Architecture is doing the algebra, before doing the arithmetic.
A good (network) architect suffers from the topologist’s vision defect. He can’t tell a coffee cup from a doughnut.
Architecture is maximizing the invariances and minimizing the discontinuities.
A field cannot consider itself a science until it can progress beyond natural history; moving from describing what is, to positing principles or theories that make predictions and impose constraints. And it shouldn’t be just any theory; we need a theory that has the fewest assumptions and the greatest breadth: a theory with the fewest concepts, the fewest special cases, that includes the extremes as degenerate cases of a more encompassing ...