{
  "authorship": "Five self-authored fictional cases, not vendor results or actual observations.",
  "cases": [
    {
      "id": 1,
      "text": " 1. Play receives focus but playback remains unverified  Fictional setup: Mira has a finished lyric video on a draft portfolio page. Her review note says that Tab reaches a triangular control. She is tempted to mark playback as working because the control has a visible outline. This example deliberately supplies no real player result.  Her proposed walkthrough separates finding the control from using it. She records the focused control label, tries its supported activation key, and checks whether the video actually begins. She then returns to that control and pauses. If activation scrolls the page instead, she records the scroll and the starting focus location rather than concluding that the video played.  The original decision rule is to keep two fields: reachable and operable. A positive entry in the first cannot fill the second. Her fictional handoff remains pending until both play and pause have an observed outcome on the delivery page. This catches a different problem from missing caption text: the viewer may never reach the intended viewing state at all. "
    },
    {
      "id": 2,
      "text": " 2. Captions exist but the menu path is incomplete  Fictional setup: Noah supplies an English caption file with a lyric video. A collaborator sees a captions icon in the player, so the delivery record initially lists captions as available. Nobody has yet completed a keyboard-only route through the menu.  Noah's proposed check begins with captions off. He reaches the menu control, opens it, identifies the current option, selects the English track and closes the menu using the interface's supported interaction. He then checks that the expected caption content appears during a known passage. Merely opening the menu would leave selection and return-to-playback unverified.  His worksheet has separate entries for menu access, track selection and return focus. If the menu closes but focus cannot be located, he preserves that uncertainty. The fictional decision is to withhold approval of this path until the whole sequence is repeatable. Having a valid caption file is useful evidence about an asset; it is not evidence about operating the destination's controls. "
    },
    {
      "id": 3,
      "text": " 3. A language choice needs a confirmed selected state  Fictional setup: Lena prepares two language tracks for a music teaching page. The intended task is to switch from one language to the other while keeping the same video revision. The example does not claim any real player supports multiple tracks.  Her proposed record names the starting track, the desired track, the keyboard path and the selected-state feedback. She checks both the menu label and a known caption passage after switching. If two options have indistinguishable labels, she records the ambiguity instead of guessing from their order. If the second track is absent, the result is unavailable for that destination, not a failed translation test.  The original handoff rule pairs a track label with a caption revision. That prevents a later reviewer from approving a familiar label while a different file sits behind it. Lena keeps the selection task distinct from whether the translated words are accurate. Keyboard operation can be reviewed without pretending that a language-quality review has happened. "
    },
    {
      "id": 4,
      "text": " 4. Seeking must change the intended playback position  Fictional setup: Omar wants a learner to return to a chorus without dragging a progress bar. His delivery requirements include seeking backward and forward. No actual timing measurements or browser outcomes are supplied in this case.  He proposes reaching the seek control, reading any available instructions and using its supported keyboard operation. His record includes the starting position, requested direction, resulting position and whether focus stayed identifiable. He does not prescribe an invented five-second increment. If an arrow key scrolls the page, he notes that behavior and verifies the focus location before trying an instructed interaction.  Omar's fictional decision rule approves only the observed direction and control path. A successful forward move does not silently approve backward movement, nor does seeking prove that captions remain synchronized afterward. He adds a separate follow-up observation at the destination passage. This gives the next reviewer a concrete action to repeat while leaving untested playback behavior explicitly pending. "
    },
    {
      "id": 5,
      "text": " 5. The viewer must continue beyond the player  Fictional setup: Priya places a transcript link immediately after a lyric-video embed. Her page review needs a complete journey from the introductory link, through the player, to that transcript. A player-only control checklist would miss the surrounding page journey.  Her proposed sequence starts before the embed, operates a necessary control, exits any open menu and continues forward. She records the next focused element after the player, then checks the reverse route. If focus repeatedly cycles within the player, she records the exact sequence and any documented exit instructions. She does not press random keys indefinitely or describe a mouse-assisted escape as keyboard completion.  The fictional handoff keeps the transcript reachable as its own task and leaves any suspected trap unresolved for investigation. It does not label the entire site compliant or noncompliant from this single exercise. The useful contribution is a boundary record: where focus entered, what was operated, and where focus went next. That boundary makes later checks reproducible after the embed or page layout changes. "
    }
  ]
}