Isn't maintainability mostly about applying basic engineering principles e.g. separation of concerns, single responsibility, dependency inversion, open-closed etc..? If these topics are addressed in the normative representation, and correctly translated by an LLM into code (especially by slicing up the specifications into measurable outcomes to set the intended foundation), then future change e.g. maintainability becomes essentially easy as well, no?
Historically I know that the majority maintenance problems occur from slow continuous evolution of a system that it initially was never designed for. And the only way to address this was continuous system design.
Historically I know that the majority maintenance problems occur from slow continuous evolution of a system that it initially was never designed for. And the only way to address this was continuous system design.