Shouldn't this line:
> "using regular expressions clocks in as taking about twice as long as the hashmap-based solution."
be written as the opposite? That the RE takes half as long/is twice as fast?
I think the point is that naive regex are a very generic purpose tool, but it's still in the same ballpark. Having a custom optimized state machine for this specific use case could bring another 5x improvement on top, leading to 2.5x faster, potentially.