And see if you can do a better job at it than the existing systems do — which is quite possible. Of course, actually having your scheme implemented would be a perhaps unrealistic expectation, but you can never quite be sure what people will do in the future.
My idea is a logarithmic system that works in a similar way to AWG, but with the cross-sectional area aligned to "round" numbers in metric. I would set the precision at 20 steps per decade (of cross-sectional area), though I would expect all but the most demanding applications to use increments of either 2 or 3 at a time. The reference point, 1mm², would be defined as size 0, while smaller wires would have negative numbers and larger ones positive.
Oh, you want an actual name for it? Try "Future Wire Gauge" (FWG). For mains cables under this system, I might try the following sequence:
Size numbers — –6, –3, 0, +3, +6, +8, +10, +12, +14, +16, +18, +20, +22, +24, +26, +28, +30, +32, +34…
CSA (mm² rounded) — 0.5, 0.7, 1.0, 1.4, 2.0, 2.5, 3.2, 4.0, 5.0, 6.3, 7.9, 10, 13, 16, 20, 25, 32, 40, 50…
Approximate free-air ampacity* — 6.0, 7.5, 10, 13, 17, 20, 24, 28, 33, 40, 47, 56, 67, 79, 95, 112, 133, 158, 188…
*Using the reference point of 10A for 1.0mm², and assuming that the ampacity co-relates to CSA^0.75 — which could be off slightly in practice, but is closer to reality than any other "simple" calculation. (For some reason not clear to me, 0.5mm² flexible cord is supposedly only good for 3A. The other thing about it, I believe, is that it is only available with light-duty insulation and sheathing.)