Ultimately, any text processing toolchain that uses ReGex for parsing will inherently be difficult to extend. And slow. And perhaps even broken. The disconnect

Ultimately, any text processing toolchain that uses ReGex for parsing will inherently be difficult to extend. And slow. And perhaps even broken. The disconnect lies in the human-machine readability gap. The only way to do it right, I believe, is custom text formats, scanned via structured strings into FSM grammers.