A bug problem I see with the methodology here is that _new_ SO questions aren't good measures of use - SO is very aggressive about duplicate questions, and for

A bug problem I see with the methodology here is that _new_ SO questions aren't good measures of use - SO is very aggressive about duplicate questions, and for anything as popular as Rails the common questions have already been asked. Anecdotally, my company uses Rails for new client projects all the time, so I don't think it has fallen as far as the author claims.