Organisation

System

With the goal of lessening the likeliness of pull requests being left hanging around, making it clearer who knows what, giving more motivation via responsibility and change for the sake of change, Minetest's organization update of 2014-01-02:

The engine is split into virtual "subsystems" based on the most easily definable but still meaningful parts of the code (mostly each file should belong to one subsystem, but this is not strictly possible).

Each subsystem is allocated to one "subsystem maintainer", who is responsible for making sure pull requests get handled and fixing the subsystem if it doesn't work. Subsystem maintainers are trusted to give their permissions and/or responsibiltiies to someone else too if they see fit.

Subsystem maintainers can be proposed to the core team by anyone (preferably by themselves or by the previous one), and they are not-so-officially majority voted. The person on "maintainer conflicts" can turn down proposals on the basis of wrong direction or too much controversy (as rarely as possible).

The old rule of merging if two core devs agree still stays valid for when it is useful (= when the appropriate maintainer is away).

Github

Each issue and pull request should be assigned to a subsystem using the appropriate section label. For example, pull request #7 would get the label "Section - Build".

Subsystems

Developer	Subsystem
celeron55	maintainer conflicts
sapier	server/client/env
kahrl	startup/config/util
ShadowNinja	script API
hmmmm	mapgen
celeron55	client/audiovisuals
sapier	low-level network
[OPEN]	build system
proller	masterserver
ShadowNinja	Documentation

Those positions will maybe be assigned in the future:
releases
voting/community

Subsystem Files

(TODO)

maintainer conflicts

Meta-subsystem with the task of resolving conflicts between other subsystem maintainers that they can't solve themselves.

Proposed

proposed: releases

Meta-subsystem with the task of maintaining a reasonable release schedule by deciding when to do it, what is missing and announcing feature freezes and poking the necessary people to make it happen. The most important tasks are to make sure every subsystem maintainer knows when there is a feature freeze and when there is not, and to manage the tags and branches on github.

proposed: voting/community

Meta-subsystem with the task of figuring out and remembering what the community wants by reading discussions and arranging votings, and being a source of that kind of information to developers who want to focus on developing.