What about that noise? This is pretty typical Perl thinking: pretend everything is ok, nothing to see here. Lots of CPAN commits, we have one metric to cling to. Everything is ok.
TIOBE's flaws are well-documented throughout the Internet (look out Ruby -- Delphi, RPG, and Logo are hot on your heels!). In particular, no third party has ever been able to duplicate their results, whereas anyone with access to any of the hundreds of CPAN mirrors can verify those numbers.
Generally I take everything I read with a large pinch of salt unless all the facts are presented and can be proven so don't believe either of these are fact especially the TIOBE one!