rem’s Twitter Archive—№ 51,130

replete —we timed the DOM parsing time and reported back (in 2005), but more to optimise the entire payload, rather than edge cases.