What are the major differences between (the latest versions of) Unix, Plan 9 and Inferno? Is there anything that Inferno and/or Plan 9 do that would be harder to achieve with a Unix-based system?
To quote from that old thread (they discuss Taos not Inferno, but the distinctions from Unix are similar):
> could you efficiently implement Taos abstractions on top of POSIX? (probably yes) Could you efficiently implement POSIX abstractions on top of Taos? (probably no)
Thank you for your response, but I'm thinking perhaps I didn't explain the reason why I was asking. I understand that Inferno has a different design based on a VM, what I'm trying to ascertain is... why is it better than what came before? The design decisions were more than likely based on a desire to improve on what came before, my questions are about what the designers seeked to improve. In other words, rather than knowing it's a different design, I'd like to know why it's a better design (or to be more fair, what are its strengths and weaknesses)?
Plan9 (and Inferno, the virtualised descendent thereof) embody a single, unifying abstraction: that everything is a file. This often stated in Unix but (beyond the /dev and on Linux /proc & /sys filesystems) there’s lots of things that aren’t files (the framebuffer, network sockets...). In Plan9 and Inferno if you want to do something to anything all you need to do is know fopen & company.
I think people like plan9 more for it's minimalism rather than for it's special features.
Also, there still isn't anything quite like 9p for other OSes. Although many of them can mount 9p filesystems it doesn't match the VFS very well so it doesn't make a lot of sense to share device files. FUSE matches Linux's VFS closely but lacks authentication support so it isn't useful over a network without a special wrapper. Furthermore, a lot of these OSes rely on the setuid bit for a number of things which you can't have in an OS like plan9.
That whole OS is just very different, it's not easy to emulate on something like Linux.
Me too, it also saddens me that many aren't aware of Inferno OS and always think Plan 9 was the end of the line, when that status actually belongs to Inferno OS.
Which implemented the original vision for Plan 9, regarding what Pike wanted to do with Alef.
Too bad they didn't continue its development. Probably choosing a different GUI than TK would have helped to spark more people's interest in the project. There's a huge need for an alternative OS on Android phones, but most of the people who would benefit from it don't know that, so that need doesn't translate to actual demand.
I don't see Tk as being a problem. Rio, however, and the continued reliance upon 3-button mice, dates both Plan9 and Inferno and makes it quite hard to use them on laptops with trackpads.
The pain of developing it was one reason I didn't do much with Hellaphone after that summer. Other reasons: lack of funding and no spare phone to use if I wanted to develop it on my own time.
