DahliaOS - a Glass Elevator?

A community fork of Google Fuchsia has emerged called dahliaOS. Apparently it runs on both the native Zircon kernel optionally the linux kernel. There is even an opportunity to run a session in the browser.

The browser session suggests that the visual design and the range of programmes are not a patch on those of Haiku. I would definitely use Haiku on the desktop in preference to Dahlia. Nonetheless the underlying technology and what it represents is probably of more interest to us.

My own understanding is that Fuchsia was written by many former Beos engineers and offers many of the technologies that Beos might have pioneered if it had kept going. As such I feel that if DhaliaOS becomes viable, consideration should be given to building Haiku on Dhalia underpinnings after R1.This would bring the better desktop design, longer established native codebase and community of Haiku to what would seem a more modern system design.

I’m sure devs will keep a look at it but, there won’t be Haiku on Dahlia.

First because, in short, it would mean that Haiku becomes another linux distribution.
Devs already answered multiple times to that, it’s a definitive no.
linux design is almost as old than MSDOS. Haiku kernel is 20 years younger and is coded in a modern language, has been designed with multi-processors, attributes and multimedia in mind from start. What is more modern here?

Second because Haiku isn’t a kernel + an X server + a display manager + a window manager. Haiku is all these things in one therefore you can’t simply pick up a part of it and copy that elsewhere. Last twenty years, already multiple attempts to make something looking like BeOS or Haiku on an X server have failed. Most of time, their only convincing resemblance was the yellow tab. Honestly, that’s not enough.
But, there are people to keep on trying; they didn’t wait that announce. Perhaps someone will come out with something elegant but it won’t be easy, it will take time. Even longer because the others before failed and they can’t make public something not ready, not convincing enough; they will have to work alone for an eternity before it happens.
So, if you’re interested by doing it, as long as you respect licenses, nothing or nobody forbids you to try but you won’t receive help from Haiku devs and you can’t call that Haiku.

I would word it differently :slight_smile:

Indeed the Fuchsia team has several former Be employees in. They are writing a new OS, 20 years later, with experience elsewhere during these 20 years (possibly at Apple, Palm, Android, …). They are writing a new OS with different needs and concepts. So there are some ideas recycled from BeOS, and there are ideas from elsewhere or entirely new. Just as BeOS got some things from MacOS, UNIX, and Amiga, and added its own on top.

Dahlia seems to be mostly userland work on top of existing kernels. And you are essentially suggesting that we remove their userland to put our own. So, the only thing from their work we would reuse is… the Linux and Fuchsia kernels. Which we can already get if we need to.

That being said, it’s nice to see new operating systems appearing, and maybe we can collaborate in other ways :slight_smile:


It is worth noting that Zircon/Fuchsia is not a POSIX system (though of course it has design elements inspired by POSIX), which is a path we are really not interested in going down. We have enough trouble porting software as it is, with POSIX…

1 Like

I heartily agree with this sentiment. The more cross fertilisation that can happen, the better. I am interested to discuss the merits of Dahlia with respect to Haiku, and I apologise for wandering off into speculation about how Haiku may or may not proceed. :slight_smile:

I am puzzled how they are able to offer both zircon and linux kernels if Dahlia is not posix compatible which I thought was part of the guts of Unix. It also raises the question as to why bother with the Zircon kernel if the Linux one is quite capable of doing everything they need.

As an armchair follower of a lot of these projects (hence the perhaps naive pondering of the two different projects might merge) I am fascinated about the Fuchsia operating system which seems to be an opportunity do finally do a lot of stuff right that frankly should have been fixed ages ago. Perhaps it is the crowning work of Google? Apple’s new silicon Mac might well prove to be the hardware equivalent. They say that any empire in history produces its most grand enduring works just before they fall, and these massive concerns which had been happy to soldier on with unix with some java on top (Google Android) or x86 processors (Apple) might realise it too and have just produced their enduring legacies…

The Portable Operating System Interface ( POSIX ) is a family of standards specified by the IEEE Computer Society for maintaining compatibility between operating systems.[1] POSIX defines the application programming interface (API), along with command line shells and utility interfaces, for software compatibility with variants of Unix and other operating systems

^ Portable Operating System Interface (POSIX) (wikipedia)

The big deal here is that API compatibility, which means that software can be written once and compiled on/in other POSIX-compliant systems and environments with, afaik, fairly minimal changes.

I don’t know nearly enough but presumably you could incorporate most or all of the Linux kernel without being truly POSIX-compliant. If an application only interacted with the kernel/core then stuff coded to be POSIX-compliant might work okay.

The DahliaOS web demo things strikes me as being somewhere between ChromeOS and an Android-like desktop.