You might also search for "eccentricity".  I'm not sure how it's calculated (maybe like your second example?), but it measures how "elliptical" a round figure is.  I believe the method behind it is that all circles look like ellipses when viewed from an angle.  Therefore looking directly down on a true circle would give an eccentricity of 1, and looking at the side of a circle would give an eccentricity of 0 (a straight line segment).

Thanks for your insight, however I don't know how the second example was calculated. The diameter difference is 0.21, not 0.19.

I believe my first definition is a weighted average whereas the second description was an attempt at expressing ovality as the difference between major and minor diameters.  I have not come across ovality expressed as a geometric tolerance and was wondering why runout would not do.  For instance pistons are machined using ovality tolerances:   major diameter minus (somethingA) +/- (somethingB).  To my surprise pistons are machined as ovals, (how this is done, I would appreciate if someone could comment). This is so that differing thermal expansions of the piston mass can assist in sealing. In general a piston's cross-section is different at right angles due to the wrist-pin. And, an oval is basically an ellipse, or a circle with all the runout at 90 degrees.  Any comments would be appreciated.

What I was thinking is something more along the lines of: (D-d)/D

The result would be between (0,1).  My last post reversed the numbers, so 0 would be a perfect circle and 1 would be a line segment.