A new model for open source: closed to pull requests, open to remix

All right, I'm going to open up a former instruction since it's been a brace of generations of application engineers since this essay, and I have to assume not everyone knows what a cathedral is. Everyone knows what a market is since they're using it day to day alongside GitHub.

GitHub took the bazaar's ideas—a decentralized, open, self-organizing open-source project—and provided a workflow that lets anyone inspect or contribute origin code via a drag request.

I've been thinking that there needs to be a new application betterment model. We're not going rear to the cathedral, anywhere a small, private, exclusive collection builds the origin code, alongside releases that are heavily planned, infrequent, and polished.

But it's apparent we need item alternatively now that AI is here. The GitHub market model, equal before AI, was already showing its cracks since it created consumer-based entitlements anywhere everyone was nagging maintainers to merge something. It created really weird social-structure expectations: that person who releases code has an duty to keep it and get paid nothing for the work.

But now that AI is here, the dams have fine and really burst.

For what's the item of study a drag petition that's been generated by code? All I desire is to peruse and comprehend the center idea. If it's a fine idea, I power merge it rear upstream, but there have to be no anticipation that I will.

You see, open origin application is open origin by definition since the licence says that it's as is, and that as is permits you to remix software.

But the issue has been that it's been perceived as too difficult to remix software, but it hasn't been. It's fair a cognition gap matter that no longer exists now that agents are here. Nix has existed for a lengthy time. If you haven't seen the part that I posted a brace of days ago concerning Nix overlays, you should peruse it.

So, I'm going to propose item that I anticipation helps another open origin maintainers. This is how I'm thinking concerning my application now that I publish it under open-source licenses.

The Remix Model

• Structure: Decentralized and compositional. Ideas dispersed through autonomous overlays fairly than converging on a sole repository and a sole maintainer.
• Workflow: Publish origin and intent under an as-is license. Others remix by penning a small Nix overlay that expresses the desired alter once, afterward reuses that derivation throughout devenv shells, Docker images, NixOS systems and tests, CI, and delegate sandboxes—without forking or maintaining a divergent artifact.
• Core Principle: Ideas complete patches. The precious component is a alter in behavior or architecture. Maintainers decide whether to merge it, reimplement it, or disregard it. The first author has no position duty to review, integrate, or keep the change.
• Social Contract: The licence already permits remixing without creating a assertion on anyone’s forthcoming time. Credit and visibility can motionless exist; there is no ongoing care duty. Upstream incorporation remains optional and deliberate.
• Discovery: Useful adaptations appear as a proliferation of overlays in the wild. Maintainers can example from that dispersed if they choose; everyone alternatively can catch and create ideas merely by updating their own overlays via an agent.

Example: A person of a project wants to disable Git's capability to power shove in that project. Instead of raising a drag petition to the upstream project, the person makes an overlay that patches Git to refuse power pushes; this is defined formerly and afterward appears consistently in developer environments, receptacle images, and scheme tests—without a drag petition against the upstream project.

In practice, from this item forward, for any code I publish on GitHub, I volition disable drag requests, keep discussions open, and publish + keep direction on how you can remix my software.

The GitHub network chart provides a awesome way to examine the remixes in the ecosystem. Heck, you can equal run agents complete it to summarize what group are doing. You can additionally use it in an delegate iteration to examine remix ideas and nourish them rear upstream into the chief project.

ps. socials