And how is testing said browser on an emulator that reasonably accurately mimics said older hardware not a valid way of evaluating it? Haven't exactly made a point there, other then possibly raising a squishy argument that using the emulator somehow isn't in the "proper spirit" of the exercise.
(Emulation/Virtualization is an extraordinary useful tool for software development and QA. Need to see how what you're working on works on a machine with 1/4 the RAM and a different version of the target OS then your development box? Beats wasting power and space on a zoo full of actual hardware.)