[T-hawk]

> Software cannot be engineered like a bridge.

It can be, as in the world of formal proofs and languages like Ada which facilitate them. It just typically isn't, as such formal procedures are far more costly and slower than than just dealing with bugs as they happen. For critical software systems like say avionics where a failure case is extremely costly (in both money and casualties), there do exist certification programs and rigorous methods as you describe for structural engineering.

It could be argued that avionics behaves more as engineering than as programming. Then we have a self-fulfilling definition: if we define fields with rigorous procedures as engineering, and fields without as software development, then of course software development doesn't have formalized methodology.

You're also comparing the largest scope of structural engineering - a massive bridge - with the entire spectrum of software. Engineering does happen on a smaller informalized scale too, like a homeowner building a shed or a Boy Scout troop building a rope obstacle course. That's still engineering, applying skills in carpentry or knot-tying to build something.