[less_less]

True, but also the work can be reduced significantly with better tooling, which is still being developed but has improved markedly over the past decade. Eg SMT solvers that can output proofs, or tactics in Coq or Lean.

I'm hoping that this will be a big application of AI actually. If an AI can be built do to this simple but very tedious work, and your verification tool is capable of catching any errors it makes, then you've covered up a major flaw of formal verification (its tediousness) and of AI (its tendency to output bullshit).

[ulrikrasmussen]

Proving safety is just a small part of the problem to be solved. The hard part is actually structuring the program such that its correctness can even be formulated as a formal property which can be proved. For a lot of software that alone is highly nontrivial.

[recursivedoubts]

From "No Silver Bullet":

> More seriously, even perfect program verification can only establish that a program meets its specification. The hardest part of the software task is arriving at a complete and consistent specification, and much of the essence of building a program is in fact the debugging of the specification

[986aignan]

On the other hand, proofs sometimes give you more than you'd expect. A proof that the implementation of a function always returns some value automatically proves that there's no arbitrary code execution, for instance.