You animate by making a bunch of these things, and then "snapping" them together, so they are joined end-to-end for example, and that becomes constantly-updated behaviour. Problem that I'd appreciate input on an idea for is how to visualize the "code"

I’ve been stuck thinking about this also which i think is an unfortunate artifact of code being portrayed as text for so long. It sounds like the Actor Model could be similar enough to the modularity you’re describing. Ironically i’ve always visualized that as single arrows (like bow and arrow) where the air the arrow flows through is actually the data being modified by the flight of the function through it