rgame 0.2.0 → 0.3.1

data/docs/api/assets.md

@@ -1,11 +1,14 @@
-# Sheets, atlases and maps
+# Assets: sheets, atlases, maps and sounds
 
-The classes between a file on disk and a draw call: a sprite sheet sliced into
-frames, a nine-slice panel stretched to any size, a UI atlas, a tile map, and
-the asset manager that loads and caches all of them.
+**The asset manager loads and caches every file a game uses**: images, sound
+samples, songs, sprite sheets, UI atlases, tile maps and plain text. This page
+covers the manager and the Ruby classes it builds from those files: a sprite sheet
+sliced into frames, a nine-slice panel stretched to any size, a UI atlas and a
+tile map. [Images](images.md) and [Audio](audio.md) cover the C-backed `Image`,
+`Sample` and `Song`.
 
-They are pure Ruby, but they live in `RGame::Core` because they hold images, and
-an image is a GPU handle. Game logic names them by id and never holds one — see
+**These classes are pure Ruby, but live in `RGame::Core` because they hold GPU or
+audio handles.** Game logic names assets by id and never holds one; see
 [Testing what a scene draws](drawing.md#testing-what-a-scene-draws).
 
 | Page section | Class |
@@ -16,25 +19,40 @@ an image is a GPU handle. Game logic names them by id and never holds one — se
 | [UI atlases](#ui-atlases) | `RGame::Core::UiAtlas` |
 | [Tile maps](#tile-maps) | `RGame::Core::TileMapRenderer` |
 
-*This page grows as the rest lands.*
-
 ## The asset manager
 
-The one place file-backed assets are loaded and cached. Every game has one, and
-does not build it — `app.assets` does, rooted at the app's `media_root:`:
+**The asset manager loads and caches every file-backed asset.** Every game has
+one, and none builds it. `app.assets` does, rooted at the app's `media_root:`:
 
 ```ruby
 app.assets.image('space.png')                # => RGame::Core::Image
-app.assets.sound('example 09/boom.ogg')      # => RGame::Core::Sample
-app.assets.song('example 09/theme.ogg')      # => RGame::Core::Song
-app.assets.sheet('example 09/player.json')   # => RGame::Core::SpriteSheet
+app.assets.sound('sounds/boom.ogg')          # => RGame::Core::Sample
+app.assets.song('music/theme.ogg')           # => RGame::Core::Song
+app.assets.sheet('sheets/player.json')       # => RGame::Core::SpriteSheet
 app.assets.ui_atlas('ui/ui_atlas.json')      # => RGame::Core::UiAtlas
 app.assets.read('data/levels.txt')           # => String
 ```
 
-Paths are relative to the media root; an absolute one is used as it stands. Two
-spellings of the same file — `'a/b.png'`, `'a/./b.png'`, the absolute form —
-are one cache entry, not three.
+Paths are relative to the media root; an absolute path is used as given. Every
+accessor returns the same object each time. A file requested twice is read,
+decoded and uploaded once. Several spellings of one file share one cache entry:
+`'a/b.png'`, `'a/./b.png'` and the absolute form.
+
+The manager gives a game one object that knows what is loaded. Setup code builds
+no paths by hand and constructs no images inline.
+
+### Listing what is there
+
+```ruby
+app.assets.glob('locales/**/*.yml')   # => ["locales/de.yml", "locales/en.yml"]
+```
+
+`glob(pattern)` returns the paths under the media root that match `pattern`,
+relative to the root and sorted. The order is therefore the same on every
+platform, whatever order the file system lists them in. A directory that does not
+exist matches nothing and returns `[]`. An absolute pattern is used as given, and
+its matches come back absolute. `glob` loads nothing and caches nothing; hand
+each path to an accessor to load it.
 
 ### Adding an asset type
 
@@ -43,21 +61,16 @@ app.assets.add_loader(:level) { |path| MyLevel.parse(File.read(path)) }
 app.assets.level('levels/one.json')          # cached and grouped like any other
 ```
 
-The built-in types go through the same mechanism at construction, so an added
-one is not a second-class citizen. It exists because some types cannot be built
-from inside `RGame::Core` at all — see [Tile maps](#tile-maps).
-
-Every path is **relative to the media root**, and every accessor returns the
-same object each time it is asked — so a file wanted twice is read, decoded and
-uploaded once. That is the point: loading stops being scattered across a game's
-setup, building paths ad hoc and constructing images inline, and becomes one
-object that knows what is loaded.
+The built-in leaf types (`image`, `sound`, `song` and `read`) register through the
+same method at construction, so an added type works exactly like them. `sheet` and
+`ui_atlas` are composites built from those; see below. `add_loader` exists because `RGame::Core`
+cannot build some types itself; see [Tile maps](#tile-maps).
 
 ### Groups, and what `release` frees
 
-Each cached asset remembers the **set of groups** that asked for it. An
-ungrouped load belongs to a permanent sentinel and survives every `release`; a
-grouped one is reference counted.
+**Each cached asset remembers the set of groups that asked for it.** An
+ungrouped load belongs to a permanent group and survives every `release`. A
+grouped load is reference counted.
 
 ```ruby
 app.assets.image('ui/buttons.png')                       # ungrouped: permanent
@@ -70,49 +83,47 @@ app.assets.release(:level1)   # drops lvl1/* unless another group still holds it
 app.assets.clear              # drops everything, permanent included
 ```
 
-An asset two levels both loaded survives until **both** release it, so two
-scenes can share a texture without either one pulling it out from under the
-other. A cache *hit* under a new group is tagged with it too — the alternative
-silently loses the second group's claim.
+An asset that two levels loaded stays until **both** release it. Two scenes can
+therefore share a texture safely. A cache *hit* under a new group also adds that
+group. Otherwise the second group's claim would be lost without a trace.
 
-Releasing drops this cache's reference. When the GPU texture actually goes is
-the collector's business; `Image.debug_live_textures` is there if you want to
-watch it happen.
+Releasing drops this cache's reference. The garbage collector decides when the
+GPU texture goes. Watch it with `Image.debug_live_textures`.
 
-`release` refuses the permanent sentinel by name, because releasing it would
-drop every ungrouped asset — the opposite of what "permanent" means. Use
-`clear`.
+`release` refuses the permanent group. Releasing it would drop every ungrouped
+asset, the opposite of "permanent". Use `clear` instead.
 
 ### Composites share their parts
 
-A sprite sheet is a descriptor plus an image, and **both are pulled through this
-same cache**. So these hand back one upload between them:
+**A sprite sheet is a descriptor plus an image, and the manager loads both
+through its own cache.** These two calls therefore share one upload:
 
 ```ruby
 sheet = app.assets.sheet('sheets/hero.json')   # names hero.png inside
 image = app.assets.image('sheets/hero.png')    # the same texture, not a second one
 ```
 
-The descriptor's image is resolved *next to the descriptor*, which is what lands
-it on the same cache key a standalone load would use. Release the sheet's group
-and its PNG goes with it.
+The manager resolves the descriptor's image *next to the descriptor*. That gives
+it the same cache key a standalone load would use. Releasing the sheet's group
+releases its PNG too.
 
-**One known gap.** A composite tags its parts with the group that first built
-it. If a *second* group later asks for the same already-cached composite, only
-the composite's own key is re-tagged, not its parts — so releasing the first
-group can drop a PNG the second still expects. Fine for the usual "each level
-owns its assets" pattern, and it would take per-part tracking to close.
+**One known gap remains.** A composite tags its parts with the group that first
+built it. When a *second* group requests the cached composite, only the
+composite's own key gains the new tag, not its parts. Releasing the first group
+can then drop a PNG the second group still expects. The usual pattern, where each
+level owns its assets, is unaffected. Closing the gap would need per-part
+tracking.
 
 ### Failure
 
-A loader's own error comes through unchanged — `Image::LoadError`,
-`Sample::LoadError`, `Errno::ENOENT` — naming the file. A load that failed
-leaves **nothing** behind: no cache entry and no group tag, so a retry is a
-clean retry rather than a half-registered asset that can never be released.
+A loader's own error passes through unchanged and names the file:
+`Image::LoadError`, `Sample::LoadError`, `Errno::ENOENT`. **A failed load leaves
+nothing behind**: no cache entry and no group tag. A retry starts clean, with no
+half-registered asset that can never be released.
 
 ### Testing without files
 
-Every asset type maps to a loader proc, and they are injectable:
+Every asset type maps to a loader proc, and you can inject your own:
 
 ```ruby
 assets = RGame::Core::AssetManager.new(
@@ -121,18 +132,17 @@ assets = RGame::Core::AssetManager.new(
 )
 ```
 
-The defaults name `Image` and `Audio` only *inside* their bodies, never at load
-time. That is deliberate: it means the caching, path resolution and grouping —
-which is all of the logic here — can be specced with no window, no GL context
-and no files at all.
+The default loaders name `Image` and `Audio` only *inside* their bodies, never at
+load time. Specs can therefore cover all the manager's logic with no window, no GL
+context and no files: caching, path resolution and grouping.
 
 ## Sprite sheets
 
-A sheet is one image plus a JSON descriptor, sliced into frames at load time and
-drawn one frame at a time.
+A sheet is one image plus a JSON descriptor. It slices the image into frames at
+load time and draws one frame at a time.
 
 ```ruby
-sheet = RGame::Core::SpriteSheet.load(app, 'media/hero.json')
+sheet = app.assets.sheet('hero.json')
 
 sheet.frame_width    # => 16
 sheet.grid           # => [rows, columns]
@@ -159,14 +169,14 @@ sheet.draw(renderer, row, col, x, y, flip_x: false, z: 0)
 }
 ```
 
-`image` is resolved **next to the descriptor**, so a sheet can be moved as a
-pair of files without editing either. `frame_width` and `frame_height` are the
-only required keys; a descriptor missing one raises `ArgumentError` naming it.
+The sheet resolves `image` **next to the descriptor**, so you can move both files
+together without editing either. Only `frame_width` and `frame_height` are
+required. A descriptor missing one raises `ArgumentError` naming the key.
 
 ### A frame can be smaller than its cell
 
-Cells sit on a fixed `cell_width` x `cell_height` grid. What gets *drawn* is a
-`frame_width` x `frame_height` rectangle offset by `origin_x` / `origin_y`
+Cells sit on a fixed `cell_width` x `cell_height` grid. The sheet *draws* a
+`frame_width` x `frame_height` rectangle, offset by `origin_x` / `origin_y`
 inside its cell:
 
 ```
@@ -178,58 +188,58 @@ cell (32x32)          frame (16x24) at origin (8, 4)
 └──────────────┘      └────┴────┴────┘
 ```
 
-That is what lets a sheet whose cells are sized for the widest pose — an attack,
-a swing — still expose a tight, centred box for walking, so a character does not
-appear to change size when its animation changes. Leave the four keys out and
-frame == cell, which is what a simple sheet wants.
+Cells can fit the widest pose, such as an attack swing, while walking frames keep
+a tight, centred box. A character then keeps its apparent size when its animation
+changes. Without the four keys, frame equals cell, which suits a simple sheet.
 
-Only whole cells count: a sheet 70 pixels wide with 16-pixel cells has four
-columns, and the six leftover pixels are ignored rather than becoming a narrow
-fifth.
+Only whole cells count. A 70-pixel sheet with 16-pixel cells has four columns.
+The sheet ignores the six leftover pixels instead of making a narrow fifth column.
 
 ### Facing
 
-`flip_x` mirrors the frame **inside the same rectangle**, so a character
-occupies the same pixels whichever way it faces:
+**`flip_x` mirrors the frame inside the same rectangle**, so a character covers
+the same pixels whichever way it faces:
 
 ```ruby
 sheet.draw(renderer, row, col, x, y, flip_x: moving_left)
 ```
 
-There is no width to add back — see
-[Mirroring](drawing.md#mirroring) for why, if you are coming from Gosu.
+You add no width back; [Mirroring](drawing.md#mirroring) explains why.
 
-### Animations are handed back raw
+### Animations come back raw
 
-`#animations` returns the descriptor's table untouched. This class knows nothing
-about time: which frame to show at a given moment is the scene layer's job, and
-it builds its own animation state from that hash. Keeping the raw form here is
-what lets the two sides evolve separately.
+**`#animations` returns the descriptor's table untouched.** The sheet knows
+nothing about time. The scene layer decides which frame to show, and builds its
+own animation state from that hash. The raw form lets each side change on its
+own.
 
-A sheet with no `animations` key gets `{}`, not `nil` — a sheet of static tiles
-is a legitimate sheet, and a caller should not have to branch.
+A sheet without an `animations` key returns `{}`, not `nil`. A sheet of static
+tiles is a valid sheet, and callers should not have to branch.
 
 ### Slicing costs nothing
 
-Every frame is cut once, at construction, as a view onto the single upload. A
-sheet of two hundred frames is two hundred small objects and **one** texture, and
-`#draw` is an array index plus one draw call. Nothing is re-cut per frame.
+The sheet cuts every frame once, at construction, as a view onto the single
+upload. Two hundred frames are two hundred small objects and **one** texture.
+`#draw` is an array index plus one draw call. Nothing is cut again per frame.
 
 ### Loading
 
 ```ruby
-RGame::Core::SpriteSheet.load(app, path)   # standalone
-RGame::Core::SpriteSheet.new(image, atlas) # from an already-loaded image
+app.assets.sheet(path)                     # what a game calls: cached and grouped
+RGame::Core::SpriteSheet.new(image, atlas) # from an already-loaded image and parsed descriptor
+RGame::Core::SpriteSheet.load(app, path)   # reads both files directly, bypassing the cache
 ```
 
-Use `.load` for a game with a sheet or two and no asset manager. The asset
-manager uses the second form, with an image it has already cached, so a sheet's
-PNG is shared with a standalone load of the same file rather than decoded twice.
+**Load sheets through `app.assets.sheet`.** Every app has an asset manager. It
+builds the sheet with `.new`, from an image it pulls through its own cache, so the
+sheet's PNG is shared with `app.assets.image` of the same file. `.load` reads the
+descriptor and decodes the image itself, outside any cache. A second `.load` of
+the same file decodes and uploads it again. `UiAtlas` has the same three forms.
 
 ## Nine-slices
 
-A bordered texture drawn at any size, by cutting it into nine pieces and
-treating each differently.
+A nine-slice draws a bordered texture at any size. It cuts the texture into nine
+pieces and treats each piece differently.
 
 ```ruby
 panel = RGame::Core::NineSlice.new(image, x: 0, y: 0, w: 26, h: 28,
@@ -248,37 +258,36 @@ panel.draw(renderer, x, y, width, height, z: 0, color: nil)
   └──┴────────┴──┘
 ```
 
-One small piece of art fills a button, a dialog or a health bar of any size,
-without the corners smearing.
+One small piece of art fills a button, a dialog or a health bar of any size, and
+the corners never smear.
 
-`(x, y, w, h)` is the source rectangle **inside** the image, so one sheet can
-hold many of them — which is what a [UI atlas](#ui-atlases) does with it.
+`(x, y, w, h)` is the source rectangle **inside** the image, so one sheet can hold
+many nine-slices. A [UI atlas](#ui-atlases) relies on this.
 
 ### Tiled, not stretched
 
-Edges and the centre **repeat**. Stretching a 7-pixel motif would blur exactly
-the detail the art was drawn for; repeating it keeps pixel art crisp at every
-widget size. Each band is clipped to itself, so the last tile in a row is
-cropped cleanly rather than spilling into the corner beside it — and the loops
-always start one more tile rather than stopping short, because a gap at the seam
-is more visible than an overhang that gets cropped.
+**Edges and the centre repeat.** Stretching a 7-pixel motif would blur the detail
+the art was drawn for. Repeating it keeps pixel art crisp at every size. Each
+band clips to itself, so the last tile in a row is cropped cleanly instead of
+spilling into the corner. The loops always start one extra tile, because a gap at
+the seam shows more than a cropped overhang.
 
 ### `border` and `scale`
 
-`border` is either a uniform integer or a hash:
+`border` takes a uniform integer or a hash:
 
 ```ruby
 border: 7
 border: { left: 2, right: 6, top: 4, bottom: 4 }
 ```
 
-`scale` is an **integer pixel scale for the chrome itself**. Source art is
-small — corners are often 7 pixels — so a scale of 2 or 3 gives legible borders
-on a 640x480 screen with no blurring at all, because every source pixel becomes
-a whole square of screen pixels. It scales the pieces *and* the step between
+**`scale` is an integer pixel scale for the border art itself.** Source art is
+small, with corners often 7 pixels wide. A scale of 2 or 3 makes borders legible
+on a 640x480 screen without blur, because each source pixel becomes a whole
+square of screen pixels. `scale` multiplies both the pieces and the step between
 tiles, so the tiling stays seamless.
 
-### Edge cases, and what they do
+### Edge cases
 
 | | |
 |---|---|
@@ -289,24 +298,29 @@ tiles, so the tiling stays seamless.
 
 ### What it costs
 
-The nine pieces are cut once at construction, as views onto the one upload, so
-`#draw` allocates nothing. It issues one call per tile, which is what makes
-`scale` worth having: a panel drawn at 3x is a ninth of the tiles of the same
-panel drawn at 1x.
+The nine-slice cuts its pieces once at construction, as views onto the one
+upload, so `#draw` allocates nothing. It issues one call per tile. That is where
+`scale` pays: a panel at 3x needs a ninth of the tiles of the same panel at 1x.
+
+Inside a scene, draw a registered nine-slice by id with
+`renderer.nine_slice(id, x, y, width, height, z: 0, tint: nil)`. Its `tint:` is
+`NineSlice#draw`'s `color:`.
 
 ## UI atlases
 
-One sheet of UI chrome, cut into named [nine-slices](#nine-slices).
+A UI atlas cuts one sheet of UI art into named [nine-slices](#nine-slices) and
+named images.
 
 ```ruby
 atlas = app.assets.ui_atlas('ui/ui_atlas.json')
 renderer.register_ui_atlas(atlas)
 
 renderer.nine_slice(:button_idle, x, y, width, height)
+renderer.image(:home, cx, cy)
 ```
 
-A button has four states, a panel has one, a scrollbar has three pieces — all
-small, and all cheaper as sub-rectangles of one texture than as a dozen files.
+A button has four states, a panel one, a scrollbar three pieces. All are small,
+and sub-rectangles of one texture cost less than a dozen files.
 
 ### The descriptor
 
@@ -319,94 +333,117 @@ small, and all cheaper as sub-rectangles of one texture than as a dozen files.
     "button_focus": { "x": 43, "y": 59, "w": 26, "h": 28, "border": 7 },
     "panel":        { "x": 0, "y": 0, "w": 32, "h": 32, "scale": 2,
                       "border": { "left": 4, "right": 4, "top": 8, "bottom": 4 } }
+  },
+  "images": {
+    "home": { "x": 0,  "y": 96, "w": 50, "h": 50 },
+    "gear": { "x": 50, "y": 96, "w": 50, "h": 50 }
   }
 }
 ```
 
-`image` is resolved next to the descriptor. Each entry is a source rectangle
-plus a `border` — a uniform integer or one value per side — and an optional
-`scale` that overrides the sheet-wide one. A sheet with no `scale` draws at 1.
+The atlas resolves `image` next to the descriptor. Each `nine_slices` entry holds
+a source rectangle and a `border`: a uniform integer or one value per side. An
+optional `scale` overrides the sheet-wide one. A sheet without `scale` draws at 1.
+
+Each `images` entry is a bare rectangle, cut from the sheet with
+`Image#subimage`. An icon draws whole, so it needs no border. The draw call's
+`scale:` sets its size, so it needs no scale either. `atlas.images` is a Hash of
+name to `Image`. Either section may be missing or `null`, and one atlas may hold
+both.
 
 ### Element names, not filenames
 
-`nine_slices` is keyed by whatever the descriptor calls each element, and those
-names are what a widget asks for. That is why nine-slices are the one asset the
-renderer resolves **by registration only** — `:button_focus` is not a file and
-never can be. `register_ui_atlas` binds every element in one call:
+**Both sections are keyed by element name, and widgets ask for those names.** So
+the renderer resolves nine-slices **by registration only**: `:button_focus` is
+not a file and never can be. `register_ui_atlas` binds every element of both
+kinds in one call. It registers nine-slices with `register_nine_slice` and images
+with `register_image`:
 
 ```ruby
-renderer.register_ui_atlas(atlas)          # all of them
+renderer.register_ui_atlas(atlas)                                  # all of them
 renderer.register_nine_slice(:panel, atlas.nine_slices[:panel])   # or one
+renderer.register_image(:home, atlas.images[:home])
 ```
 
 ### When an entry is wrong
 
-A descriptor holds a dozen of these, so a broken one **names itself**:
+**A broken entry names itself.** A descriptor holds a dozen elements. A nine-slice
+whose border does not fit, or an image rectangle past the sheet's edge, raises
+with the element's name:
 
 ```
 ArgumentError: ui atlas element :button_idle: nine-slice borders (40, 40, 40, 40)
                do not fit in a 26x28 rect
 ```
 
-Without the element name the failure is arithmetic from inside `NineSlice`, and
-finding the culprit means bisecting the JSON by hand.
+Without the name, the error would be bare arithmetic from inside `NineSlice`.
+Finding the culprit would mean bisecting the JSON by hand.
 
-Parsing happens once, at load. Nothing here is touched again per frame.
+The atlas parses once, at load, and touches nothing again per frame.
 
 ## Tile maps
 
-Draws a Tiled map: the static layers baked once, the animated tiles drawn each
-frame and culled to the viewport.
+`TileMapRenderer` draws a Tiled map. It bakes the static layers once, and draws
+animated tiles each frame, culled to a rectangle of the world.
 
 ```ruby
-tiles = app.assets.tilemap('map/island.tmx')
+tiles = app.assets.tilemap('map/island.tmx')   # => RGame::Core::TileMapRenderer
 
-renderer.tilemap('map/island.tmx', 0, camera_x, camera_y, view_w, view_h, elapsed: seconds)
+renderer.tilemap('map/island.tmx', 0, cull_x, cull_y, cull_w, cull_h, elapsed: seconds)
 # ... the scene draws its actors here ...
-renderer.tilemap('map/island.tmx', 1, camera_x, camera_y, view_w, view_h, elapsed: seconds)
+renderer.tilemap('map/island.tmx', 1, cull_x, cull_y, cull_w, cull_h, elapsed: seconds)
 ```
 
-### One call per layer, because the actors go between them
+**A game rarely makes these calls.** [`TileMapLayer`](components.md#tileworld)
+mounts one node per layer and draws it.
 
-A layer is drawn on its own, in the order the caller asks for — which is what
-lets a scene put its actors between two of them, trunks under and canopies over.
-Which layers those are is a question about the scene, not about the map, so no
-`z` is passed: in a game it is [`TileMapLayer`](components.md#tileworld) mounting
-a node per layer, and the scene tree deciding the rest.
+**Tiles draw in world coordinates.** A tile at column 3 lands at
+`3 * tile_width`. The caller's transform, usually a `WorldView`'s camera, puts it
+on screen. The rectangle is only a **cull rect**: the part of the world worth
+drawing. A camera supplies it but does not move the result, so one map can be
+drawn through several cameras in one frame.
+
+### One call per layer, so actors fit between layers
+
+**Each call draws one layer, in the order the caller chooses.** A scene can put
+its actors between two layers: trunks under, canopies over. Which layers those
+are depends on the scene, not the map, so the call takes no `z`. In a game,
+[`TileMapLayer`](components.md#tileworld) mounts a node per layer, and the scene
+tree orders them.
 
 ### What it costs
 
-Within each layer, every tile that is **not** animated is baked into a
-[recording](drawing.md#recordings-bake-once-replay-cheaply) the first time that
-layer is drawn. Scrolling it afterwards is one call per texture, however many
-thousand tiles went into it. The handful that *are* animated are drawn
-individually, **culled to the viewport** — so a map far larger than the screen
-costs only what is on screen.
+**The renderer bakes each layer's non-animated tiles into a
+[recording](drawing.md#recordings-bake-once-replay-cheaply)** the first time it
+draws that layer. Scrolling the layer then costs one call per texture, however
+many thousand tiles it holds. The few animated tiles draw individually, **culled
+to the viewport**. A map far larger than the screen costs only what is on screen.
 
-Two maps sharing a tileset share one GPU upload, because the tiles come through
-the asset manager rather than being loaded by the map.
+Two maps that share a tileset share one GPU upload, because tiles load through the
+asset manager, not through the map.
 
-### Animation is advanced by you
+### You advance the animation
 
-`elapsed` is seconds, and it is an argument rather than a clock this reads:
+**`elapsed` is seconds, passed as an argument.** The renderer reads no clock:
 
 ```ruby
-def update(dt) = @elapsed += dt
-def draw(renderer)
-  renderer.tilemap(@id, @layer, camera.x, camera.y, w, h, elapsed: @elapsed)
+def on_update(dt) = @elapsed += dt
+
+def on_draw(renderer, view)
+  camera = view.camera
+  renderer.tilemap(@id, @layer, camera.x, camera.y, view.width, view.height, elapsed: @elapsed)
 end
 ```
 
-Stop accumulating and the water freezes; accumulate slower and it runs slow; a
-spec passes `0.15` and gets the second frame. See
-[the frame loop](app.md#the-frame-loop) for why nothing on a draw path reads a
-clock.
+Stop accumulating and the water freezes. Accumulate slower and it runs slow. A
+spec passes `0.15` and gets the second frame. [The frame loop](app.md#the-frame-loop)
+explains why nothing on a draw path reads a clock.
 
-### It is wired up, not built in
+### Installed, not built in
 
-`RGame::Core` cannot parse a `.tmx` — that is the engine layer's job, and Core
-is not allowed to know the engine layer exists. So the type is *installed*, by
-the one class that may name both:
+**`RGame::Core` cannot parse a `.tmx`.** Parsing belongs to the engine layer, and
+Core may not know that layer exists. `RGame::Game`, the one class that may name
+both, installs the type:
 
 ```ruby
 app.assets.add_loader(:tilemap) do |path|
@@ -417,9 +454,10 @@ app.assets.add_loader(:tilemap) do |path|
 end
 ```
 
-Until that runs, `app.assets` has no `tilemap` accessor and a tilemap draw id
-raises `KeyError` — which is the honest answer, rather than a half-working
-subsystem.
+Every `RGame::Game` installs this loader when it is built. A plain
+`RGame::Core::App` has none: its `app.assets` has no `tilemap` accessor, and a
+tilemap draw id raises `KeyError`. A clear error beats a half-working subsystem.
 
-`TileMapRenderer#map` hands the parsed map back, for the scene's own collision
-and world-bounds queries.
+`TileMapRenderer#map` returns the parsed map, for the scene's own collision and
+world-bounds queries. [Tile maps](tile_maps.md) documents `TileMap` and `Tileset`,
+and which Tiled features rgame reads.