The mass-action dynamics analysis for e.g. Petri nets uses PolynomialSystem; older ODE semantics analyses, such as Lotka–Volterra and linear ODE, do not.
To make @kasbah 's work on #766 easily applicable to these older analyses, we should refactor them so that they too use PolynomialSystem.
