In the software industry we're always saying things like, “scheduling software is inherently difficult because it has never been written before, so it's science. It's not like the building industry, where everyone involved has done the same thing 100 times before and it's possible to make good reliable schedules. The software industry needs to become more like the mature trades with predictable schedules and budgets.”

Well, what I've learned from my first large construction project is that this is hogwash