
ALT 1. The correct architecture varies wildly on the lifespan of the implementation: if you know with absolute certainty something is going to live for ten days, your approach to building it is going to be wildly different than if you know with absolute certainty that it is going to live for ten years. 2. (Absolute certainty, of course, does not exist.) 3. People in large organizations invariably underestimate a line of code's lifespan; people in small organizations invariably overestimate. 4. A big part of maturing as an engineer is not just improving at your ability to context switch *between* correct architectures if given a presumed lifespan, but your ability to *guess* the presumed lifespan. 5. A tricky confounding factor: engineers are humans, and they care about not just *the lifespan of the implementation* but *the lifespan of their having to interact with the implementation*. 6. A sufficiently self-interested engineer is going to, in practice, care much less about the lifespan of























