Structural bets
- [1] There is no drawn spine. The sketch panels themselves, stacked and joined by a short seam, ARE the line. This is the literal verbaltovisual vote nothing in v1 built.
- Sketches are large and central, the load-bearing element; text is demoted to annotation hanging off each sketch (alternating left/right).
- The hero question sits above the first sketch so the page still opens message-led, not image-led.
- Mark-and-return is woven INTO the sketch: you mark elements on the drawing itself (a boulder, the bulb), and marks live as annotations on that sketch, not in a separate tray.
- [3] Programmes become a row of small sketches on the same axis, keeping the "everything is a drawing on one line" logic intact.
- The final milestone is the form, introduced by the cable-to-bulb sketch: the invitation is drawn as the last station.
- On 390px, annotations reflow below each sketch and the sketch goes full-width, so the drawings never shrink to unreadable.
- Ground changes still mark section shifts; the sketches carry the narrative so grounds can stay very quiet.
Main risk: this shape is hostage to the artwork. Only two sketchnotes are locked-original; the rest are concept-only or absent. If the real drawings are not consistent in weight and framing, the spine visibly breaks, and the page has almost no text scaffold to fall back on. Highest ceiling, highest dependency on assets that do not yet exist.