Depending on how one chooses to print (either for maximum black or for high value tonal range as the primary beginning point), is probably the determining factor of which contrast extreme to begin with. I was taught and continue to expose first for the highlight values by doing a test strip at the soft contrast. When this exposure is determined then I make my next exposure first at the soft contrast time and then follow this with a test strip at the high contrast exposure on that second piece of paper. I let the high contrast exposure time be the determining factor for my low values. When I have the low values determined then I make my third print with the low contrast exposure first (at the previously determined time) and follow this with my high contrast exposure (at the previously determined time. From that resulting print, I then make determinations of burning and dodging.

I admire people who can make arbitary determinations of ratios of low contrast to high contrast exposures. I have found that in my experience that it just isn't that "cut and dried". Maybe I don't have my system developed well enough, or perhaps I want more out of a print.

I disagree that split developing is primarily used to salvage thin or overexposed negatives. In my experience excellent printers use this as one of several means to arrive at a fine print. It seems to work very well in that application apart from thin or overexposed negatives.

Interestingly enough my enlarger compensates for differing contrast grades insofar as exposure time. Perhaps not everyone has not been exposed to that technology yet.