I genuinely appreciate your explanation, and I'm sorry that it appears as if I rage-baited you into wasting your time answering my question, which in retrospect does look as if I asked "why does this exist [at all]?"

I don't even write Haskell actively, but I still wanted to understand. If I've learned anything about your work, it's that whatever experimentation you do is worth paying attention to, even if it may end up not going anywhere later.

The "JVM has no tail-call optimization" angle was one of the first questions that came to my mind, and TIL about your Cadenza trick - very interesting.

Another thing I'd be curious to explore is the polyglot interop. My guess is that with THC I can call Haskell code directly from Clojure, and it looks like it might even be possible already today. The other direction, I suppose, might be slightly more difficult, since Clojure is not a Truffle language.

Anyway, I'll try to find some time to dig deeper; this effort of yours is enormously intriguing.

ps. I just realized I misspelled your name in my prior comment. I'm not sure how that have happened, sorry about that. It's not like that I don't even know who you are, I have met you in person.