The multi lines are quite useful, I made something similar in a spreadsheet. It helps to see illogical values sometimes better than the 3D graph.
And it is good to see where parts of the maps are linear; that means you can spread out the breakpoints wider since an interpolation will give the correct value anyway. Of course for that you must check both injection and ignition timing maps to see if the linear behavior is in the same ranges. So far it seems to work nicely in the upper rev ranges.
I also make the multi line graphs in the other plane, that helps to see funny dips and crossings as well.
As for optimal breakpoints: I can imagine a software routine that analyses one or more maps and determines the amount of linearity and steps between current cells. Then it would be possible to rearrange the breakpoints such that they concentrate around the areas with the 'wildest' behavior. Of course that would always be a compromise and it should find a sort of overall best fit within a certain bandwith. You can see it as a wire frame where you position all 16+15= 31 lines of nodes so that you get the best fit over the mapped surface, with least total stress in all the connecting edges. A max entropy algorithm would do that I think.
Hmmm.... From chicken wire forming to quantum mechanics in two small steps.
It's probably complete overkill, but if it's possible, why not
cheers
Tom