Notes from the deep end.

Long form pieces on the parts that fought back. Parsers, task queues, and benchmarks that argue with their own conclusions.

Arguing with one of these?

Good. Send the counter example, or the system you want measured the same way.