From 2f8dab4537dda4205763ef2a48f298da9765a408 Mon Sep 17 00:00:00 2001 From: Peter Hutterer Date: Wed, 10 Mar 2021 16:46:43 +1000 Subject: [PATCH] doc/user: explain why we are doing motion normalization Signed-off-by: Peter Hutterer --- doc/user/normalization-of-relative-motion.rst | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/doc/user/normalization-of-relative-motion.rst b/doc/user/normalization-of-relative-motion.rst index 99c62f8..b7059be 100644 --- a/doc/user/normalization-of-relative-motion.rst +++ b/doc/user/normalization-of-relative-motion.rst @@ -27,6 +27,11 @@ libinput applies a dpi-dependent acceleration function. At low speeds, a movement speed increases, acceleration is applied - at high speeds a low-dpi device will roughly feel the same as a higher-dpi mouse. +The reason for the normalization is convenience: a caller can assume that a +delta of 1 should result in a movement of 1 pixel on a traditional +(low-dpi) screen. On screens with high resolutions, the caller must scale +according to the UI scale factors. + This normalization only applies to accelerated coordinates, unaccelerated coordinates are left in device-units. It is up to the caller to interpret those coordinates correctly. -- 2.7.4