In the WebGPU specification "Editor’s Draft, 16 February 2024" as currently posted on gpuweb.github.io, § 3.3 Coordinate Systems:

• Normalized device coordinates (or NDC) have three dimensions, where: … The bottom-left corner is at (-1.0, -1.0, z).
• Framebuffer coordinates …
  • The top-left corner is at (0.0, 0.0).
  • x increases to the right.
  • y increases down.
• Viewport coordinates combine framebuffer coordinates in x and y dimensions, with depth in z.

This implies that, when NDC are converted to framebuffer coordinates or viewport coordinates, the Y coordinate must be negated to convert from "increases up" to "increases down". But in § 23.3.5 Rasterization, the function for this conversion is given as:

framebufferCoords(n) = vector(vp.x + 0.5 × (n.x + 1) × vp.width, vp.y + .5 × (n.y + 1) × vp.height)

This does not negate n.y. This seems to contradict both the definitions of the coordinate systems, and what implementations actually do.