I realized how large a success of engineering/standardization/design patterns it is that lumped circuit model is usable for typical circuits only after I played with water-based fluidics (where I would consistently make choices that caused my devices to deviate greatly from models with similar assumptions).
@niconiconi
Which mental model? Lumped components, or water flow?
@robryk@qoto.org This mental model was found useful even in early VLSI design. Mead & Conway's book has a section about how to imagine a chip layout as a water bucket.