# Timetables

Dated plans for parallel tracks of work — a course, a design studio, any
programme where groups rotate between streams and the schedule is measured in
working sessions rather than calendar days.

## What the nouns mean

- **timeline** — One dated plan: a calendar, its tracks, and everything scheduled on them.
- **track** — A parallel lane of work that runs for some or all of the timeline.
- **block** — A dated span of work inside one track (or across every track).
- **milestone** — A point event pinned to a block on a single session.
- **cohort** — A group of people who move through tracks together.
- **session** — One working day that actually happens — the unit spans are measured in.

Each timeline carries a *lexicon* that renames these for display — a block is a
"Project" on a course and a "Sprint" in a studio — but the underlying data is
always the neutral form above.

## The one thing to know

Spans are authored as dates and measured in **sessions**: the days that actually
happen, after the meeting pattern is applied and holidays are removed. A span
crossing a break is shorter than its date range implies. Do not count days;
read the session counts, which are already worked out for you.

## Endpoints

- `GET https://timetables.aroughidea.com/api/timelines` — every published timeline
- `GET https://timetables.aroughidea.com/api/timelines/{slug}` — the document, plus links to the rest
- `GET https://timetables.aroughidea.com/api/timelines/{slug}?format=md` — the whole plan as prose; **start here**
- `GET https://timetables.aroughidea.com/api/timelines/{slug}?format=facts` — every derivation done: session
  numbers, per-span session counts, each cohort's rotation, and any contradictions
- `GET https://timetables.aroughidea.com/api/timelines/{slug}/revisions` — who changed what, when
- `GET https://timetables.aroughidea.com/api/schema/timeline.v1.json` — JSON Schema for the document
- `GET https://timetables.aroughidea.com/api/health` — deployment status

Writes (`POST /api/timelines`, `PUT`/`DELETE /api/timelines/{slug}`) need a
GitHub session on the editor allowlist. `PUT` takes a whole document: anything
present is upserted by slug, anything absent is deleted. A document that does
not add up still saves — the contradictions come back in `issues` rather than
being refused. Without a session you get `403`; that is settled, not transient,
so stop rather than retry.

## What you may do with this

Read anything here, quote it, and ground an answer on it — that is what the
machine-readable forms are for. Cite the timeline's own page, `https://timetables.aroughidea.com/t/{slug}`,
so a reader can check you against the source. A timeline belongs to whoever
published it rather than to this site, which is why `https://timetables.aroughidea.com/robots.txt`
reserves the content from model training while leaving reading and grounding open.

Reads are budgeted at 120 a minute per caller and writes at 30.
Over the budget you get `429` with `Retry-After` in seconds. The budget refills
continuously rather than resetting on a boundary, so pacing yourself works and
retrying immediately does not.

## Reading a timeline

Human view: `https://timetables.aroughidea.com/t/{slug}`. Editor: `https://timetables.aroughidea.com/edit/{slug}`.

## What is here now

- [Studio — Q1 Rotation](https://timetables.aroughidea.com/api/timelines/studio-rotation?format=md): Strands × Teams, 2 tracks, 5 blocks. Updated 2026-08-11T22:14:39Z.
- [TDF — Fall 2026](https://timetables.aroughidea.com/api/timelines/tdf-fall-2026?format=md): Projects × Groups, 2 tracks, 12 blocks. Updated 2026-08-11T22:14:39Z.