I don't really agree with the article. I often break up my complex regexes into different variables that get re-assembled before passing it off to the parser.

I don't really agree with the article. I often break up my complex regexes into different variables that get re-assembled before passing it off to the parser. The thing is, with regex, maybe worst is better. It's the best we've got (for now), and I can't imagine parsing complex text without it. Manually parsing it (tokenize, etc) without regex is much harder.