Plural OS is a multitasking, microkernel operating system being developed based on the XOpen Single Unix Specifications. The project only started a month ago. Another hobby OS, Clicker, a multitasking open-source operating system project, is based on a modular microkernel and a brand-new architecture. With its new release today, Clicker made a leap, the OS was able to run its first binary application.
New Releases for Clicker and Plural
2001-10-29 OS News 24 Comments
Is it just me that thinks there are too many of these ‘Hobby’ OS’s?
I know its fun to do, I wrote one a few years back, 32bit/PM/VM etc, so I realise how slow development can take. After a few months I just got disheartened with the required work and I am sure that is what happens to almost all of these projects. One notable exception is AtheOS which seems to have progressed very far considering it is only really Kurt working on the kernel/GUI.
Shouldn’t we all be joining forces? After all most of the ones being developed end up just feeling like another Unix with what the coders think is a smart gui.
What I would love to see and would to help out on is a **completely** new design for an OS, something that really breaks the mould instead of rehashing old ideas, I have no idea what that is but somebody a LOT smarter than me must have an idea somewhere!
Just my 2 pence.
Just to follow on from what I just wrote.
What I would like to see is a freely available kernel, with good drivers/VM/FS support/Multithreading that people can use as a base model for their GUI/WM/Environment on top. Kinda like a good foundation so people dont have to complete reinventing the wheel. I know there is Linux but I feel this falls short in some areas.
That way people could rapidly develop thier OS knowing a good solid continously developed kernel is there to use. Think BlueOS and what they are doing (but I dont think Linux is the best choice for them, but thats another discussion).
“I know there is Linux but I feel this falls short in some areas.”
Ha ha. What areas? What you just described *IS* Linux.
Linux is just a kernel. And a very good one at that. Whatever you put ontop of it is entirely up to you. The reason “BlueOS” chose linux is the same reason as every other embedded systems/application systems/any other systems company uses linux – because its stable as all fuck, actively maintained, and supports a huge amount of hardware.
If you’re not happy with linux (and you don’t really provide such a good argument, but anyway), try FreeBSD or NetBSD or OpenBSD or HURD or the multitude of excellent kernels with source code available out there.
Maybe I am just biased from my experience with the Linux distributions built around the Linux kernel, is there any project out there using the Linux kernel for something different than Unix type OS’s?
But I have been reading that its VM is far from ideal (FreeBSD is better I believe) and its multithreading capability is less than optional.
The Linux kernel is a Unix (OK a unix clone if you wanna be correct) kernel (for example its tied to the unix directory structure). Therefore it’s a bit of a problem using it in anything other than a Unix OS. I don’t get one thing, What’s wrong with Unix?
to the above ‘why don’t we all work together’, what would be the fine in that, plus not everyone wants the same thing from an OS.
Why don’t we all just use Windows, it’s stable-ish, well supported…
> > But I have been reading that its VM is far from ideal (FreeBSD is better I believe) and its multithreading capability is less than optional. < <
These are the main problems in Linux at present, but there is a considerable amount of work being done in those departments. Multithreading can be significantly improved with Robert Love’s kernel preempt patch. It is not in the official kernel yet, but it is slated for 2.5. For now, the patch works brilliantly, and I don’t know of any real drawbacks (there’s an almost negligible decrease in overall throughput but that’s about it). The VM situation is a little thornier, as we now have two VMs (the old Rik van Riel one, which Alan Cox still uses, and Andrea Arcangeli’s new one). This will be rectified over time, as one VM proves to be better than the other.
As a developer in the advanced stages of OS development, I too get weary of seeing all these new OSes sprout like mushrooms. I wonder if it’s an ego or fame thing. Lots of people see Linus as their hero role model and want to replicate it I guess, lucky fella 😉 I remember before Linux was big and doing a OS was quite a daunting task, and being truly amazed at the 386bsd effort which occurred at around the same time.
There seem to be a lot more people doing it for the fun of it now, but I guess it has always interested CS people. Heck, I remember in my undergrad days a couple of close colleagues pseudo coding up a bit of sample OS code for the mainframe that we used – this was well *before* the days of PCs and stuff when writing software was a real art – we didn’t even have the standard VDU that mid 70’s people took for granted, and just getting access to a terminal was really tough. I even remember a bit later when 3 or 4 or of us wanted to win a commercial software tender by building the hardware and software for a P-machine emulator – boy were we naive. Those were days dominated by the minicomputers like the PDP-11 and where a business had to invest tens of thousands of dollars in a single system. I also remember comparing the latest 16 bit micros coming out at the time – 4 big ones were contenders. The Z8000, the PACE, the 8086 and the 68000. Of these the PACE was the worst. It was an oversized PDP8 – really ugly. The next worst was the 8086 – we thought it wouldn’t amount to anything – and it probably wouldn’t have if IBM hadn’t chosen it as their processor of choice. I was truly astounded when that decision came through a few years later. The Z8000 was a little better but the 68000 come out on top because it looked more like the traditional computers than anything else. Not surprising it was widely used until the x86 line finally took hold.
Ah those were the days…. real pioneering. OS work is pretty tame these days – so much info and cheap hardware at ones fingertips.
Dont get me wrong, I dont see anything wrong with Unix, I would just like to see an OS project go in a new direction, an environment that pushes some new boundary.
EVERY time the topic is kernel development, the Linux kernel inevitably pops up! ANd then takes center stage, even though there are other, maybe even superior kernels, out there. Well, screw that, I had enough of this Linux HYPE!
I’m gonna scrap linux from my work-workstation, as long as it’s an x86 Unix, it’ll do, and it won’t have a Linux kernel. Because.
When you see a headline about a “kernel developer” it’s only and exclusively about a Linux kernel developer! If it’s not working on a Linux kernel, it’s just a developer, well, screw that!!
HURD was supposed to replace Linux for GNU/Linux. These guys have been working on HURD for AGES now, what the heck are they doing? Look at all the development in kernels, Clicker, and even Plural, seem to progress with giant steps, compared to HURD. Either the HURD guys are just a bunch of tossers, or the project went bezerk and is approaching the Great Grawe of Butthead-Developed Projects.
My only hope is NewOS and maybe OpenBeOS. Oh, and SkyOS looks promising!
Plural OS started just one month ago? It has been registered at SourceForge in 1999 and most of the files in the cvs are 21 month old!
You do realise you can get a complete debian install based on HURD *right now*, right?
>to the above ‘why don’t we all work together’, what would be the fine in that, >plus not everyone wants the same thing from an OS.
The great thing with linux/bsd/”your favorite Unix” is that is ~opensource. The thing that is great with linux is that linux is the Unix version/clone/derivate with the most drivers…
What do you want that Linux doesn’t provide?
>Why don’t we all just use Windows, it’s stable-ish, well supported…
But It isn’t GPL/LGPL/BSD Licensed. I do not want to run a OS without sourcecode that I can change..
Just my two kronor..
>Dont get me wrong, I dont see anything wrong with Unix, I would just like to >see an OS project go in a new direction, an environment that pushes some new >boundary.
I understand what you mean but Is it something that Unix doesn’t do good/can be made better that you can list?
Maybe, but when I mean something new, I mean different File structure/completely new idea for a gui. I know its hard and I dont say I can do it, but I am sure there are people out there with new ideas.
>”I mean different File structure”
I agree with you partly. I have created a somewhat different file structure on my “Customized Linux/Glibc system” (I have compiled and enhanced from only sourcecode). I have for example removed the usr,local and opt dirs. My system proves that a different file structure doesn’t mean you can’t use the Unix system.
>”completely new idea for a gui”
I can agree with you there. But a new GUI doesn’t mean we cannot use
Unix. On the contrary look at MacOSX.
>I am sure there are people out there with new ideas
I hope so.
okay, this sure gives hope: http://www.urbanophile.com/arenn/hacking/hurd/hurd-hardware.html
Time to toss Linux?
For a different type of OS you could Try Plan9 http://plan9.bell-labs.com/plan9dist/ I’ve heard it’s a bit different and I haven’t got around to trying it myself.
>What do you want that Linux doesn’t provide?
A nice, coincident interface.
A wealth of applications and games from my local shop.
“Double click” or less installing of drivers.
Now I could fork of Linux to do this, but then why? Windows\MacOS do all of the above.
And even then, why should anyone write there own, The authers might not like the GPL or the direction of the Linux kernel. Should there freedom to do what they like be taken away inorder to bring forward Linux?
> I do not want to run a OS without sourcecode that I can change..
And I don’t want to worry about having to change the source code to do what I want.
What do you want, to force every one into using the same OS?
 USB forever 😉
 But it still will not support as much hardware as Windows
“should” should be “should not”
>>What do you want that Linux doesn’t provide?
I wrote “Linux” but meant “Unix”.. Sorry for that.
>What do you want, to force every one into using the same OS?
What I meant was that the _Unix_ type of OS (and philosophy)
is great. If I would create a new OS I would implement a new Unix OS.
>A nice, coincident interface.
That doesn’t have anything to do with Linux or Unix for that matter. It’s just a matter of
replacing X with something different.
>What do you want, to force every one into using the same OS?
Hehe, Nope I am not Microsoft. Seriously I love that there are a lot of OSes
I just don’t think that there are any _big_ things with Unix that needs fixing.
And If you like many OSes you gotta love Unix since there are so many flawors
>And I don’t want to worry about having to change the source code to do what I want
Thats a good point but I personly wan’t to be able to edit the source (for customization for example).
PS I’m sorry for my bad english
Hello all, I’m Pype, team leader of the Clicker project.
I liked the messages from Darly Dudey:
>but when I mean something new, I mean different File structure/completely new
>idea for a gui. I know its hard and I dont say I can do it, but I am sure there
>are people out there with new ideas.
That’s truly what we’re trying to do. So much ideas about new concept for GUI and filesystem (including native metadata support like author/notes/fileclass, intelligent directories etc.)
See the “idea” page if you want to see what we plan to do
Why we do a new kernel from scratch is that, very often, the existing Operating Systems are now so huge that you can’t figure out how it works (really). Moreover, some of the “new concepts” of Clicker (like resources, memory objects or Kernel Dynamix System) could not be installed in the Linux kernel (or another one) without completely re-designing it …
For those who tell us that “we’re re-inventing the wheel”, just remember we’re still using 30-years ago wheeles (unix) concepts everywhere. Roads have changed! Files do no more hold text, but music, pictures, sourcecode … That’s why something new must appear in filesystem. And no existing OS (neither Win nor Lin) will make it appear because it’s too much breaking innovation (they prefer step-by-step innovations)
Same for the GUI: it becomes nicer and smarter every release of toolkits, but why can’t we do flash within a window ? why can’t i easily make a remote control for Winamp if this remote control is not a Winamp plugin ??
Something here too has to appear to make the programs interact more easily within the same GUI in order to make the whole environment more intelligent.
I agree with the fact that there are currently many startups kernels, and that it would be better if we had a single whole team working on a single OS. but unfortunately, we’re not an huge enterprise, but rather several standalone people. Merging all the projects would need documentations effort and many
I have seen a couple of days ago a “webring” project about ASM-made operating systems. If someone want to make such merging page with “3rd generation OS”, count me in !