> I looked into the eweek article and found, that it lacks certain
> statistical qualities, like the selection criteria and the response
> percentage. I do not consider the article itself useful at all.

I guess you have to buy the Evan's report which is being discussed
(http://www.evansdata.com/n2/surveys/db_toc_05_2.shtml). There you
will have all the needed information.

Personally I see no reason to say that the methodology is incorrect,
otherwise all those big guys mentioned there (like Microsoft) would
have complained already.

Roman