So, I’ve been thinking about just what it would take to make a Linux-based Haiku. I think it’s downright staggering, to put it like that.
- It would need to reimplement all of the kits, in the style of, say, winelib.
- It would probably be a Wayland desktop environment, for simplicity’s sake. Making the actual app_server run on Linux looks like too much work, by far.
- It would need to have a production-ready OpenBeFS driver on Linux, on account of the fact that queries and attributes aren’t supported by the current filesystems (Correct me if I’m wrong, people!)
Those could all be achieved… On the other hand, is it really worth it just to have access to AMD and Nvidia drivers?
Also, you’d essentially be trading in the uniqueness of Haiku for a position as the strange, divergent Linux distro that does things which fundamentally make it incompatible with the Linux mainstream. Let me just say: Divergent Linux distros, unless they are funded by billionaires (cough Ubuntu cough) Tend to die off.
If it’s a simple matter of hosting everything that is Haiku on top of the Linux kernel like AROS-hosted, well… That involves too many tradeoffs to make it worthwhile.
I’d say Haiku ought to stay unique, although the potential for a fairly close companion distro that uses Linux should not be disregarded. It’s simply something for someone else to handle, possibly through doing 1) and then gravitating towards the other points. I wonder how 1) Would even arise, though. Several projects have tried in the past, and they have failed.