Should we really fork?

Kevin · 2026-08-20T13:42:59+00:00

I think it is beneficial to have some structured conversation around the topic of forking since there have been multiple people talking about this, and infact some already have (i.e. Tabby for PPC). Note: I don't want this thread to develop into a flamewar about AI or the current Haiku developers, that's not practical or beneficial. Over 25 years the Haiku devs have built something truly special, and I am very appreciative of that. If it's time to part ways then let's do it in a respectful manner. I want to talk through what it would take to make a serious fork of Haiku, not just a one person maintained repo on GitHub, but an actual fork maintained by a team of people with the goal of distributing a usable operating system. My big question is would there be enough commonality among the community to make it worthwhile to fork? Or is everyone's vision so separate that it would result in more endless debates? Saying "oh just fork it!" is probably a lot easier than it sounds, so what would it really take? Here are my thoughts, I am curious what everyone else thinks. Goals Build a fast, fun, and reliable operating system based on Haiku Have rich cross-platform support. We already have pretty good x86 and x86_64 support from Haiku, maybe bring in PPC from Tabby (if the developer wants to), I am working on a SPARC port, others have made good progress on RISC-V and ARM Accept LLM contributions, but have a reasonable policy (i.e. must adhere to coding guidelines, must only copy license compliant code if sourcing from other OSes, must update documentation on what change and why, must be tested physically if possible, etc) Have rich branding and a good name. We want to attract more users and developers. Have open and transparent communication about the development and goals of the OS - not just comments buried in commits or chat on an IRC channel. Continue to support "obsolete" hardware (i.e. x86 and PPC) as long as practical or as long as their is interest from users and developers Maintain a high level of code quality through code reviews and testing Questions Would we maintain BeOS application compatibility? (I strongly vote for no) Would we maintain Haiku application compatibility? Do we add multi-user support? Even if it breaks BeOS and Haiku compatibility? (I vote yes) Are we forking and running away? Or do we want to bring in upstream changes from Haiku? What is our release structure? I am a fan of moving to a set schedule of 2 or 4 releases per year and abandoning the "one day we might make it to R1" approach. I think more frequent releases builds momentum. Just call the version based on year and month i.e. 26.8 for an 2026 August release, 27.2 for a 2027 Feb release. What would we need GitHub / source control Website hosting Forum hosting Build servers Devops tooling - how can we get automated builds and ISOs? Non-profit organization (eventually?) for donations and to fund paid developers (or at least pay for server costs) and handle trademarks What does everyone else think? What are we missing? This post is meant to start a conversation, I am no way trying to dictate the future.

Zardshard

I've linked this topic from https://discuss.haiku-os.org/t/the-ongoing-genai-debate-vol-1/19716/149. It would be impossible to please everyone, but a fork carries the promise of being able to please more of the people more of the time. Some people are primarily interested in LLMs. Others are primarily interested in avoiding them. And it is unlikely that either side will really ever change their minds much on the issue.

Certainly, as others have pointed out, a fork can take a lot of manpower, so it will be work. But I wish you all success!

Atomozero I think it would be a waste of resources not to use their code, written as it is by unblemished Tibetan monks.

I dislike name-calling. I prefer respecting other people and only arguing against their opinions where I disagree.

Atomozero Freedom of speech is definitely a tough topic with that team, especially if you cross their vision, whether it's right or wrong.

Eh, I don't think that's the whole story. I doubt that we've ever hidden a post with an opinion we disagree with where the opinion was advanced respectfully. It's when posts start getting somewhat disrespectful that moderation becomes harder (did this post cross a line or not?) and more susceptible to human biases. I, for my part, do not believe the moderation team is perfect. No team can be. Certainly, I am not. We can just work on our flaws together, respectfully.

Wishing good luck to all of you and peaceful coexistence with the main project! 😃

Anonymous1

Kevin
Good luck with that.

With participants who already demonstrated they are quick to call people disagreeing too much with their views "talibans" or just ending with "STFU", the starting base is not looking pinky here.

Kevin

I think a tweak suite would quickly become a distribution or fork. Haiku is modular, but some changes touch multiple modules.

For example, if I want to add true multi-monitor support to my RadeonHD driver (not the BeOS style dual head support, but proper independent displays) I would need to modify more than just the RadeonHD driver: the app server, accelerant API, the screen preferences app, the interface kit (so BAlert centers on primary monitor, not center), etc.

A distribution that brings in the latest changes from Haiku could work, at least at the beginning. Perhaps we could have an LLM agent review all the commits every X number of days and determine which ones conflict with code we have changed, then flag those for manual review.

In the long run I think we would out grow that. We may have driver changes, kernel changes to support PPC/ARM/RISC-V/SPARC/etc, improve the installer to complete the UEFI process, USB and media kit changes, make tracker and deskbar changes, maybe someone wants to adds transparency and floating kittens 😃 , etc.

I think the code bases might diverge rapidly. If that is the case ,it becomes a full fork and we just use an agent to review and analyze Haiku commits to determine if there is anything we still want to bring in.

Atomozero

Zardshard I dislike name-calling. I prefer respecting other people and only arguing against their opinions where I disagree.

I'm so sorry, I didn't mean to be offensive! I was just trying to bring a smile to your face 🙂

Zardshard We can just work on our flaws together, respectfully.

Everyone wanted it, but then they hit a brick wall.

Thanks, but we're still in the discussion phase, so it'll take a lot more time to see anything—especially since we have at least 25 years to reach R1 without feeling guilty 😃 (just kidding)

Kevin

Anonymous1 Good luck with that.

With participants who already demonstrated they are quick to call people disagreeing too much with their views "talibans" or just ending with "STFU", the starting base is not looking pinky here.

I know it will not be an easy task. Maintaining a health community will probably be harder than the actual work on the OS 😆 Being able to disagree in a healthy and respectful manner is a skill that will need to be developed and nurtured.

Kevin

clasqm I'm not worried about legal issues.

I agree. There should be no legal issues as long as we don't use the name Haiku, don't do anything intentionally confusing from a brand perspective (i.e. we can't call it HighKu), and don't use/modify any of their trademarks (logo and leaf). kallisti5 even said on the forums he would ensure the Haiku Inc doesn't fight a fork.

Atomozero

Kevin For example, if I want to add true multi-monitor support to my RadeonHD driver (not the BeOS style dual head support, but proper independent displays) I would need to modify more than just the RadeonHD driver: the app server, accelerant API, the screen preferences app, the interface kit (so BAlert centers on primary monitor, not center), etc.

The interesting thing is that in the community's underground, there are already two projects modifying app_server to implement transparency and multi-monitor support.

Kevin A distribution that brings in the latest changes from Haiku could work, at least at the beginning. Perhaps we could have an LLM agent review all the commits every X number of days and determine which ones conflict with code we have changed, then flag those for manual review.

I think this is the heart of this project's innovation, and I fully agree with this idea/need.

Anonymous1 With participants who already demonstrated they are quick to call people disagreeing too much with their views "talibans"

That awful person opened this forum and abandoned the old one after yet another attack against me by the masked penguin. I don't feel the atmosphere has improved at all since my departure, if anything it's gotten worse. 🙂

And let one thing be clear, I'll keep repeating it to the end: I am the first to be sad and bitter about the situation, but I definitely don't feel responsible.

Anonymous1

clasqm > we are talking about ONE fork.

It always starts with one fork.

But history show that the community who agreed on this one fork will also, given time, have its own division, leading to another fork.

Linux distro s is a good example.
But even Haiku history is. It's not the first fork from Haiku. People who were initially agreeing with the Haiku project ending to disagree enough to do their own fork too, already.

May this fork, if it happens, be successful enough, and one day it will face the same fate too. It's kinda unavoidable.
That why each fork leads to an higher risk of even more later.

It's way way more difficult to merge, to reunite , to keep people together on a shared vision than to fork, to divise, to split communities.

ewninja

A thing I haven't seen anyone bring up is that with AI:

Triage is way, way easier. Any kind of PR / issue could be immediately triaged & analyzed by the LLM before it even gets to human eyes, dropping down the amount of "junk" issues to deal with to nearly zero.
Testing is way, way easier. Smoke tests before every release can test literally every single item in the menu to see if it's correct, all without any human involvement.
Forking and merging is way, way easier. Both the infrastructure (like HaikuPorts) as well as "merge branch X with Y" are nearly effortless. This means that accepting patches from Tabby and other forks (as well as keeping up with upstream Haiku) won't be any problem.
Porting software can now be almost fully automated. That includes complex software like Blender or Node.

Many of the limitations we've been struggling for the last decades can be automated away now. It's important to not underestimate the amount of work involved in forking. However, it's also important to not overestimate it. Programming did change a lot in the last half a year.

Anonymous1

Atomozero > And let one thing be clear, I'll keep repeating it to the end: I am the first to be sad and bitter about the situation, but I definitely don't feel responsible.

That exactly why I wish "good luck with that".

Because the seeds for unhealthy community are still there: a trend to resort to name calling without any remorse, regrets but most importantly no accountability.

You're no the only ones I was referring too, btw.

What I mean is the community is not only divided around the AI topic. It's also, and my point is that the strongest issue, divided on disrespect. The goal doesn't matter anymore much, what matter is to be the one right on how to achieve it, even before it's actually achieved (if ever, like if softwares are ever "achieved").

These kind of division doesn't heal easily, if ever.

Don't be naive to think that it won't happened again in any new fork of a project. The issue is not the project, not it's goal, not even the tools to achieve it, the issue is, like so often in human history, communication and disagreement resolution.

Atomozero

Anonymous1 These kind of division doesn't heal easily, if ever.

If there ever is another Begeisterte, I hope instead that we can all have a beer together peacefully 🙂 maybe even having a laugh about it.

Anonymous1 Don't be naive to think that it won't happened again in any new fork of a project. The issue is not the project, not it's goal, not even the tools to achieve it, the issue is, like so often in human history, communication and disagreement resolution.

But this project was also created to try to curb this loss of amazing resources:

https://github.com/Chrizlys/SCHWARZPAUSE
https://github.com/enricopesce/haiku-arm64-port
https://github.com/rainygirl/haiku-sony-vaio-p-patch
https://github.com/atomozero/haiku-vaio-vpceb3k1e (Yeah, count me in too! 😃 )

and others that are private and don't want to come forward for now

I accept that Haiku's core shouldn't be 'dirtied' by these weekend programmers, but from the perspective of a user who uses Haiku every day and sails the seven seas of the internet, I think this is absolute gold and needs to be preserved, regardless of what tool was used to create or write it.

Anonymous1

Atomozero most of these forks have a small diff size vs the Haiku one.

The issue is the lack of will to push human-made change requests to Haiku. Some did.

As I said, it's far easier to start a division than to merge, to rejoin. It take more effort, at least at communication level.

@dcbeckman initial change request was rejected because the communication was LLM generated, a very long one, quite time consuming to read while the root issue and the fix was legit, a deadlock bug in devfs when a device is removed ved (USB one for instance).

He did the effort to redo another change request, and it was accepted and merged.

Working as a team has a cost. It's great to have more manpower, but it comes with a cost.
A cost that single-handed projects/forks are not necessarily ready for.

Null

Anonymous1 It's also, and my point is that the strongest issue, divided on disrespect.

Respect needs to be earned.

I am not your typical human being. I can read humans as an open book. I am a walking bullshit detector.

I also refuse to play the social game. If you say something inappropriate, usually a lie, I will correct you with facts. I do not care if you hate me for it. I only care to speak the truth.

My soul is very kind. But if you are not kind to me, or others, you are my enemy and I will expose you. I've lost count on how many psychopaths I've made cry or narcissists wanting to hurt me because they cannot manipulate me. I always stand my ground. I am also very intelligent, so I am almost impossible to fool (I wrote almost because some fools are impossible to detect without prior knowledge).

How come my posts are constantly being deleted or my account suspended when I do not break the rules? All I do is speak my mind without censoring myself. Am I the one dividing the community? Am I the one that is disrespectful? Should I respect those who do not respect me?

I disapprove of what you say, but I will defend to the death your right to say it.

– Not the Haiku forum moderators

Note that I have never had any problems with you. You are one of the good ones. I just replied to one of your statements.

Atomozero

Anonymous1 He did the effort to redo another change request, and it was accepted and merged.

See? There's a glimmer of hope! Differences between smart people are never final. Maybe it'll take a little time, or we'll simply end up standing next to each other at a booth, joking about the Walter trout 🙂

Anonymous1

Null

Respect needs to be earned.

Change requests submitted to Haiku are reviewed by humans, who ask currently that their review effort being respected by submitting changes written by humans.

Don't they earned anything, with some of these reviewers being involved for decades in Haiku?!

How come my posts are constantly being deleted or my account suspended when I do not break the rules?

I dunno, and as I don't know which username you're using on Haiku forum, I can't tell.

Should I respect those who do not respect me?

I guess, by your own logic, if your feel disrespected, you should earn it, then.
Debunking wrongs or even lies by facts is a way to do it, name calling is not.

I disapprove of what you say, but I will defend to the death your right to say it.
– Not the Haiku forum moderators

That's about freedom of speech in a time when people were killed arbitruary by the power in place (a king in case of Voltaire). Doesn't means that Voltaire though that freedom of speech, or any freedom, can be absolute anyway.

And yes, I know it's not actually a quote from him, it's from "The Friends of Voltaire" about what would be Voltaire attitude regarding freedom of speech.

Anonymous1

Atomozero I'm pretty sure we actually meet you and me at some BeGeister 20+ ago.

Do you remember the guy who demoed, from a floppy disk, the first boot of (non graphical) OBOS (mostly NewOS at this time) ? That was me.

Atomozero

Anonymous1 I have to head out now, but if I look properly I'm sure I have a photo of that moment 😃

Anonymous1

Atomozero

BTW, these projects are not lost, not until their respective creators decided to delete them before anyone made a clone of these.

weekend programmers

I'm a weekend programmer too, it's not pejorative.

I think this is absolute gold and needs to be preserved, regardless of what tool was used to create or write it.

And I think what missing the most is a collaborative effort to merge them, and that's the part that take time to us, humans: working together.

I'm pratically sure that if multiple AI agents driven by different users were fighting over a common code base, it will be also a mess quickly turning into a war (of who succeed to push before the others ones). Collaboration needs communication, and therefore time, because clearly we're all animals, and communication is not our best skill. We mosy all walks the same way, swin the same way, but after millenia we still have different languages, different way to communicate together and still a lot of miscommunication.

Null

Anonymous1 Don't they earned anything, with some of these reviewers being involved for decades in Haiku?!

I've never argued for or against that. So I don't know why you ask.

Anonymous1 I guess, by your own logic, if your feel disrespected, you should earn it, then.

No, that was not my logic.

I do not try to earn respect. If individuals like me and respect me that is fine. If not, that is also fine. I am always myself anyway.

It is human nature to defend your own beliefs, no matter if they are wrong. Since I am of the small minority that do not have that trait, I challenge those believes and for that I am attacked, ridiculed, censored and words are put into my mouth (meaning they make up things I never wrote and use that to attack me).

My point is respect needs to be earned not by being polite or "correct", but by speaking your will. We don't need to agree on everything.

Anonymous1 Debunking wrongs or even lies by facts is a way to do it, name calling is not.

I agree. I've never name called anyone. Name calling is something you do when you cannot argue for yourself. But I have called people untruthful, evil, manipulative etc. That is not name calling, those are my beliefs.

Anonymous1 about what would be Voltaire attitude regarding freedom of speech

Which is also how I used it; to illustrate the Haiku forum moderators attitude regarding individual forum members opinions that differs from theirs.

Kevin

Anonymous1 I'm pratically sure that if multiple AI agents driven by different users were fighting over a common code base, it will be also a mess quickly turning into a war (of who succeed to push before the others ones). Collaboration needs communication, and therefore time, because clearly we're all animals, and communication is not our best skill.

I agree! That was one of might thoughts in the original post, is there enough commonality in vision for multiple people to be working together towards a common goal? If everyone has goals so varied, or conflicting with the objectives of others, it may not make sense to join forces.

And I do think there needs to be some structure involved in the project. Even if vision is similar, how to execute on the vision may be a cause for debate as well. Hypothetical: What if the community tires of C and C++, but differs on approach. What if you say you want to rewrite every driver in Rust and I want to rewrite every driver in Zig? How do we decide?

« Previous Page Next Page »