Should we remove species properties (ex. diffusivity) from Phase object?

[ma-sadeghi]

I think we should remove species properties such as pore.diffusivity from Phase objects. The reason is that it's very ambiguous. Of course, for binary mixtures, D12 = D21, but still, it's ambiguous for Phase object to have such property. It might make sense for a BinaryMixture class to have pore.diffusivity as a compromise.

I propose the following:

• Remove species properties from Phase objects altogether, and by default use Mixture class for mass transport problems.
• Depracate the use of pore.propname for species properties altogether, and use pore.propname.X instead where X is the species name.

PS. For backwards compatibility, we could still support pore.diffusivity, and eventually get rid of it in v4.

[jgostick]

I agree completely, but the solution results in the code being significantly more confusing to use, so we remove one source of confusion, but add another. This is why we're stuck with what we have.

Incidentally, the new 'Phase properties' functions and classes to support your proposed approach. But this would require everyone using mixtures, which is more complicated.

Anyway, definitely worth discussion, and I'd love to hear from active users their thoughts on this thread too.