I wonder if it makes sense to produce some sort of neo-COBAL that produces a bridge between flow and declarative syntax that better aligns with expressing intent that maps from human input. That is harder to do with lower-level languages, but if we are optimizing for gluing together components it seems like it is very common for systems to build their own "DSLs" to solve a task. Is there a more generalizable DSL (which would no longer be domain-specific, I suppose) but still maintains the properties of such a system without generalizing to any kind of programming task?