p_mp vs. pmp

[toddkarin]

I was working on some modeling with pvlib, and it uses a relatively consistent naming scheme of the form p_mp, v_oc throughout. For pvterms, we decided not to use one-letter abbreviations from @bmeyers suggestion. Instead we use variables like pmp, and pmp_ref.

I just want to double-confirm this important decision, since it seems like our new scheme is somewhat better, but will make integration with pvlib much more difficult. @cwhanse , @mdeceglie , @kanderso-nrel what do you think?

[cwhanse]

There's a lot of user inertia to consider, and personally I don't see that replacing p_mp with pmp is needed for pvlib.

It is important to distinguish between p_mp as a time-dependent output, and p_mp_ref or similar as a fixed parameter as an input. I think this quantity appears in pvlib as pdc0. Why pdc0? Because that's the symbol used by PVWatts. pvlib tries to maintain notation from the references, although that's not a hard rule.

[toddkarin]

Yes, this change seems like a big one and it might not be worth a such a disruption just to have a slightly more elegant code. Do you imagine the variables being of the form p_mp, p_mp_ref, gamma_pmp? Do we want to create single-letter lower-case abbreviations for the similar quantities: i_mp, e_dc (energy?), v_dc, p_ac_a? Or do we want a mixed bag, where we write out some of the quantities in full, e.g. power_real, v_mp, voltage_ac, power_inv_ac_max?

I'm not going to worry too much about the notation from references, since this seems like it would be necessarily outside of a unified scheme.

[mdeceglie]

I'm in favor of pmp, voc, isc. I agree with @cwhanse about user inertia. For example in RdTools, we aligned things where it was a clear improvement over existing names, but in other cases decide to stick with our existing naming.

[toddkarin]

All right, which way should we go for pv-terms? At least new software will start following a convention. I would have a slight preference for p_mp only because of greater automatic compatibility with pvlib.