bill-auger 7fa493303c update overview doc 2 months ago
..
EXAMPLE_WORKFLOWS.md db7d36676e update overview doc 4 months ago
README.md 7fa493303c update overview doc 2 months ago

README.md

ForgeFed - Federation Protocol for Forge Services

Motivation

Distributed version control systems (VCS) were created to allow maximal flexibility of project management structures and code hosting, in contrast to the client-server version control systems that were most widely used at the time, which denote one replica as the canonical master source. Existing project management / code hosting websites (forges) soon began supporting these, and some new ones sprung up as well; but even the new ones were modeled upon the centralized "hub" paradigm (star topology, in networking lingo); with a single canonical master "upstream" parent replica, and all other replicas implicitly and permanently designated as "downstream" child "forks". This type of website well serves the traditional purpose of facilitating release distribution, collaboration, and end-user participation; but in discordance with the decentralized nature of distributed VCS.

Indeed, although distributed VCS such as Git have become widely accepted as the de-facto state-of-the-art, it is still, for good reason, the standard practice that one replica will be designated, albeit logically by convention, as the canonical upstream source; so this retro-fitting of distributed VCS onto the traditional centralized model, is not often contested. Philosophically speaking though, this has the consequence of casting all software development as intrinsically hierarchical in nature; which is often undesirable, as it is antithetical to truly open, non-hierarchical project management "structures" such as adhocracy.

The goal of the ForgeFed project is to support the familiar collaborative features of centralized web forges with a decentralized, federated design that, by fully embracing the mostly forgotten merits distributed VCS, does not rely on a single authoritative central host, does not impose a hierarchical, master/fork collaboration structure, and can be self-hosted by anyone; with all such independent peers cooperating to form a larger logical network of inter-operable and correlated services.

Design Goals

  • Transparent authentication and collaboration across federated instances
  • Participating servers and their repos may be either private access or public
  • Users should never need to trust any server in the network other than their home-server
  • Users should never need to send any login credentials to other participating servers
  • Users should never need to run any JavaScript from other participating servers
  • Users should be able to interact with foreign repos in all of the typical collaborative ways, just as if they had an account on each foreign host
  • Most (or ideally all) collaborative interactions should be accessible with or without a web browser (e.g. via email, custom clients, etc)
  • The preceding, closely related, three bullet points are intended to allow interfaces to be maximally customizable; so that for examples:
    • A) People who rely on accessibility features could run a home instance or client which is particularly well suited to screen readers
    • B) The cool kids can use or create snazzy CSS/JS websites for web browsers
    • C) Yet others could interact with the same services on a headless server using mutt

How It Works

  • Everyone can view repos on public hosts without logging in (just as you would expect)
  • Users can create an account on any public instance or may host their own - (this server will be henceforth referred to as that user's "home-server")
  • Users can create repos on their home-server only
  • Users can fork foreign repos to their home-server without signing-in to the foreign host
  • Users can send merge requests (eg: "pull-requests", raw patches), open tracker issues, post comments, subscribe to updates (eg: "watch", "follow"), and endorse foreign repos (eg: "favorite", "star"), all without signing-in to the foreign host
  • All non-trivial user interaction with foreign hosts are mediated by the user's home-server
  • Users never need to interact directly with any foreign host; though it is possible for it to appear transparently as if doing so, albeit indirectly in reality
  • All of the above interactions will be possible with or without a web browser (e.g. via email)
  • Savvy admins and users can interact with the system by implementing parts of the protocol in custom services and clients
  • Home-servers will verify the identity of foreign users against credentials supplied by that user's home-server
  • Home-servers will vouch for the identity of their users using HTTP signatures
  • Home-servers could optionally sign submissions to foreign hosts with the user's GPG key, such as commits and comments, and send the public key along with the submission for verification
  • Home-servers could log actions of foreign users for display purposes (listing remote forks, listing remote subscribers, displaying validated signatures, etc), as phantom, non-login users in their database
  • Home-servers should notify foreign hosts of events to which their local users are subscribed, so that it may notify their local users (e.g. on-site alerts, email alerts, activity-pub "toots")
  • Home-servers should also periodically poll remote hosts, which host repos and/or users to which their users are subscribed, for events that may have been missed due to one of the servers being offline at the time the even occurred, in order to ensure maximal robustness
  • The preceding bullet point is intended to encourage users to self-host their own forge, even if they do so with a laptop, and also interact with their forge while offline, without missing incoming federated events nor dropping outgoing federated events. This would also allow forges to implement advanced features such as mirroring and synchronizing the entire data set of remote repos robustly; so that their users could interact with foreign forges in all of the expected ways while offline and/or without ever contacting the foreign directly.

See EXAMPLE_WORKFLOWS.md for some general ideas on how users could interact with the system.