Ummm... how did this develop into a language war? So far as I can see, the survey is quite narrowly defined: which language do you primarily develop in? Some people will spread themselves very evenly, but most of us tend to use one language in particular. For the people on here, using Slickedit, if we assume no-one fibbed then the results are very very accurate
Over on that other site the criteria is different. Their survey is presumably very very accurate too, but it's a
different survey and measures something very different. You cannot compare that one to this one, and in particular you can't extrapolate from that one and suggest that this one is inaccurate - how many of the language mentions on all those websites were from SE users?
As an aside, I'm deeply sceptical about using the results of Google searches to prove anything. For all we know, 300million of those hits could be along the lines of "doesn't java suck!". But you'll never find out because Google won't show you. If you do a search for "Java" (does that automatically exclude coffee or what?) you get results 1-10 of "about 349,000,000". Unfortunately, the number you can actually look at is only 984:
http://www.google.com/custom?hl=en&cof=&q=java&start=990&sa=N&filter=0And I seriously doubt that the tiobe automated collator actually reads 1 of those 984, just to check.