Java Conference: Soundly choosing language features

After a philosophical discussion with my brothers on programming languages, I posted my thoughts on how it is about time to add a little more science into the art of designing programming languages. Instead of relying on the individual tastes of Wirth, Gosling, Wall, or these days von Rossum or Matsumoto, we need languages that help less IT projects to fail. And we need objective metrics of real programmers for evaluating and quantifying our hypotheses on what works.