Implementation status#

Document

Contents

Language specification

The DSL — every construct, with examples

Panel Core

The resolved intermediate format

Shot types

Normative camera framing table

Asset contract

What a character puppet must declare

This specification is free to implement. The project is released under 0BSD, which imposes no conditions whatsoever — but to be explicit: anyone may build a competing compiler, editor, renderer or tool for this language without permission or attribution. A notation is only worth having if it is not owned.

Implementation status#

Phase

Scope

State

0

Toolchain, CI, specification, project scaffolding

Done

1

IR, Panel Core schema, YAML frontend, puppets and forward kinematics

Done

2

Camera scale resolution and actor placement (Cassowary)

Done

3

Text metrics, balloon placement, reading order, tails

Done

4

SVG emitter, debug overlay and CLI

Done

5

Comic-script frontend; multi-panel over inheritance

Done

6

Browser playground (Pyodide); VS Code extension

Done

7

Machine-readable diagnostics: scenet check, in SARIF

Done

8

Captions, faces, and the setting layer — places, masses, planes, atmosphere

Done

9

Emanata — plewds, squeans, grawlixes, briffits — as a soft placement cost

Done

10

The agent-facing surface: spec pack and llms.txt, Agent Skill, MCP server

Done

11

Pages: tiers of panels at varying sizes, columns, insets, slanted tiers, one type size per page, Page Core

Done

12

Lettering: emphasis, kerning, type size, an author’s own face; sound effects in the display face

Done

13

Comic scripts as publishers write them, and in French; off-panel speech; small type; a script’s pages; notes in scenet check

Done

14

Accessible output: every panel an image a screen reader reads whole, in reading order; descriptions written, kept from a script or generated

Done

15

Supplied artwork: a picture behind the cast, linked or embedded; figures the picture draws, declared in an artwork contract and lettered around

Done

16

Right-to-left reading, for manga: panels, pages, strips and lettering read from the top right, the scene never mirrored

Done

17

Exports for image models: an OpenPose pose, boxes with phrases, a depth map, a mask and the semantics, from Panel Core

Done

Single panels compile end to end. Constructs described in language.md are the specification, not a report of what is implemented — the table above is authoritative on what actually runs.

Planned#

Each of these is scoped to be a release on its own. The order is deliberately not fixed, and the dependency between them is stated in the tickets rather than implied by this table.

Scope

Ticket

VS Code: scenet check findings, and highlighting for comic scripts

#77

scenet build --watch

#78

Print: a PDF emitter with trim, bleed and a black plate

#80

Still further out, and not yet ticketed: the interpretation layer that would give a panel a style.