Don’t over-engineer; novelty is not a goal

A clever design costs more to build, to test and to keep working than a simple one, and it is only worth that cost when a measurement shows it is needed. Ideas are judged by how well they fit the project, not by whether anyone has done them before.

The user, 2026-10-10: “dont over engineer things”, and “the goal is not simply to build something that is not built elsewhere, that is a cool stunt, but good ideas even when built elsewhere are great if it fits our rules, visions and goals”.

• Prefer the smallest design that meets the requirement reliably. A simpler exact mechanism (for example deduplication keyed by the whole tree id) beats a clever fine-grained one (per-file read tracing) unless a measurement proves the clever one is needed.
• Judge ideas by fit and value, never by novelty.

How it is enforced: by design review. It is a working rule with no mechanical guard.