Basically a perfect description of the problem. Seems to me the default is obviously getting things wrong whether that be over or under engineering. How would

Basically a perfect description of the problem. Seems to me the default is obviously getting things wrong whether that be over or under engineering. How would one go about reliably making choices that strike exactly the right balance consistently? One thing I'm starting to really internalize is that to do so requires a deep understanding of software engineering as a domain, the domain and existing system, and the technical vectors in the business.