Last year I wrote an article that argues against using “search engine hits” as a proxy for “popularity.” My conclusion is that, statistically speaking, it’s not a good measure. Although I don’t address the Tiobe rankings, several comments bring them up. See http://blogs.sas.com/content/iml/2011/08/19/estimating-popularity-based-on-google-searches-why-its-a-bad-idea/