My opinion: the only rational way to handle 2 non-coincident loads is to in effect do 2 load calculations, one calculation with one load and another calculation with the other, and then use the larger. [Obviously alot of the calculation will be the same, so you can just pay attention to what the difference will be.]
220.60 has gotten a bit garbled lately, so the applicable version of it may be open to other interpretations that could result in a higher number than the above. Those interpretations do not represent what is possible in reality.
2017 NEC 220.60 was fairly simple: "Where it is unlikely that two or more noncoincident loads will be in use simultaneously, it shall be permissible to use only the largest load(s) that will be used at one time for calculating the total load of a feeder or service." The reason for the subsequent changes is some ambiguity in the term "largest." It needs to be understood to mean "resulting in the largest final load."
Say I'm doing a standard load calc (220.60 does not apply to a 220.82 load calc), and I tell you a house has an A/C system with an MCA of 30A, where the largest motor in the A/C is the compressor with an RLA of 20A, and there is also non-coincident electric heat that draws 26A. Then I ask which is the "largest" of the non-coincident loads, the A/C or the electric heat?
The answer is "we don't know until you provide some information about the other motor loads". If the largest other motor load has a 20A FLA, then the heat is the largest load. [Heat only, get 26A+25A = 51A. A/C only, get 30A + 20A = 50A.] Whereas if the largest other motor load has a 10A FLA, then the A/C is the largest load. [Heat only, get 26A + 12.5A = 38.5A. A/C only, get 30A + 10A = 40A.]
Cheers, Wayne