As an outsider, I understand what wpietri argues, but I fail to understand yours --- even though I've read other branches of the discussion. Do you argue that wpietri's method still estimates tasks but the estimation is always around two weeks?