Programmers would implicitly use a box filter when sampling from an underlying process (like 3d rendering, image scaling, scanning) into screen coordinates. It

Programmers would implicitly use a box filter when sampling from an underlying process (like 3d rendering, image scaling, scanning) into screen coordinates. It's one thing when writing a compositing window manager or blitting sprites to the screen; it's another thing entirely when doing any kind of image processing to make that assumption. It's similar to assuming that RGB (0-255) is perceptually linear and consistent among monitors (not understanding monitor profiles and color spaces)