I guess the point is to prove this novel attack vector. And they found interesting things that don't always show up in a buffer attack: there are certain restri

I guess the point is to prove this novel attack vector. And they found interesting things that don't always show up in a buffer attack: there are certain restrictions on patterns that can be viably encoded physically, parallel sequencing can chunk up the input in unpredictable ways, the input might be read backwards by the sequencer.