News for hackers. Stuff that matters.
Re:Utterly Unpersuasive: Formal Methods and Law (2019) by mbrodersen on Thursday April 21, 2022 @08:07PM (#31116706)
A great read. I had the same experience proving (toy) software correct: the act of formally specifying what you want to prove and proving it makes you understand the problem and the solution much better. It’s like your brain gains super powers.
Re:Utterly Unpersuasive: Formal Methods and Law (2019) by User23 on Saturday April 23, 2022 @01:21PM (#31136208)
Even an informal process like TDD is a huge improvement in this space. You can’t write tests first without an at least implicit specification.

It is really wonderful though when the algorithm just falls out of the specification. Dijkstra’s derivation of smoothsort[1] is a nice example.

[1] https://www.cs.utexas.edu/users/EWD/transcriptions/EWD07xx/E...