Guest post by The mostly-morubund Hurd project is well known for what it’s not: the kernel at the heart of the GNU/Linux system. But there’s a long and interesting story about what it could have been, too. From Linux User magazine: “The design of the Hurd was an attempt to embody the spirit and promise of the free software movement in code.” Those are mighty ambitions, and this story is as much about competing visions as competing kernels. Says Thomas Bushnell: “My first choice was to take the BSD 4.4-Lite release and make a kernel. I knew the code, I knew how to do it. It is now perfectly obvious to me that this would have succeeded splendidly and the world would be a very different place today.” This is a well-written and fascinating read.
-> moribund
Can we stop parroting the “GNU/Linux” propaganda? Linux was a fully functional OS before it took any GNU code. It just replaced some of it’s code with GNU code to become POSIX compliant and generally better. The FSF’s propaganda that says Linux “started as a kernel”/”is just a kernel” is incorrect. It started as an OS. Oh, I see, just because only the kernel part of the original codebase survived, the FSF has the right to rename the project from “foo” to “GNU/foo”. Where is the relevant clause in the GPL that defines something like this? Oh yeah it doesn’t exist. Can you imagine that happening on other projects? If you incorporate too much code from GNU upstream, we renane your project! For people that started a whole fight about the “give credit” clause in BSD, the FSF are very annoying with their demand to appropriate Linux to themselves.
Unfortunatetly, this won’t go away anytime soon. The FSF zealots have more time in their hands that anyone else, and heaven forbid any opinion non-compliant to the FSF propaganda be heard. Woe onto anyone that says “open/closed source” instead of “free/nonfree software”. Never mind that open source is a legally protected trademark and has a robust definition while “free software”… not. But i guess that will explain the comments down below.
Oh, and Android doesn’t contain “Linux”, it contains the kernel part of Linux.
Browser: Mozilla/5.0 (Linux; U; Android 2.3.4; el-gr; LG-P990 Build/GRJ23) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1 MMS/LG-Android-MMS-V1.0/1.2
Linux is just a kernel. All the utilities and commands you run on your Linux machine are GNU. Linux is just a kernel image. Thats it.
I don’t like the GNU/Linux naming crap any more than you do, but credit where its due, the GNU project is a massive part of any linux system. Linux didn’t just take parts of it, most of the system is GNU.
No, *some of* the utilities and commands you run on your Linux machine are GNU. Important ones, to be sure – glibc, coreutils, sed, etc – but GNU don’t get *all* the credit. None of the init daemons are GNU, nor things like util-linux, most of the networking tools, process-management, etc..
Well, what I meant to say was that Linus Torvalds and the Linux project are NOT the maintainer of all those utilities. That was what I was getting at. Yes I know you can get utilities from wherever you’d like
I don’t remember the exact percentages, but I read an article several months ago which broke down various Linux distros and showed how, if you’re going by percentages, you’d at LEAST have to call it X11/GNU/Linux since the X.org constitutes at least as much code as all the GNU stuff on an average Linux system with GCC installed these days.
…but, seriously, GNU/Linux is NEVER going to catch on because it has too many syllables. Hell, a lot of novice users don’t even say “Linux” these days, just thinking “Ubuntu” is the more significant moniker. (Also, when people say Linux, their intuition is generally interpreting the name as an application platform, so X11/Linux would be more accurate.)
Personally, I’d just like to see someone put in the effort to build a fully-functioning desktop distro which replaces GCC with LLVM/Clang, glibc with something like uClibc, etc. so we can just cut these GNU/Linux whiners’ last leg out from under them.
You might try building Gentoo with those. It might take some doing, but from what I’ve read it’s doable (provided you only use packages without GNUisms). I know I’ll try sometime in the future
FreeBSD 10 might be such “distro” with PC BSD as a desktop version.
Whilst you’re not wrong, the post you responded to is clearly an accurate correction of the OP, who was spouting complete nonsense. GNU existed before Linux, it enabled Linus to create Linux, and thus when Linux first became a bit of a hit, it was the GNU OS with Linux as a kernel. Naturally, over the nearly-20 years since, a lot of GNU has since been made obsolete or diverged from its GNU roots, but I don’t recall anybody claiming otherwise.
Edited 2013-01-08 00:59 UTC
A name is not the credits at the end of a movie. It’s just a convenient way of referring to something. GNU/Linux is a cumbersome, awkward (and therefore bad) name. Linux is fine. And it’s more commonly used anyway.
Whatever Linux was at it’s infancy, Linux today is a kernel, which is a component of what we call an operating system. A very fundamental component, but a component nonetheless.
So:
Android is an operating system running a Linux kernel
Ubuntu is an operating system running a Linux kernel
Debian GNU/kFreeBSD is an operating system running a FreeBSD kernel
Darwin is an operating system running a XNU kernel
Windows is an operating system running a NT kernel.
Ironically your rant about Android has hit upon the very reason why many people do called desktop distros as GNU/Linux. Because you can then have Android/Linux, GNU/Hurd and so on. It makes the distinction across the different forks easier to describe concisely.
Plus, you’re overstating just how much of complete OS the early versions of Linux was. Even the earliest of versions was developed on Minix and compiled used GNU C Compiler. Linux was never anything significantly more than a kernel (not even the terminal emulator project that gave birth to Linux). Even before the renaming of Linux (originally Linus called it something like Freknix – I’m so glad the FTP host had better ideas!), Linus’s kernel and development was heavily dependant on the user land from other POSIX OSs. So I think it’s somewhat overstating to argue that Linux had a complete user land before RMS (reluctantly) adopted the kernel.
Except that’s not the distinction at all.
X11/Linux would be more accurate and, even IF you’re running console apps that are more constrained by their GNU-isms than their dependency on something like X11, that’s still glibc/Linux, not GNU/Linux.
“GNU/Linux” came about because Stallman draws the line between “operating system” and “extras” at the bare minimum you need to run a terminal with a Bourne-family shell and emacs… which means that X11 (which is the single biggest component by KLOCs on a “Linux” system) is an “extra” (and, therefore, it’s not “X11/GNU/Linux”) and it’s “GNU/Linux” because “it’s a GNU userland on top of the Linux Kernel.”
(He ignores the fact that hybrid embedded uses are becoming very popular and they often replace all the GNU userland except glibc with busybox while retaining binary compatibility.)
In fact, the article I mentioned previously showed that, if your system doesn’t have GCC installed, GNU isn’t a noteworthy component of modern Linux “by volume”.
Hell, My “X11/Linux” example actually IS how browser User-Agent strings do it. (“Linux; X11” on desktops; “Linux; Android” on mobiles)
If Android and DirectFB weren’t around, even X11/Linux would be redundant since it’s not as if there are any OTHER GUI subsystems in use on Linux that have browsers built against them. (The GUI mode of links2 can be compiled against DirectFB and the gpm console mouse daemon. It’s quite useful for Googling up solutions when X11 won’t start for some reason.)
Edited 2013-01-08 12:26 UTC
Oh I know how and why GNU/Linux came about and I’m well versed on the workings of Linux. And I do fully agree with the point you’re making, but sadly that’s all irrelevant as it’s a term GNU/Linux is now out there and recognised.
It’s a bit like how I think “cloud” is a dumb term, but people (nerds) understand it so it’s stuck. If I was to constantly refer to Facebook as a web2.0 re-imagining of TSS then few people would understand. However most computer literate people know what a “cloud” is.
Sometimes our search for literal correctness is held back by our ability to explain concisely to a wider audience.
Edited 2013-01-08 13:49 UTC
Laurence,
“Oh I know how and why GNU/Linux came about and I’m well versed on the workings of Linux. And I do fully agree with the point you’re making, but sadly that’s all irrelevant as it’s a term GNU/Linux is now out there and recognised.”
Meh, I benefit from both and I couldn’t care less what people want to call it. It’s rare that people cannot figure it out from the context. If non-gnu userspace linux kernel distros became widely popular, then there’d be a plausible case of ambiguity, but until then I don’t care about pedantry.
“However most computer literate people know what a ‘cloud’ is.”
I’m not convinced it means anything at all
Seriously it’s been used for SAAS, CPU Virtualization, outsourced file storage, web apps, and even for streaming music. It seems that anything running on the internet might be described as “running in the cloud”. Hey folks, now you can talk to your friends in the cloud – it’s called IRC. It’s all marketing, IBM has been pioneering “clouds” for decades.
That was largely my point about TSS
Fair point. I like that attitude.
Laurence,
“That was largely my point about TSS”
Haha, I’m afraid that I didn’t know what that meant.
Duck duck go helpfully suggested these definitions:
Toxic shock syndrome (fatal illness)
Total suspended solids (fluids)
Total sum of squares (mathematics)
Time Sharing System
Task State Segment
Transcription start site (RNA)
Tromsø Satellite Station
By the process of elimination, I see what you mean about facebook being a reimagination of Toxic shock syndrome
hahaha
If you hadn’t already guessed, I was referring to Time Sharing Systems: http://en.wikipedia.org/wiki/Time_sharing
It’s basically the fore-running principle behind UNIX’s conception and gave birth to the computing paradigm of many users on a single platform sharing resources – often connected via thin clients / dumb terminals.
The only commonality between all the different cloud usages which I’ve encountered was what I’d have described as TSS; shared centralised resources which typically (though not always) handle the majority of the data processing.
In some ways, and specifically with the raise of web apps, it feels like we’re going back to a 60s computing paradigm.
Edited 2013-01-08 15:59 UTC
I often hope that we, ~geeks, would get together and popularise another term: “fog”. Seems kinda more apt
“Fog” can be more accurate, perhaps…
(and I would also ask about TSS, but I see it was already asked & answered)
Sometimes when you don’t know what you’re talking about, it’s best to say nothing at all.
That way you don’t waste knowledgeable people’s time with nonsense, you don’t spread nonsense to those who don’t know, and you don’t look a fool.
Sadly it is too late this time around.
Better to remain silent and be thought a fool, than to speak out and remove all doubt. – Abraham Lincoln.
HappyGod,
http://ask.yahoo.com/20010115.html
” Who said, ‘Better to keep your mouth closed and be thought a fool than to open it and remove all doubt’?”
“This was a tougher nut to crack than we originally thought. Not only could we not find a definite answer to your question, we couldn’t even confirm the exact wording of the quote … Other pages suggested a number of other authors for the saying, including: Abraham Lincoln, George Eliot, Groucho Marx, Albert Einstein, and a mysterious figure named Silvan Engel.”
The only thing we can know for sure is that whoever said it was undoubtedly a fool
Oh boy. You have zero clue about this matter, don’t you?
Linux is just a Kernel So android is in fact “Linux” but sans GNU userland and tool Android/Linux would be a accurate name.
GNU/Linux is the Linux Kernel with the GNU tools and userland.
Lastly we could argue that Linux would not exist without GCC.
Yet another one infected with FSF propaganda. If you stop mindlessly copying stuff you ‘ve read in fsf.org, you ‘ll find out that in the previous posts we enstablished that Linux was an OS before it took any GNU code. Hence, it wasn’t “just a kernel”. It was an OS.
And the “wouldn’t exist without GCC”? Hilarious man. Not even Microsoft requires programmers that compile with visual studio to call their programs “MS/foo” or “VisualStudio/foo”
PS: Android (the new defense for the GNU/Linux nonsense) is a fork of Linux, that uses the kernel part of Linux only (and even that is modified).
I know Stallman acts super offended everytime someone says “Linux” to refer to the whole OS and he managed to convince you it should be true. I know Stallman says that Linux was “not usable by itself” (WRONG wrong wrong, Linux was BOOTABLE and usable before it took any GNU code). Sorry, Linux started as an OS, was an OS before it took any GNU code. Cheers
PS: What’s hilarious is that non-kernel non-GNU code like init still exists in Linux. So, calling Linux “just a kernel” is silly beyond comprehension.
Edited 2013-01-08 19:41 UTC
Essentially, Linus took massive amounts of GNU code to advance his Linux OS project without giving any credit to GNU, just like the GPL allows. Stallman tried his own medicine and doesn’t want to admit it’s bitter. For that reason, he tries to spin the facts by claiming Linux was a kernel project and not an OS project (never mind Linux had non-kernel code like code for booting the kernel from day 1, and hence was an OS project not a kernel project from day 1*). Essentially Stallman tries to convince people GNU used Linux, while it’s the other way around. Linux used GNU to become a better Linux.
Failing to do that, he just pounds the table and demands credit while his own GPL doesn’t have a give credit clause. Failing that, he just resorts to “Waaahhh!!! I did all that work that yiu took. Do me a favor and put GNU next to the Linux name”
*for people who still don’t get it: If it has kernel code only – kernel project. If it has kernel and booting code – OS project.
PS: Fun fact: Stallman started a whole crusade to remiove a give credit clause from the original BSD license.
RMS spent a lot of time fighting admins over computer time and access rights. So he wanted a system where he didn’t have to fight them and could just do what he wanted. PC’s kind of killed off that need for a while. But virtualization would have solved his problems as well.
Hurd hasn’t been ported to Coyotos, they ceased before even writing any code. Also, there is no “port” to Viengoos, as Viengoos isn’t even useable (btw, Viengoos’ main reason to exist is to be Hurd’s microkernel)
Well, the bottom line is that micro-kernels are hard to build and some efforts like HURD just lack the momentum to build a proper one. With Linux on board the GNU OS was a practical reality and there was no pragmatic need for HURD anymore. Thus the momentum was lost even further.
Efforts like MINIX are completely different. It’s an academic project and every year there are new students to pick it up. Also, I’ve met Tanenbaum once and he seemed like a really nice guy. 🙂
If the FOSS coders were at least a bit more competent into kernel development and not wasting their time forking for the 1000s time another port of a Windows software that Linux always lacked, it would ease things to evolve a bit. Just my two euro cents…
Kochise
Well, different people have different interests, so you can’t really force people to spend their volunteering time on a project which they don’t care about, development-wise.
That’s sort of like saying that, if all these fanfiction authors weren’t so obsessed with their Twilight and Harry Potter shipping, maybe we’d have the next great American novel by now.
Kernel-mode development requires quite a few competencies over and above user-mode development even if they’re done in the same language… not to mention, as a programmer myself, I can say that some developers just have no interest in low-level stuff.
Yes, I’ve written low-level stuff like a partial GIF parser to separate static and animated GIFs at high speed but I wouldn’t care enough to write a whole application in C, let alone a kernel. If I couldn’t use higher-level refcounted or garbage-collected languages like Vala and Python alongside ready-made libraries and frameworks like GTK+ and Django, I’d probably have never graduated from writing DOS batch files in Windows 9x and instead learned to do something like 3D modelling using a warez copy of 3DS Max.
Edited 2013-01-08 23:43 UTC
Then you cannot moan just because nobody dared to code Hurd if you’re ready to give a helping hand. It’s always easy to complain but pretend you cannot get the knowledge or you have little to no interrest in the matter.
Kochise
So those suggesting it to be dead are being ignorant.
Of course, it is marginalized (understatement) but it is still being worked on – just mostly as a research OS IIRC.
http://lists.gnu.org/archive/html/commit-hurd/2013-01/threads.html
Seriously, why bother?
Hurd would be in the same pool as Linux, but Linux is light years ahead. Even if Hurd becomes mature enough to be useful, why pick it instead of Linux? I’m afraid Hurd will have more developers than users.
Why not put in the effort in to Linux instead?
MOS6510,
“Why not put in the effort in to Linux instead?”
I target linux myself because of it’s market potential, but I’d be remorseful if we lost alternatives all together.
Linux works because it’s got tons of man hours going into it, but that doesn’t mean it’s always the greatest approach. Plan-9 was an OS designed with much more care going into well designed interfaces. The FreeBSDs are often leading the curb as well. Sometimes it is linux. I like the variety and wouldn’t want to end up having only a few mainstream operating systems.
Speaking of hurd specifically, I concede that I’m not familiar with it
but that doesn’t necessarily mean it doesn’t have merit in some way for those who have worked with it.
I’m sure someone somewhere can find some merit for it, but I wonder how many people can and what kind of basis does this merit have.
If it’s based on “I like it, just feels nicer, dunno” I don’t give this kind of merit much merit.
You listed a couple of valid choices, but it’s just hard to imagine Hurd taking up a valid place amongst them. Considering how long it has already taken I seriously doubt anyone could seriously have much faith in a happy marriage when going for Hurd.
I don’t know anyone using Minix, but even that is a complete and working system.
If people like to work on Hurd then good for them, but it’s a sizable project and I just wonder it the time and energy spend isn’t better spend on something like Linux. Linux users are, probably, their target audience anyway, but I don’t see Linux users leaving a mature and working system for Hurd.
Because Hurd is another approach to design micro-kernel operating systems:
– More security
– Not another UNIX clone
These two points alone are worth the further work some people might be willing to invest into Hurd.
Point one is valid, I guess, even though OpenBSD and even Linux are or can be made pretty secure.
Not sure what value point two has, it sounds more like preference than technical advantage.
An operating system can have its advantages and cool features, but you also need software and hardware support. A game console without games is pretty useless.
Linux, to pick one, has hardware support, software, documentation, “community” and a bunch of qualified experts. A company can implement Linux and be pretty sure it will work and if it breaks can find someone to fix it.
If Hurd comes alive my feeling is it will be much more difficult to make it useful and find experts on it. The ones you can find will probably be pretty expensive.
Is it worth all that trouble for the extra security and not being a UNIX clone? Organizations that require above average security usually have a lot of money. Why spend that on a “hobby” project and not get some hardcore Linux guru’s and use Linux, which has proven itself over the years.
MOS6510,
“Why spend that on a ‘hobby’ project and not get some hardcore Linux guru’s and use Linux, which has proven itself over the years.”
Yes, the majority do choose that path, which reinforces the linux platform. Isn’t it true that your criticisms of hurd would apply equally to any independent projects as well? Let me ask you, why does anyone bother building something different when there’s already technology on the market that’s been proven for years?
For me, the answer is that a technology landscape with the same players infinitum is boring… This is more of a mind exercise than a serious proposal, but I’d be extremely interested to see what we could build next if everyone just dropped linux and started from scratch with the benefit of the past 20 years of computers under our belts. What would be different?
Well, it’s okay to come up with new stuff, sure. But this is a kernel, meaning it’s very complex and when it’s done it’s supposed to power an operating system. To have hardware drivers, software running well on it, having people know how to use and support it takes a lot of time and effort, making me wonder if that will ever happen.
Linux started like this too, but I guess it was a bit lucky. It provided a cheap UNIX alternative to the expensive real UNIX systems while BSD was off the map for a while.
It seems to me the audience for Hurd is the Linux crowd. For one, only the Linux crowd ever mentions Hurd.
Of course it would be nice to see what an operating system/kernel would be if one didn’t make a UNIX clone and tried to make something that doesn’t reproduce the flaws, problems or odd things UNIX/Linux systems have.
But would it be enough to take the crown? AS/400 is different, users swear by it, but it’s nowhere as popular as Linux.
UNIX and the UNIX reincarnation Linux have been around for so many years, it must be a good way of doing things.
Apart from Hurd there are a number of small operating systems around. Yesterday I installed Icaros and I liked it a lot (mostly for nostalgic reasons being a former Amiga user), but I’d take OS X, Linux and Windows over it (in that order). I’m afraid Hurd will also become a member of that “fun to play around with, but not to actually use for real” category.
Inertia and easiness of porting existing code have a lot to do with it.
I doubt Linux would have ever picked up steam if some companies did not saw on it a cheap way to stop paying big bucks for commercial UNIXes while keeping existing code bases. Thus is was worth the investment of helping to improve the Linux kernel.
Yes, and I doubt if such a situation will arise for Hurd.
I fully agree with you, but that should not stop people to work on it, if they like to do so.
Oh, I don’t mind them doing it. If I google I’m sure I can find countless hobbies that are really much worse.
I just think their goals aren’t realistic, but I probably would have thought the same thing had I read Linus’ historic post.
The thing about IT and probably a lot of things in live in general is that great ideas don’t often work out, but strange enough bad ones more often become hits.
It was in the unix haters handbook “worse is better principle”.
You hate UNIX so much you bought that book? :-p
It is freely available online, and is somewhat of a classic.
Thanks for the tip! I’ll look it up. Not because I hate UNIX, but it’s free and I’m Dutch.
http://web.mit.edu/~simsong/www/ugh.pdf
They could have taken and extended Apple’s Mach as an interim solution. The bonus, IOKit. But they are not realistic.
If the FSF did that, I’d have to wonder if someone spiked my drink with something very potent indeed.
Actually, you do have a nice idea there. The Darwin codebase isn’t GPL compatible (not necessarily a bad thing imho), and it’d be quite interesting to take the Darwin base (Mach, Launchd, IOKit) and build something really complete out of it. Something with the technical advantages of OS X (at least the lower level parts of OS X) without all the Apple hardware lock-in and perhaps with GNUStep/Etoilé as the GUI. Only problem is, we’d need a graphics stack far better than X11 to keep it stable. But anyway, I’m diverging from the topic at hand. 
Gotta love the responses of the FSFies… Guys, the “Linux” name was meant for an OS project, not a kernel project, sorry. Linux clearly started an OS project, sorry. Why do you guys insist sticking the “GNU” ever? Linux just used GNU upstream to complete his *OS* project, just like the GPL says. I don’t get why he has to give credit because he used GPL code. On what grounds do the FSFies think he is obliged to give them any credit, when the GPL doesn’t have such clause? Stop trying to help the FSF extend the GPL at will by adding arbitary obligations.
BTW if Linus had stated he had started a kernel project calked Linux and hadn’t written any non kernel code, only then the FSF would be right.
PS: Android uses a fork of Linux which includes only the kernel.
Browser: Mozilla/5.0 (Linux; U; Android 2.3.4; el-gr; LG-P990 Build/GRJ23) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1 MMS/LG-Android-MMS-V1.0/1.2
BTW it’s called the “vote up/down” button, not the “agree/disagree with poster” button geniuses.
Browser: Mozilla/5.0 (Linux; U; Android 2.3.4; el-gr; LG-P990 Build/GRJ23) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1 MMS/LG-Android-MMS-V1.0/1.2
Exactly. And there is also a “reply”-button; why are you spamming so many threads?
Here is if you still don’t get it: If i use code from Handbrake to complete my video clip converter (assume i have clearly stated i want to make a video clip converter not a video stream compressor like Xvid), do i have to rename it to Handbrake/foo? Of course not. The situation of video clip converter vs video stream compressor isn’t any different than OS vs kernel.
Show me proof that Linus started off to make a kernel – not an OS and that he hadn’t written any non kernel code, and I will admit I am wrong. (good luck with that, we all know Linux had NON KERNEL code and was bootable before it received any GNU code. A kernel only is not bootable genius. Yet Linux was bootable before it received any GNU code. So, Linux was *an OS* before it received any GNU code. This makes the “Linux was just a kernel before it received GNU code” excuse of the FSF moot.
Browser: Mozilla/5.0 (Linux; U; Android 2.3.4; el-gr; LG-P990 Build/GRJ23) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1 MMS/LG-Android-MMS-V1.0/1.2
Edited 2013-01-08 12:29 UTC
Uhm, “reply”-button is for replies.
This is because Stallman is a honorary professor (whatever this is) and not a real professor. A real professor knows the “bare minimum” for “operating system” is the ability to boot and allocate resources. And maybe to interact with I/O or disks (not necessarily both). Linux met those requirements before it took any GNU code. Hence, Linux was an OS before it took any GNU code. Read my previous post.
Instead, the bare minimum for “kernel” is to “allocate resources”. No boot.
I am an undergraduate and even I know that!
PS: The “run terminal” thing is completely silly. Symbian OS didn’t offer a terminal with bourne yet it was an OS…
Edited 2013-01-08 12:41 UTC
There’s a keyboard shortcut in Firefox to incremental-search all links. You can look for “reply” to find where the “reply”-button is located.
And the only Free-Software-Trolling I see on this very OSnews article is yours.
The whole naming thing is silly, just call it Linux. Yes, technically that’s only the kernel, but we all know what someone means when he says he’s running Linux on his server or desktop and when it’s running on his phone.
Calling it GNU/Linux just wears your keyboard out more and it’s, well, kind of listing the spec/features of the operating system with no added value. Why not call it GNU/Linux/ext4 or GRUB/GNU/Linux/ext4/swapon?
To me it seems GNU/Linux is an attempt by Stallman to stick his name on it making it look like a team effort between him/FSF and Linus. While GNU plays a big part in most Linux distributions a lot of parties also do. Stuff like Apache, Python, Postfix, KDE. GNU makes a Linux OS work, but all those third party projects give Linux its real use.
Typing ls, grep, ps, etc… (pun intended) is rather useless without anything real running like a web or email server.
Stallman likes to take a lot of credit where he might have had some input.
I generaly try to avoid him and his opinions. Idealists are seldom, if ever, practical in their suggestions and sollutions. And he’s annoyingly unfunny in general.