This actually sounds like a moderately good idea. But we should go about it in a slightly more scientific way.
I would be willing to bet that with the combined efforts of some list members, we can produce a "box" that contains the stuff needed and instructions on how to accurately measure the voltage of a particular coil. Pass the box around, and in a few months we should have tons of data. Data which could be correlated based on coil input power, physical characteristics of coils, frequency of operation etc. Match that data up with a maximum spark length each particular coil could produce. and we should now have a pretty close model of voltage to spark length characteristics . maybe some fancy graphs to show the world.
This would require team work, I for one would be up to participate. =)
I mentioned this earlier, but i don't see why a huge voltage divider would not work, since we all seem to be interested in spark lengths relation to voltage.
tons of resistors, in a wax or epoxy filled pvc pipe, mini toroid on top, a little tripod to set it up. Ground one end let the farthest spark hit the other, rectify and filter the last resistor and measure the peak voltage. I have done something similar before with my mini coil a long time ago, way long ago, before i really even knew how to make an actual coil.