That was me, shooting my mouth off. Glad to know the situation has improved. Still, I think I couldn't be bothered with anything less than a Haskell front-end for it.
It's possible to use anything as a front-end to Csound. I used Perl + Csound for many years since the early 1990s and that's still a powerful combo IMO. You can easily use sclang to generate Csound scores. It's not much different than using the Score/NRT paradigm to write your scsynth directives as a list of OSC messages. And then you gain access to Csound's massive Ugen library.