Impact is a big factor in how useful an activity is, even if it's tied to a big company.
Impact is a big factor in how useful an activity is, even if it's tied to a big company.
From a purely utilitarian perspective:
- The same amount of effort spent trying to document will result in more documentation being produced on a free system, since free systems are easier to document.
- The same amount of documentation produced will result in more benefit from end users for a free system, because more people will be able to use a free system.
Documentation effort towards Linux would help get that number up from nobody plus epsilon.
Also, at this point that's just not true because of Android. Most of the userspace is different from normal GNU/Linux distros, but not all: for instance, if your question is "how the heck am I supposed to connect to this WPA-Enterprise network," it's wpa_supplicant under the hood whether you're using Android's network dialog or GNOME's, and documenting wpa_supplicant's abilities and bugs helps Android users just as much as GNOME users.
The NDK APIs are quite constrained and only meant to be used for Java native methods, real time audio and high performance graphics.
Google can replace the kernel for anything else POSIX like, with the same set of NDK APIs.
Since Android 7, they have been locking down access to anything else not part of the official NDK APIs.
(But maybe I missed the point and we're specifically talking about developer documentation here?)
As for user documentation, any OEM is free to do whatever they feel like, so you don't have any guarantee how much of it is actually like AOSP.
Samsung is well known among Android devs for breaking AOSP compatibility.