I wonder if there's any way to conclusively rule out the placebo effect. Perhaps some way of using the unixy 'time' command to measure the loading time of apps. I doubt that this would be possible, but imagine that you could pass an argument to microb that would cause it to start up, load a page, display the page and then exit. That would allow a tester to run something like "time microb -load www.bleh.com -exitafter" at the command-line and get a time before and after the tweaks. I have no idea if anything like this is remotely plausible, but perhaps someone else knows more.