CountNook Methodology

CountNook treats a counter as a stateful workflow, not just a number and two buttons. The checks start with what an action means and end at the page a visitor can use.

Last updated:

How a CountNook behavior is established

Each layer answers a different question. Passing one layer does not silently stand in for the others.

  1. Define what each action changes

    Increment, correction, round completion, repeat movement and Clean or Retry attempts are represented as explicit actions. Limits and invalid inputs have defined outcomes rather than relying on visual state alone.

  2. Exercise repeatable examples

    Repeatable examples cover ordinary use and meaningful edges: signed scores and ties, repeat boundaries, target completion, undo, imports, invalid saved data and storage limits. A one-off success is not enough.

  3. Check browser saving separately

    Saving is checked separately from what appears on screen. Checks cover reopening saved work, recovering from an interrupted write, older saved formats, multiple tabs and protection against two tabs overwriting each other.

  4. Inspect the rendered and keyboard surface

    Page reviews cover labels, announcements, dialog structure, focus, keyboard use, narrow screens, recovery from failure and the information shown to search engines.

  5. Check external services separately

    CountNook can check how it handles Google Analytics, Turnstile and Brevo. That does not prove those services or Vercel protections are configured. Contact and Feedback keep their delivery-pending state until live setup is complete.

The same foundation does not make the tools interchangeable

The main workspace owns multiple independent tallies. Stitch & Row owns a project with stitch position, row or round history and optional repeat progress. Game Night owns participants, complete score rounds and a chosen high- or low-score direction. Music Practice owns one ordered stream of user-judged attempts and a consecutive-clean target. Each tool keeps its saved work separate, so one tool cannot accidentally open another tool’s data.

JSON and CSV are also checked by purpose. JSON is the restorable form for the three tools that expose it. CSV is a readable snapshot with columns suited to each tool; it is not accepted as an import. Music Practice exposes neither format. These facts are kept consistent across controls, help copy, page descriptions and the published limitations.

Automation is evidence, not a replacement for judgment

Automated checks

Automated checks cover the working tools, saved data, downloads, forms, page links and accessibility basics. Every required page and example must be present; a missing surface does not count as a pass.

Known-error check

When a check changes, it is first tried against a known bad case and must catch it. The bad case is removed before the final run. This confirms that the check still detects the problem it is meant to find.

Human review

Visual hierarchy, comprehension and overall readiness still require human judgment. Automated checks do not amount to owner approval.