This is not very productive, so forgive me…
Did anyone consider using a well established unix/linux distro as the base of haiku?
This information is probably readily available and I just didn’t see it, but…
It seems like the haiku team is re-re-inventing the wheel.
BeOs was always nix based, or atleast structured like nix.
I just see Haiku in alpha 3, and it being so limited. Even a chop job of an os based on Ubuntu and using Ubuntu repos could run the plethora of available software in the Ubuntu repos.
Not that it has to be Ubuntu, I just happen to like it over the other 2 big dsitros, currently.
When it comes down to brass tacks, I guess I am just wondering if this idea was considered, and if it was considered, what was the driving point behind the decision the Haiku team made.
And old BeOs fan happy to see the Haiku project keeping the dream alive,
If you mean the BeOS API on top of a Linux kernel, yes, somebody tried that, it was called Blue-Eyed OS. Didn’t work out. IIRC someone else tried to put it on top of BSD as well.
If you just want a Linux distro with a BeOS skin, you can download one right now. It’s called ZevenOS.
There was a Linux-based initiative a long time ago, called Blue Eyed OS, but people gravitated towards OpenBeOS which became Haiku.
A few key arguments, from a hazy memory:
- binary compatibility with closed BeOS software (e.g. Gobe Producive)
- the filesystem folder structure
- the filesystems support for typed, indexed, extended attributes
- MIT license (instead of GPL) would lend itself easier to commercialization of the project
- Using a Linux-kernel would by sheer force/size/flow influence the direction too much in a non-BeOS direction.
Most people thought it would be easier to recreate Haiku from scratch than it would be to mold Linux into the desired BeOS-shape. (Or they preferred to do so as a challenge.)
If one were to make a focus shift to a Linux based Haiku today I’m not sure how to proceed. A Haiku on Linux layered/insular approach could keep much of the Haiku user experience intact, but it would be more of a linux Desktop Environment than an OS in its own right. From a Linux POV its probably more useful to develop on of the existing toolkits and desktop environments, like Qt/KDE, Gnome or XFCE.
Quite a few people have suggested similar ideas, like using different APIs such as QT as standard, etc. Personally, I think that the danger is that the end user will think ‘Why not just get Ubuntu/Linux/some QT OS and make life easier?’ This is a general problem in computing. Be innovative or be compatible.