Discussion related to expanding the CCPP framework to do variable conversions

[grantfirl]

The SDFs consist of many interstitial schemes to couple the various parameterizations.

What about extending unit-conversions to do variable-conversions?
A tad bit more complicated, given the conversions will have dependencies.
Could greatly reduce interstitial code.
Less interstitial schemes => closer to “plug and play” physics.

Examples of common variable conversions.
Mass mixing-ratios ⇔ Volume mixing-ratios
Specific-humidity ⇔ Relative-humidity
Temperature ⇔ Potential Temperature
Temperature and Moisture ⇔ Virtual Temperature
Hydrometeor mixing-ratios ⇔ Water paths