There are three primary problems with mirrors:
1) bending deformation(compression of mirror front surface, stretching of the back surface).
2) shear deformation(adjescent parts of the glass move parallel to the direction of force).
3) thermal expansion.
Im working on an idea , this is incredibly simple and i cant go into all the details yet.
I think the 6:1 rule is total B.S.(allmost total actually), i think something like a 16" diameter/.25"(thats right .25! inch thick) will work, and not only that, will perform better than a 16" X 2.66" thick mirror ever could, and here's why:
1) the 16 X 2.66 has alot more thermal mass than a 16 X .25
2)the 16 X .25 if supported on a 27 point floatation cell and each of the nine 3-point supports are spaced a distance greater than .25" this will negate any shear deformation.
3) typical mirror cells are constructed of materials such as aluminum,wood,plastic, why support a low expansion mirror with a high expansion material?, no matter how the mirror is attached to the cell(sling/side support aluminum clips,etc,when the high expansion cell expands the mirror expands as well!.
4)if this thin mirror is slumped over a convex tool the front surface being concave and the back convex, a cell made of low expansion material as well as every support point, having a concave face on one side and flat on the other(the convex back of mirror sits in the concave front side of low expansion cell with 9-3 point supports interposed between).
you dont need a thick mirror, in fact alot of the heat distortion you see which is attributed alot of times to the atmosphere is actually the ridiculously thick mirror you have, never fully thermally equillibrating.