New derivation of equations, including weak form statement. mshr and spherical mesh only work on python 3.6 for now. Changing initial conditions, time step, grid size can all influence stability of algorithm.

Droplet can now have different size than one and still start at zero. 
Other minor changes prior to changing units.

MAJOR BUG FIXED. x domain now starts at zero. pdepe enforces no-flux for spherical coordinates at absolute zero, not at the beginning of the domain.

Adding peaked initial conditions to ternary FRAP. Also adding figure comparing time course with steady state.

Frank's solution to probability problem via Laplace transform. Weirdly only works for one side right now.

Matlab code for ternary model for unbleached molecules seems to work properly now. Problem before: volume fraction not defined properly and thus low mobility inside.

Deleting simple Cahn Hilliard example. First trial Flory Huggins with many terms (simple substitution of y=laplace(phi_tot) instead of the entire chemical potential) still doesn't work properply. For now continuing with amended Cahn Hilliard which accounts for Flory Huggins free energy. The other Flory Huggins code probably has a coding/fenics issue I don't see.

Adding new file with Cahn Hilliard extended to Flory Huggins binary mixture. This was used in email to Chris.

Amended Cahn Hilliard code in a minimal way to account for Flory Huggins binary mixture with the corresponding mobility scaling. Seems to work, but not sure about the volume fraction to concentration conversion.

2D/1D Flory Huggins works in principle. Currently output only via paraview, need to figure out plotting from fenics.