Of course I was immediately reminded of this wise remark by Marcel Minnaert:
If one does not know how a phenomenon should be measured, one can always start with a random, qualitative scale measure for it.
For a star that does not sparkle I use the number "0". The strongest sparkle that I ever saw - close to the horizon - I call "10". Then I distinguish the intermediate values by the other digits.
It is remarkable how useful many of these provisional scales turned out to be in the natural sciences. Much quicker than one would expect it, one gets used to the significance of each number of the scale, and there will soon come a moment when someone finds a way to calibrate the qualitative scale and turn it into a well defined quantitative measure.
Marcel Minnaert, 41. Hoe meet ik het fonkelen der sterren? in De natuurkunde van 't vrije veld. Deel I (1937), Licht en kleur in het landschap [Physics of the natural surroundings, Light and colour in the landscape]