That makes sense!

I was actually wondering whether its enough to batch Schnorr verify every N events to resolve performance issues on low-end hw. Strikes me as simpler (or are we doing this already?)