This context defines the vocabulary for composing, updating, and rendering function plots in the next public API.

Scene: A user-owned composition of ordered visual marks plus shared plotting state. It is independent of a mounted DOM target and is mutable until rendered. Declarative changes create a replacement render rather than updating an existing output. A scene may contain no marks when only coordinate-system decorations or an empty plotting surface are needed.

Avoid: chart instance

Scene specification: The declarative composition and render inputs that describe what a plot should contain, including marks, sources, scales, and presentation options.

Avoid: view state, chart instance

Mark: A visual participant in a scene that turns plot state and source data into renderable geometry through the public extension contract.

Avoid: graph type, layer

View: A mounted presentation of a scene that connects it to a target, interaction state, and a renderer. It is not a declarative update handle.

Avoid: plot, chart

Render result: A detached SVG, Canvas, or other renderer output produced from a scene. The caller owns inserting, replacing, and removing it from the UI; interaction listeners and internal view state are initialized before the result is returned.

Avoid: mounted chart, view update

View state: Runtime state of a mounted presentation, including zoom, pan, pointer position, viewport dimensions, scales, and renderer metadata. View-state changes may trigger internal rerenders without changing the scene specification; pan and zoom can require marks to recompute their geometry. Pointer-driven helper behavior is also owned by view state rather than by mark-owned event subscriptions.

Avoid: scene specification, update spec

Coordinate state: The mutable runtime coordinate state owned by a View, including current domains after pan or zoom, instantiated scales, viewport dimensions, and derived metadata. It is initialized from a scene specification but is not written back into that specification. A new declarative render initializes fresh coordinate state; transferring it between views is an explicit advanced operation.

Avoid: scene domain, options cache

View-state snapshot: A serializable capture of portable interaction state, such as coordinate domains or zoom transforms, that can be explicitly applied to another render. It has a schema version and coordinate-system tag, and excludes DOM nodes, renderer instances, workers, compiled functions, and other runtime resources. Applying an incompatible snapshot is rejected rather than clamped or partially applied.

Avoid: scene specification, renderer cache

Group: A possible composite mark that owns nested marks when real use cases require nested composition. It is not currently a required general-purpose hierarchy.

Avoid: scene node

Source: Mathematical or static input consumed by one or more marks, such as an expression, point collection, parametric pair, or polar expression.

Avoid: datum, graph type

Coordinate system: The built-in plotting context that maps data coordinates to the viewport and owns scales, projections, inverse mapping, axes, ticks, grids, and other coordinate-aware decorations. Coordinate systems are internal implementations in the initial public API.

Avoid: axis mark, grid mark, chart options, public coordinate-system adapter

Prepared geometry: Renderer-neutral visual primitives produced by a mark for a particular view state, such as paths, points, rectangles, text, or arcs.

Avoid: SVG selection, render output