This might have been the worst possible way you could perform this experiment. What if he accidentally drew over the same line twice? What if there are uneven darknesses of the filled-in letters? What if one pen has a lower flow of ink than another? What if he left a gap somewhere?<p>Why not just use calculus to determine the area inside each of the letters?
Now, What really needs to be done is to compare "Ink Efficiency" with "Legibility". I wonder if there are other factors which can be used to determine whether one font is "better" than another?<p>BTW, the ink efficiency is better determined using weighted average of distribution of letter frequency in language use over the alphabet and should be trivially determinable by parsing font files.<p>Sounds like an a valuable mini exploration project suitable for print media.<p>[EDIT]. Additionally, for display optimizations you need to maximize "Legibility", "Font Size" at various contrast ratios on screen. It would be interesting to be able to quantify an optimal "Font Face" and "Font Size" for various contrast ratios (screen power consumption) for mobile displays.
Here's something that I am using (on Ubuntu FYI) - <a href="http://www.ecofont.eu/ecofont_en.html" rel="nofollow">http://www.ecofont.eu/ecofont_en.html</a><p>This is the Spranq Eco Font . It's a TTF font, designed with holes (invisible to the naked eye) that saves printer ink ... they claim 20% lesser ink, I havent benchmarked my usage.<p>The ecofont Pro is a software (rather than a font) that punches holes into any font of your choice - dunno about its compatibility with Linux/OpenOffice, so havent tried it.
It's cool that somebody can actually draw different fonts by hand, but wouldn't it be fairly easy to make images of fonts at a large size and count the black pixels? That way you could get numbers.