[KaeseEs]

Here's where the analogy falls apart - civil engineering does not provide for coordinated efforts among thousands of construction workers. Construction management does. Any large building project will involve a civil or structural engineering firm, which designs the building or structure, and a construction firm, which builds it.

The construction firm is your compiler toolchain. The benchmark to meet is how long it took to design the Empire State Building, not how long it took to put it together.

[ChristianMarks]

Where would construction management be without civil engineering? The analogy between construction firms and compiler toolchains is either laughable or serious publishable research--I've never seen it mentioned once in the civil engineering literature.

[jerf]

Civil engineers are self-confident and don't go comparing themselves to all the other engineering disciplines trying to convince themselves they are real engineers. Software engineering is prone to a significant minority of people who are not so self-confident.

Personally, I am. In my opinion, whenever someone tries to drag in a methodology from other engineering disciplines they do so with a blinding ignorance of "cost" side of the cost/benefit ratio (and often non-trivial ignorance on the "benefit" side too), and the real differences between software and physical things. If car engineers could smash a car into a wall, observe something they didn't like, tweak something and 30 seconds be smashing another car into a wall effectively for free, their design methodology would be different too. (And if you feel inclined to get into a really detailed argument about how it would affect them because you think I'm assuming something about exactly how it would be different, you're missing my point, which is simply that it would indeed be different. Especially towards the beginning phases of the design.)