A camera represents how to render the virtual world. It's essentially a transformation from virtual game coordinates to pixel coordinates on the screen.

A camera that always shows the same area. This is useful in a top down game. You probably want to specify the area as a multiplication of width and height:

fixedArea (16*10) (x, y)

This means the camera will always show 160 square units of your game. In practice, this means that on a 16:10 viewport, 16 by 10 units of your game will be visible. But on a 4:3 viewport it would show 14.6 by 10.95 units. (sqrt(16104/3)=14.6, sqrt(16103/4)=10.95)

A camera that always shows width units of your game horizontally. Well suited for a side-scroller.

A camera that always shows height units of your game vertically. Well suited for a vertical scroller.

The custom camera allows you to use a function that maps viewport size (in pixel) to game units. E.g. here's an implementation of the fixedWidth camera using custom:

fixedWidth width =
    custom (\( w, h ) -> ( width, width * h / w ))

Move a camera by the given vector relative to the camera.

Move a camera to the given location. In absolute coordinates.

Smoothly follow the given target. Use this in your tick function.

follow 1.5 dt target camera

Gets the transformation that represents how to transform the camera back to the origin. The result of this is used in the vertex shader.

Given the screen size, gets the width and height in game units

Convert coordinates on the canvas element to coordinates in the game. Coordinates on the canvas element are given relative to its top left corner.

viewportToGameCoordinates camera (elementWidth, elementHeight) (positionX, positionY)

Element click coordinates can be extracted with a package like Elm-Canvas/ElementRelativeMouseEvents