Researched and Dropped Shipped
Two investigations closed on YAGNI grounds, one design that outran its team, and one sweep with no relationship to anything. Listed because a research section that hides these is not a record.
Graph analysis, dropped deliberately
Two files on graph clustering and 3D visualization were researched thoroughly and then closed with an explicit note: no pipeline code was written, the key findings were folded into reference material rather than scheduled as work, YAGNI invoked by name. One finding survives as genuinely useful — Leiden clustering is purely topology-based and only sees edge density, never node meaning, so injecting model-extracted semantic edges is what makes semantic similarity BECOME topology.
A planning team that outran its research
A pack on project-management disciplines fed a merged design document proposing seven discipline specialists. The shipped team has one, and it is the simpler pre-existing one. That gap is why projectteam ships an explicit "what is not wired" entry, marked in flight, rather than a roster it does not have.
Backend integration notes
A live inspection of an external API — roughly 364 routes and 453 schemas — mapping its storage primitives onto a future learning loop. Nothing in the engine calls it. Exploratory notes on a system that has not been built against, and labelled as such in the file itself. It is file five of the same six-file code-review programme that reshaped the review pipeline, which is the point rather than a contradiction: a body of research can be load-bearing overall and still contain a document nothing was ever built from.
And one that is simply unrelated
A search for MIT-licensed macOS applications sits in the same directory with no relationship to this project's architecture at all. It is there because the directory is a notes directory, not a curated one, and pretending otherwise would misrepresent how the material was actually produced.