My point was not that we shouldn't work to produce even small efficiencies (which, yes, cascade into larger aggregate ones).

It was more wondering whether speed in markdown parsing is such a concern that this would merit a marquee 'selling' point.