Companion to Discoverability Engineering
“A Discoverability Score Measures Whether Finders Who Don’t Know Your Name Meet You” gives the argument and the short version of each tool. This companion holds the rest: the formulas, the examples in full, the skills, and the toolkit that computes the score. The Discoverability Score is a working method of Intellectual Frontiers Press. It is not an industry standard.
Method
-
The score is nine sub-scores and a rule against guessing: the formula for every dimension.
-
Question classes carry weights, and the benchmark is frozen before the first run: the ten classes, the rules against gaming, and the freeze.
-
A run keeps the conditions, the raw answer, and the normalized observation apart: what is captured, and the folder layout.
-
Evidence is rated on seven attributes, and the claim type decides which count: the rubric and the evidence chain.
-
Every intervention comes from an observed gap and names what would show it failed: the five classes and the spam rule.
-
A retest runs the same frozen benchmark and reports what got worse: the nine steps.
-
The report comes in two forms, and a section with no data says so: the executive page, the practitioner report, and the 30/60/90 plan.
Worked examples
-
The first example scores the author from public evidence and finds recognition without inclusion
-
The engineering firm is named correctly and absent from its own service lines
-
The consulting firm shows why you define the finder before you read a zero
-
A complete run on invented data shows the loop, a regression, and an intervention that did nothing
The toolkit
The download is a zip of the method’s data and pages. The commands that run it are archon score commands, which need only uv; they call no model and no network.
| Folder | What it holds |
|---|---|
|
Every weight and rule the score uses |
|
A schema for each input and for the scorecard |
|
A blank copy of every file |
|
The three worked examples and the fixture, with their raw, normalized, and analysis files and scorecards |
|
The method pages as AsciiDoc |
|
The three skills, with the nine prompts inside them |
From a case folder such as examples/fixture-end-to-end/, run:
archon benchmark check benchmarks/v1/questions.yaml
archon score check runs/2026-05-04-baseline
archon score generate runs/2026-05-04-baseline
archon report generate runs/2026-05-04-baseline
archon comparison generate runs/2026-08-03-retest --against runs/2026-05-04-baselinearchon check score validates and scores every example against its committed scorecard and its frozen benchmark.
Capture answers by hand or through an interface the engine’s maker permits. The toolkit contains no code that drives an engine’s website.