The discussion rail — <muse-discussion> on a published docv1.1 · ZEN REVISION

v1.1 — zen revision. Design mockup for the discussion rail mounted beside a published document. At rest the rail is almost nothing: breath marks in the margin, and a thread blooms only when you rest on its mark. Engaged, the host page owns the prose and injects the highlight marks; the sealed rail opens with one composed sentence — the weather of the discussion — never a dashboard of counts. Marks degrade honestly as the text changes: solid gold while the quote matches, dashed when re-anchored fuzzily, and a rewritten passage moves its thread to Detached instead of deleting it. The muse is a participant you mention — @muse — never a feature you operate. Copy at Zen depth throughout — it says "your team", never anything colder.

AT REST — THE DEFAULT STATE
Workshop — published · candle-kit-launch-guide.wdoc

Candle Kit — Launch Guide

Published · last edited yesterday · David & Mara · 5 min read

The candle kit is the first thing we ship that lives entirely outside the studio. It gives a first-time maker everything needed to pour two 220 ml candles at their own kitchen table, with the same wax and the same fragrance blends we use in the Tuesday workshops. The kit is the workshop, folded into a box.

Inside the box: 600 g of natural soy wax, two braided cotton wicks with steel sustainers, a brass pouring spout that clips onto any saucepan, 30 ml of our signature cedar-and-bergamot fragrance, and a fold-out instruction card illustrated by Mara. Every component is packed in paper — there is no plastic anywhere in the kit. a brass pouring spout that clips onto any saucepan David — Are we still committed to the brass spout? The quote was six weeks back in March… ✦ your muse replied · 2d

The first pour is the whole product. If someone ruins their first candle, they don't blame themselves — they blame the kit. The instruction card therefore walks the maker through temperature in three plain checkpoints — melt, cool, pour — with no thermometer jargon. Test kitchens told us the checkpoint language cut failed pours to almost none.

Curing is where patience pays. Let each candle cure for a full fourteen days before its first burn — soy sets slowly, and the fragrance needs that time to bind. The card says this plainly, and the box lid repeats it, because the single most common complaint in testing was a candle burned on day two.

We launch on the first Saturday of September, in the shop and online the same morning. Kits from the first pressing are numbered by hand. The launch price sits alongside the workshop ticket rather than under it — the kit is a workshop you keep, not a discount on one.

The rail at rest — breath marks only; a thread blooms when you rest on its mark; the full rail (below) is the engaged state.

ENGAGED — THE FULL RAIL
Workshop — published · candle-kit-launch-guide.wdoc

Candle Kit — Launch Guide

Published · last edited yesterday · David & Mara · 5 min read

The candle kit is the first thing we ship that lives entirely outside the studio. It gives a first-time maker everything needed to pour two 220 ml candles at their own kitchen table, with the same wax and the same fragrance blends we use in the Tuesday workshops. The kit is the workshop, folded into a box.

Inside the box: 600 g of natural soy wax, two braided cotton wicks with steel sustainers, a brass pouring spout that clips onto any saucepan1, 30 ml of our signature cedar-and-bergamot fragrance, and a fold-out instruction card illustrated by Mara. Every component is packed in paper — there is no plastic anywhere in the kit.

The first pour is the whole product. If someone ruins their first candle, they don't blame themselves — they blame the kit. The instruction card therefore walks the maker through temperature in three plain checkpoints — melt, cool, pour — with no thermometer jargon.2 Test kitchens told us the checkpoint language cut failed pours to almost none.

Curing is where patience pays. Let each candle cure for a full fourteen days before its first burn3 — soy sets slowly, and the fragrance needs that time to bind. CommentThe card says this plainly, and the box lid repeats it, because the single most common complaint in testing was a candle burned on day two.

We launch on the first Saturday of September, in the shop and online the same morning. Kits from the first pressing are numbered by hand. The launch price sits alongside the workshop ticket rather than under it — the kit is a workshop you keep, not a discount on one.

ANCHOR MODEL — THE QUOTE LADDER

Every thread anchors as a text quote + prefix/suffix context. On load the rail re-anchors each thread down a ladder: exact → context-scored → normalized → fuzzy (≥ 0.75 similarity) → detached. Fuzzy shows the dashed mark and the "text changed" whisper; detached keeps the original quote as the record. A thread is never deleted by an edit.

SEALED RAIL, LIGHT-DOM MARKS

The rail is a sealed Shadow-DOM component — host styles can't leak in and rail styles can't leak out. The highlight marks and breath marks are the one deliberate exception: light-DOM injections into the host's own prose, so the document's typography carries them and selection behaves natively.

STORAGE — THE HOST OWNS THE FILE

Comments are Brainy entities in an embedded PortableGraph; the whole discussion serializes as a .wthread file that the host persists alongside the document. The rail edits the graph in memory and hands it back — it never owns storage, credentials, or the network.

THE MUSE IS A PARTICIPANT

No labeled AI buttons anywhere. The reply row carries a bare and nothing else; you reach the muse by mentioning it — @muse — exactly the way you'd mention anyone on your team. It replies in-thread, under its own name, signed with its glyph.

THE WEATHER HEADER

The rail opens with one composed sentence — the weather of the discussion — built from deterministic narration templates, never an LLM. The tiny counts line beneath it is Normal depth; the sentence alone is Zen. Counts never headline.

SETTLEMENT

A settled thread collapses to one pinned calm sentence at its anchor — the outcome, not the transcript. Settlements distill into a decision memory via the know↔agree loop, so the next conversation starts from what was settled rather than from scratch.