the scalar conservation equation module.
• Reword some variables and simplify some code in the HyperbolicModule.

[ejtovar]

@tamiko Shall we merge this?

[tamiko]

@ejtovar Yup.