What if someone made a Plan 9 that could run Linux apps?
liam-on-linux.dreamwidth.org
liam-on-linux.dreamwidth.org
http://9p.io/wiki/plan9/Linux_emulation/index.html
It seems to be rather outdated by now (only supports 32-bit x86 and might be problematic with current binaries), but it's probably a less resource-intensive solution than integrating a VM. Of course, this depends on the kind of Linux applications you intend to run, emulating an X11 or Wayland environment probably won't make too much sense.
Yes, I did find out in the responses to the post -- which was part of the point of the exercise.
Yes, there is indeed, but as you say, it's neglected, it's 32-bit only, and updating it would probably be a substantial amount of work. It was for the one in openSolaris, which is one of the jobs that Joyent did in making SmartOS.
Then, if it were updated, it would need continuous maintenance to keep parity with the Linux kernel. It's a moving target, and an exceptionally complex and difficult moving target as well.
Whereas what I'm suggesting -- and that's all it is, a suggestion, not even a proposal -- needs neither, it provides a much higher level of compatibility for little work, and merely by updating the kernel in the VM you get 100% Linux compatibility essentially for nothing.
The FreeBSD bhyve hypervisor does sparse memory allocation for guests: only guest RAM that's used is allocated on the host. So the guest instances could remain pretty small.
An unstated assumption is that this is an idea for 64-bit OSes (host and guest both). 32-bit Linux is on the way out now anyway.
It might use more computer resources but the idea is that it uses less programmer resources, which is a large part of the point.
People running 9front does it for convenience, to avoid the bloat, dinamically linked libraries/binaries and so on.
If any, 9front people run 9vx and OpenBSD on it with icewm or any sensible wm to run Firefox because of some JS web pages for gov/corpo related (and often personal) needs.
The point of Plan 9 was to eliminate the cruft of Unix. So I am not proposing adding the bulk or complexity of Linux to Plan 9.
What I'm trying to suggest is something that's small and simple to implement, but brings the richness of Linux apps to the more elegant next-gen OS... without incurring a huge cost in plumbing and maintenance.
The point being to try to attract some Linux users and developers over to Plan 9, by letting them bring their most essential apps with them...
And then seduce them into becoming Plan 9 users and developers instead.
Answered in my other comments.
> If Plan 9 had the same amount of hardware support as Linux has, then there would not be any need to run Linux apps in Plan 9
I do not see any connection between those things, TBH.
Better hardware support won't help with that. It's not even linked.
> Oh wait, there is no need to run Linux apps in Plan 9 anyway.
On a Linux box I can do a lot of boring essential daily stuff that I need in order to do my job that I can't do on Plan 9.
Like, you know, open 9 tabs onto different complex Javascript-driven websites, as I am doing at this moment in order to reply to you (among other things).
Maybe you don't need those things, but I do.