Arising from int-0001 — The why of a decision survives the session that made it
Long decisions use a summary/detail split; every refusal stays visible with its grounds
Every alternative stays visible, struck through, **with a one-line ground beneath it**. Full reasoning expands per item via `<details>`/`<summary>` — no JavaScript, the same collapse mechanism the chain adopted.
The distinction that matters, and the reason this beat progressive disclosure: the summary line carries **the grounds, not just the title**. A reader scans all twenty refusals *and why each lost* without a single click. Expansion is for the full argument, not for the fact that an argument exists.
Chosen from rendered alternatives rather than argued in the abstract. Measured page heights: 6.9 screens untreated, 3.7 progressive, 3.8 for this. The near-tie between the last two is why the deciding factor was which device survived, not which was shortest.
The chain collapses in the same idiom — `… 16 further refusals on record` — so the instrument half and the prose half now use one disclosure grammar instead of two.
Alternatives on record — 2
✗Let it run — no truncation, trust the measure ceilingsrefused Honest to the record and it never hides reasoning, but a 20-alternative decision measured 5,316px, about 6.9 screens, for one entry.
On the surface a stranger lands on, that is a scroll nobody finishes, and reasoning nobody reaches is not meaningfully published.
revisit ifa reader study shows people do scroll it
✗Progressive disclosure — show the first N, hide the rest behind detailsrefused Halves the page, but it hides the thing that makes the page persuasive.
A reviewer testing this design as a stranger said the struck-through refusal with its reasoning still attached was the only element that made them want the tool. Putting most of them behind a click optimises the page against its own strongest device.