herein lies the problems of infrastructure. infrastructure has no layers of abstraction. things go all the way deep - nuts and bolts matter, galvanization processes matter, the temperature at which mortar sets matters, documentation matters. the flat ontology par excellence is the material world, hey, in which there are no theories here, not even models, just the ruthless ping-pong game of one-to-one physics in which if it works, it works, and if it doesn't, it doesn't. there's no undo.
in that setting it is fascinating, it is fascinating, to build (or to point fingers and to watch others build) and to see a space slowly grow. it is also fascinating to realize that infrastructure is boring precisely for the reasons it is interesting. infrastructure and buildings and spaces are slow. they are heavy. there are no layers of abstraction to pile layers of abstraction. no layers of reliability.
(well, there are some. the screw is reliable. you hang your life on nails and joist hangers. you hang your life on engineered wood structures and pre-tested fire assemblies, and margins of error and safety factors that have become baked into objects and their assembly documents alike. you trust in the solidity of a wireless access point, the repeatability of a switch.)
is this boring? is this fascinating? all of the above. it is slow, though, and so easily ignored. software infrastructure is so fast, so blindingly exciting. it seems easier and harder to reconcile architecture and technology, and software seems more dangerous, more seductive in its coherence, platforms in which if it works, it works.