I'm firmly in this camp as well. A strong code quality metric is how easily understood something is - we write code in all sorts of states of minds at all hours

I'm firmly in this camp as well. A strong code quality metric is how easily understood something is - we write code in all sorts of states of minds at all hours of the day. If others can read your code and understand it, it means you can too when it's time to extend/modify it. This is also why things like Ruby's over reliance on metaprogramming bugs me - sometimes duplication is fine, and I'd much rather have some duplication than a wrong (or hard to discover) abstraction. The argument that "all code gets sloppy so let's just have very verbose code from the get-go" is pretty insane to me.