Measuring Genetic Eccentricity from the Cecilian Standard

Newspapers keep printing the percentages of commonality of genome we other species share with the human ideal (?). I believe it goes something like:

chimpanzee 99.7
round worm 70
yeast 30

I know very little about how genotypical
structures and sequences are translated in nature into the physical features and behaviors of correlate phenotypes. I'm not really fearful of being mistaken for a bit of yeast, or even for a chimp, now that the scores are known. However, given all the possible parametric differences that must exist as reasonable candidates to be selected for the measuring of some recognizable similarity among different species' genomes, what sort of algorithm might I suppose they have selected to produce all these nice catchy numbers every time a news reporter calls them up? (No need to show your work.)