[mumblemumble]

Agreed. If odd scoping rules are the problem, we can possibly disqualify most lisps that have ever existed. Including all of the ones that John McCarthy had a hand in building.

For my part, I had similar "this doesn't feel quite like lisp" moments with both Hy and Clojure. And I ultimately realized that it's the mere presence of the base platform that gets me. With Racket, I'm used to feeling like I'm sealed off in my own lispy bubble. With Hy and Clojure, there are always bits of Python or Java (read: Not Lisp) hiding in the background. Many differences from other well-known lisps that I know are attributable to that base platform.

And that's a catch-22 situation. Objectively, they might be (and, IMO, are) good design decisions. But they're also covered in non-lisp cooties. I don't know why I should expect otherwise. Brains are weird.