Plan 9 is worse than snake oil. Plan 9 was created to make distributed computing look like snake oil. Like UNIX compared to Multics, it arrived years after working distributed computing protocols and is still worse. AT&T's "research" is inferior to what was already on the market and will always be, no matter how much time and money is put into it. It's "research" that results in an inferior OS, increases costs on every single level, decreases productivity, replaces simple solutions with complex problems (sometimes so complex that nobody believes they can be solved), makes things worse for users, and prevents real researchers from doing what works because of "compatibility" with horrible interfaces. UNIX and Plan 9 are anti-research that shit on 60 years of computer science. That's why UNIX technology like Plan 9, BSD, and Linux in 2019 is worse than what Multics did in the 60s. Plan 9 uses "tar" to copy directory hierarchies not for any "philosophical" reasons, but because the version of UNIX it was based on didn't have "cp -r" and they were not capable of adding it. It doesn't have dynamic linking because they weren't smart enough to copy Multics or another mainframe system that does it properly. It still has the same bc calculator and all that other UNIX bullshit. What sucks is when Plan 9 weenies point to something that UNIX does wrong and blame the entire thing itself instead of the broken UNIX implementation.
Plan 9 is worse than snake oil. Plan 9 was created to make distributed computing look like snake oil...
Ok
>>He mad.
Golang is good
Yeah it's all bullshit. If it's not using a microkernel it's not real OS innovation. Thankfully the botnet is now funding and developing one, and there's also based seL4
>the version of UNIX it was based on didn't have "cp -r" and they were not capable of adding it
what the fuck
>microkernel
Aren't they inherently slower?
Yes, and retards even put the optionless plan9 commands in a side by side with the GNU commands to show how "bloated" GNU is.
I'm a minimalfag, and even I realize this is stupid. Having more options and features isn't bloat. Having more code than is necessary to implement said features is bloat, or having unnecessary onions-inspired features could be, but just having options for simple shit like recursion is not.
That's a FUD argument brought up because people can't seem to forget mach for some reason. Things got a LOT faster after the innovations of L4.
Also HURD is dead. The new gen microkernel projects are Fuchsia and Genode. Genode is often used with the seL4 kernel, which is an incredibly secure and yet high performance microkernel. It's secure because of formal verification, and it's fast because it's based on the L4 design and also parts of it are in assembly. I heard the U.S. military is starting to use it. Fuchsia is another new microkernel project based on Zircon, a fork of littlekernel. It's being made by Google, so it has a high chance of success due to having such a massive company behind it.
There's also Redox, which is the Rustfags' attempt at an OS, and also that Huawei Hongmeng/Oak/Ark/whatever that there were rumors about. That's apparently a microkernel as well.
You talk like a sane, reasonable human being. Why do you identify as a minimalfag then?
>Aren't they inherently slower?
Not necessarily.
The problem of micro-kernel fags is that they are purists who try to do everything that way, a bit like OO purists try to do everything the OO way (or functional programming purists, etc.)
But you have modern systems that are hybrids. Systems like Dragonfly BSD, etc.
Glowniggers fear 9front is what my takeaway from OP is.
He's not wrong, but the only major alternatives were turned to shit by similar big Corp incompetence, silos, and log rolling.
>t. Windows and Gentoo dev
>Like UNIX compared to Multics
maybe because the unix creators worked on multics
Good to know that the one meme OS is, in fact, a meme.
no, just Mach
seL4 is kinda shit for running Genode speed/feature wise. The better ones are NOVA, Fiasco.OC (another L4), and Genode's own microkernel they call base-hw. Fiasco.OC in particular is interesting because it's fast and works on 64 bit ARM. Genode on Fiasco.OC on an RPi >=4 would be a hell of a thing.
slideslive.com
skip to 20:40 for the performance part
I'm disappointed DragonFly gave up on single image unix clustering, although I suppose that's what 9front is. Another cool project is ChrysaLisp, basically attempting to revive the old TAOS distributed OS as proper Free Software and written partially in a Lisp dialect.
github.com
Drew DeVault is working on porting a Wayland compositor to 9front. Throw in a kernel plumbing layer for GPU drivers capable of wiring up to wlroots and Mesa, and suddenly 9front becomes a viable target for damn near the entire body of open source software. If 9front had GPU accel, Firefox, and mpv I'd switch my laptop to it in a heartbeat.
>multics
Are you that multics guy from 8ch /tech/? Never thought i'd see you here.
Unix and Plan 9 are the peak of computing you childfucking MIT nigger.
Plan 9 gave us acme and that alone makes it the greatest thing to ever come out of bell labs
Rob Pike is one of the most arrogant low IQ hacks in this industry. Also Plan 9 was a failed project and distributed systems are best implemented in userspace like Kubernetes
Yes, Hurd is.
> (s)he doesn't know what a 'research os' is
