squint VS twinspark-js

Have you considered https://github.com/squint-cljs/squint ?

Personally I couldn't let go of Clojure's other advantages but at least using the syntax would let you step off the syntax churn bus

Interesting, in the CLJS space we recently got two new libraries in this space:

https://github.com/squint-cljs/squint Which is a thin layer on top of JS

https://github.com/squint-cljs/cherry Which is closer to CLJS semantics and data structures but compiles to .mjs files without any advanced optimizations etc.

I've also experimented a bit with Solid. Here's a SolidJS demo in squint: https://github.com/squint-cljs/squint/blob/main/examples/solidjs/src/App.cljs It runs over here: https://squint-cljs.github.io/demos/squint/solidjs/ (The JS output is smaller than the favicon in that project!)

Maybe Squint https://github.com/squint-cljs/squint or Cherry might fit the bill. https://github.com/squint-cljs/cherry

Should you need a minimal sprinkle of JS, you can still generate this server side using squint (which works as a JVM library) as well, while shipping a pretty small JS library to your client.

Sounds like there's some overlap with squint there!

That's a valid criticism of htmx and there is a developer who created a conceptual fork of the idea eliminating this design decision:

https://github.com/kasta-ua/twinspark-js

I view the issue pragmatically: I think developers should consider LoB strongly when building systems and when using htmx, but I also recognize there are other design considerations (principally DRY). I can definitely see both sides of the argument, however, and would encourage people not to abuse attribute inheritance, just like I would encourage people not to abuse generics in OO languages.