Reach Is A Node In A Mesh Rather Than A Rung On A Ladder, Questions Divide Into Ones That Identify And Ones That Measure, And Each Carries A Reliability Because A Wrong Answer Is Not Noise But The Finding
Summary
An evolution of the guessing game on one correction and one implication. The correction: a capability's reach is not a rung on a ladder but a node in a mesh, because reading a file at the local disk, the user's home, a corporate share, a network mount, a cloud store and a container's own layer are six exposures wearing one verb, none above or below another; each reach node carries its own edges to environments, products and obligations, which makes the traversal bidirectional, and walking inward answers which products touch my share. The implication: a player asked whether the thing can reach every credential on their machine may say no because they do not know, so every answer is already a prediction; questions divide into ones that identify and ones that measure, each carries a reliability, a low-reliability answer barely moves the belief and fully counts toward the gap, and the gap is a set of per-capability disagreements collected throughout. The detective panel shows asserted, inferred and possible in three treatments never mixed. Built the same day: the mesh as files with one ontology, the engine, the two-column play screen, the mesh navigator, the report and the sources page.
Key concepts
- The play screen — the transcript and the inspector
- The mesh, start anywhere — which products touch my share?
- The report — one page, every class kept apart
- The sources of truth — one file per node; a correction is an edit
- The ontology — every edge typed
Key ideas
- Reach is a node in a mesh rather than a rung on a ladder, because six file-system reaches are elsewhere from each other.
- Coarse and fine views must be the same shape, or every boundary needs a special case.
- Questions divide into ones that identify and ones that measure, and only the second may contribute to the finding.
- A low-reliability answer barely moves the profile belief and fully counts toward the gap, which turns the user's error into the product.
- The inspector shows asserted, inferred and possible in three treatments that are never mixed.
- The most valuable guesses are things nobody calls an agent; an obligation reached this way is a question worth asking, never a finding.
- A model earns three edges and never sits in front of an arithmetic step.
On this site
Adds probes/mesh/ (ontology, reaches, envs, vendors, obligations, questions as one file each), refine maps and two non-agentic profiles, admin/build/gen_mesh.py with its gates, guess/engine.js, the rewritten play screen, guess/graph.html (the mesh navigator), guess/report.html and guess/data.html.