Hey look, Mitteneague ran his in Firefox. Don’t you just LOVE bugzilla 915?
That said, your string add + no <? ?> result is a bit shocking… I’ve never seen those report faster than using comma delimits before – at the same time it’s also funny how the same thing with the opening/closing PHP is inline with the results I’d expect.
Though I’ve noticed some strange variances across hardware… Hardly surprising – JHVH knows what affect different CPU and disk caches could have on this.
Just curious, I’ve mostly tested this on Intels, your numbers are more inline with the one AMD I have – what were the system specs you tested that on?
The difference is there – common sense says the difference should be there – it’s really no more or less characters either way… so why use the slower method at all?
@Jeff – You complain I’m saying it’s stupid to choose the other way once you know better – what would you call intentionally choosing a method that runs slower with a higher memory footprint that’s no more or less work to implement?
Oh, and I have run multiple tests in the past on this – pretty much every time this comes up, so don’t pull the routine of making up claims about me and my conclusions! Yes, I’ve run tests in the past, YES, that’s how I came to this conclusion. People complain I’m insulting just by my very nature – well what in the blue blazes would you call your posts making insinuations and accusations?
Maybe you’re the one who should be running tests and possibly understanding how to test BEFORE running your mouth?