The desktop browser statistics in the article value Internet Explorer higher than in other articles I have read such as this one: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers
Other stats count page hit totals of each Browser.
Then you have ones that mess around with not including all of the particular Browser's versions.
And there is no telling who's getting a world-wide sample, who's taking out some coutries, who's samples over-represent a country because their traffic collection base is there, etc.