Alright, but one of the central findings of software engineering of the 80s is that software engineering cannot be driven rigidly, unlike other engineering fields.. Such as railways.
Think about it, rail is an eminently standardized piece of infrastructure that has existed for more than a century and millions of kilometers have been laid out. Don't you think _some_ effort has gone into automating the process and making it predictable?
Don't you think _some_ effort has gone into automating the process and making it predictable? Here is a (french) example
This really underscores Klein and Thompson's argument: this infrastructure sclerosis seems to be a uniquely American problem.
This review of Klein and Thompson's book sums it up pretty well:
Adding a kilometre of subway track in the United States now costs twice what it does in Japan or Canada, and six times what it does in Portugal; in the past fifty years, the inflation-adjusted cost of a mile of interstate highway has tripled
Think about it, rail is an eminently standardized piece of infrastructure that has existed for more than a century and millions of kilometers have been laid out. Don't you think _some_ effort has gone into automating the process and making it predictable?
Here is a (french) example of laying up to 2km of continuously-welded rail a day: https://www.youtube.com/watch?v=o97SB8c7Ezk