GNU OS
- Anmelden oder Registrieren um Kommentare zu schreiben
I am a member!
Why do we not have a distribution packaged by the GNU Project, simply called GNU OS?
Probably the ongoing commitment of time and personnel, and not wanting to reinvent the wheel. http://www.gnu.org/distros/free-distros.html
I am a member!
The GNU project builds the components of an operating system, but until it combines the components into an operating system with the GNU branding, the project is not complete.
If we consider all those fully free GNU/Linux distributions, we realize we have enough money and personnel to make it happen.
Having an official GNU distribution wouldn't achieve anything, because the current free distros have the same aims as GNU anyway - to produce a completely free system.
I am a member!
"Having an official GNU distribution wouldn't achieve anything, "
It would accomplish the stated mission of building an operating system.
"because the current free distros have the same aims as GNU anyway - to produce a completely free system."
Then why not join forces?
Trisquel has already joined forces with the FSF and GNU projects. They share the same views and have similar policies.
In fact, if you go to the GNU.org homepage you will see a screenshot of Trisquel running OpenOffice.org.
I am a member!
There is a lot of cooperation between projects already. What is gained by merging the branding exactly? Something more tightly integrated? I'm not so sure even that would be true. We're all working together as it is. Trisquel works with ThinkPenguin. We're working with Atheros. Trisquel works with Ubuntu. Trisquel works with the FSF. And so on...
What your talking about seems to me to be more of a marketing / public relations move. I think there could be some advantages in regards to web development though and public relations. Merging resources to pay one person or small team would likly help. However I'm not so sure that there is the money to do that. The FSF probably has the money and a few of the most popular distributions. ThinkPenguin also has a little bit... that is probably about the extent of it though.
yes, running a GNU OS would be pretty awesome.
But, the main page of GNU says, the kernel hurd is not yet ready for daily use. Like Andrew said, nothing would be achieved by building OS on librelinux as there are already a lot them. Unless hurd is ready for current day usage, i dont think GNU OS will be built.
So, that should be the answer to your query.
Hope that helps.
I think systemovich may have been talking about an official GNU/FSF distro (correct me if I'm wrong).
I fully support the HURD project though. I think Debian GNU/HURD is an interesting project. But I probably won't use it until it works properly on my hardware. Linux-libre enables me to run free (but lacking WiFi firmware) and it's good enough for me.
The GNU Project distributing a new GNU/Linux distro would not do any good. Trisquel, Ututo, Parabola, and all the others sponsored by the GNU Project are already completely free software. The only thing that
adding and maintaining another distro would do is split the community more than it already is.
Not really.
The good things would be:
1. It could exist as long as FSF exists.
2. It could become the main libre distribution, effectively obsoleting other libre distributions, which in turn unites the shattered community.
At least two very good things exist. If very good reasons exist, then it is not in vain to do something.
> 1. It could exist as long as FSF exists.
There are many GNU packages with not enough developers and many distros
with not enough developers, how do you know it won't be one of them?
> 2. It could become the main libre distribution, effectively obsoleting
> other libre distributions, which in turn unites the shattered
> community.
Would it solve the problems users of other libre distros see in Trisquel
without introducing the problems Trisquel users see in other distros?
That sounds like quite the fantasy. There's a reason I gave BLAG Linux and GNU a try: my dad prefers Fedora over other distros like Ubuntu. I, on the other hand, can't stand Fedora and much prefer Ubuntu-based distros. Then you get into life cycle of versions and rolling release. Different people want different things.
I could volunteer for GNU OS.
I am a member!
The best way to create GNU OS is for an existing free distribution to give its copyright ownership to the FSF and change its brand to GNU.
Giving copyright ownership to the FSF will ensure that GNU OS will always be free. GNOME started out with the freedom mindset, but has become open source, even though it is supposed to be a GNU project. The FSF cannot fix GNOME because, as far as I understand, it does not own the GNOME copyright.
Trisquel seems to be the prime candidate to morph into GNU OS, because it is used by the FSF.
> The best way to create GNU OS is for an existing free distribution to
> give its copyright ownership to the FSF and change its brand to GNU.
Seems pointless or impossible when most software in a distribution isn't
copyrighted by its developers (it distributes software written by
others).
> Giving copyright ownership to the FSF will ensure that GNU OS will
> always be free.
Licenses like the GNU GPL are designed to do this, even if the copyright
holder later decides to make newer versions nonfree (if there are
multiple independent significant copyright holders then it might be
difficult to get their permission to make a nonfree derived work).
> GNOME started out with the freedom mindset, but has
> become open source, even though it is supposed to be a GNU
> project. The FSF cannot fix GNOME because, as far as I understand, it
> does not own the GNOME copyright.
Explain how copyright assignment could help in such a case. (It's
interesting for me as a contributor to a different non-FSF-assigned GNU
package.)
> Trisquel seems to be the prime candidate to morph into GNU OS, because
> it is used by the FSF.
They also use other distros for significantly different uses, like
gNewSense or LibreWRT. How would Trisquel replace them?
----- Mensaje original -----
> > The best way to create GNU OS is for an existing free distribution to
> > give its copyright ownership to the FSF and change its brand to GNU.
>
> Seems pointless or impossible when most software in a distribution isn't
> copyrighted by its developers (it distributes software written by
> others).
>
> > Giving copyright ownership to the FSF will ensure that GNU OS will
> > always be free.
>
> Licenses like the GNU GPL are designed to do this, even if the copyright
> holder later decides to make newer versions nonfree (if there are
> multiple independent significant copyright holders then it might be
> difficult to get their permission to make a nonfree derived work).
>
> > GNOME started out with the freedom mindset, but has
> > become open source, even though it is supposed to be a GNU
> > project. The FSF cannot fix GNOME because, as far as I understand, it
> > does not own the GNOME copyright.
>
> Explain how copyright assignment could help in such a case. (It's
> interesting for me as a contributor to a different non-FSF-assigned GNU
> package.)
>
> > Trisquel seems to be the prime candidate to morph into GNU OS, because
> > it is used by the FSF.
>
> They also use other distros for significantly different uses, like
> gNewSense or LibreWRT. How would Trisquel replace them?
I thought this was talking about the gnu herd!
gnu.org/software/hurd/
it is a microkernel? that was being produced for gnu software. Seems linux (monolithic? Kernel ) became mainsteam before they got it to work. But seems to be working now!
But to the subject at hand:
i like fragmentation. It means there is a free market and you have choice. I do not want only one brand of cola (coca cola) nor two (pepsi-cola) nor three (ubuntu-cola) but four (whole earth organic cola) and five (fentimans curiosity cola). (i am sure there are more)
I consider it my job to recomend products i like (fentimans curiosity cola is great by the way!). I would not like to say that we will no longer have many different flavours but they all got toguether and did one single flavour. And also i do not think they will.
hmmm, thats a first i heard of this idea.
+1 for out of the box idea.
But, i dont think you took my point about Hurd. Hurd is part of GNU project. Linux is not.
If you use Linux kernel, it will be just a GNU/Linux OS that will be created. Not, a pure GNU OS.
> But, i dont think you took my point about Hurd. Hurd is part of GNU
> project. Linux is not.
> If you use Linux kernel, it will be just a GNU/Linux OS that will be
> created. Not, a pure GNU OS.
https://www.gnu.org/gnu/manifesto.html explicitly calls TeX "our text
formatter", would such a "pure GNU OS" support printable documentation
of GNU programs? (Linux-libre is a GNU package.)
Hurd has useful technical features not available or not as integrated in
Linux, there are many other reasons to use Linux-libre now like device
support. (Including Linux code under the GPL version 2, Hurd doesn't
have licensing advantages.)
Weird that you are upset of having an OS with Hurd Kernel.
Everyone knows Libre-Linux is now a GNU package. Did you know it was non-gnu before FSF Latin America joined in? And, Linux-libre = Linux minus Shackles(non-free). So, it is directly dependent on Linux.
btw, you are yanking the wrong chain, as i see you are whipping out debate on Hurd vs Linux. Query was about "Why there is no GNU OS rolled out".
If GNU project is working on their own kernel, then they must be doing it because,.. ummm, [1]maybe they want to use it in GNU OS, or.. [2]just for fun.. nah, i think its the first one.
> Weird that you are upset of having an OS with Hurd Kernel.
We are discussing promoting a single OS over all other ones, this might
imply choosing a single kernel.
> Everyone knows Libre-Linux is now a GNU package. Did you know it was
> non-gnu before FSF Latin America joined in? And, Linux-libre = Linux
> minus Shackles(non-free). So, it is directly dependent on Linux.
IceCat is another example of a GNU package dependent on an upstream
nonfree program. (There was also an important GNU package with a
maintainer having very anti-FSF views, seems similar.) Hurd also is, it
has drivers from old Linux and software to use new Linux network drivers
in userspace.
> btw, you are yanking the wrong chain, as i see you are whipping out
> debate on Hurd vs Linux. Query was about "Why there is no GNU OS
> rolled out".
Somehow these debates usually lead to Hurd vs Linux.
> If GNU project is working on their own kernel, then they must be doing
> it because,.. ummm, [1]maybe they want to use it in GNU OS,
> or.. [2]just for fun.. nah, i think its the first one.
The motivation for writing a kernel is much more complex. (Technically
Hurd was chosen so that they could use an existing and working Mach
kernel instead of writing and debugging their own.)
For a man of so many words, you are yet to answer the query of OP.
"Why do we not have a distribution packaged by the GNU Project, simply called GNU OS?"
> "Why do we not have a distribution packaged by the GNU Project, simply
> called GNU OS?"
As you answered: there is no need for it when we have free
GNU/Linux-libre distros. Unless we discuss the meaning of words we use,
or Linux vs Hurd technical features, there probably isn't more to
answer.
There's certainly some would agree that a fresh approach is needed -
http://www.theregister.co.uk/2012/11/09/opensource_festering_hacks/
If they can make Hurd clean, lean and contemporary...there's real potential there. From the look of those pages on GNU/Hurd it seems to be progressing very slowly though.
The GNU project has always preferred to reuse available free components instead of recreating them. That is why GNU does not have a display system (Xfree was available under a free license) or a text processor (TeX was available under a free license). That is also why the effort on the Hurd was weakened a lot once Linux was released under the GPL license and proved to be a better solution a few years after its beginning (debugging, hence developing, micro-kernels is very hard).
All that to say that a 100% GNU distribution would not only inefficiently work on very few systems (because of the Hurd) but it would also be very incomplete. The graphical GNU packages would not even be available because GNU does not have a display system! Who would prefer such a system to Trisquel?
If it looks obvious that a non-GNU display system (Xorg or Wayland) is to be used, why not also preferring Linux-libre to Hurd, i.e., what is the argument to prefer a GNU package to another *free* alternative that is technically better?
I am a member!
Yea- the only reason I can see to use Hurd over Linux-Libre is technical. I'm not sure what those technical advantages are. The Linux-Libre are well known though (hardware support). From the free software perspective it doesn't matter which (free) kernel you use.
> Giving copyright ownership to the FSF will ensure that GNU OS will always be free.
The GPL was designed to ensure that users of GPL-licensed software would always be free.
> GNOME started out with the freedom mindset, but has become open source, even though it is supposed to be a GNU project.
Sadly that is true. But I don't think the GNU project can force their views onto everybody. If the GNOME developers chose to use that term we can't force otherwise. We can encourage them to use the term "Free Software" though.
> The FSF cannot fix GNOME because, as far as I understand, it does not own the GNOME copyright.
But even if they did own the copyright to GNOME, the GPL gives the GNOME devs the freedom to then modify and then share a fork of GNOME.
> The best way to create GNU OS is for an existing free distribution to give its copyright ownership to the FSF and change its brand to GNU.
Most of the software in Trisquel is written by people that aren't part of the Trisquel project at all. For example, LibreOffice, Abrowser (Firefox), X.org, Linux patched by FSFLA, etc. There wouldn't be much copyright to transfer because Ruben and other Trisquel maintainers haven't written a significant part of it. It is all licensed by others.
GNU asks developers to transfer copyright ownership so that they can stop large businesses from distributing it without the source code and preventing users from running freely. That's the only reason, as far as I know.
I am a member!
The aim of GNU is to provide a complete free operating system.
There are some very old images of GNU but there won't be any official public release until the Hurd is ready.
Old images (2006): http://i-hug.sarovar.org/downloads/GNU/extended/
Some info: http://www.gnu.org/software/hurd/hurd/running/gnu.html
I have actually years ago installed GNU Hurd. Good times lol.
Edit: I did write on the CD "GNU". I still have that CD somewhere.
I am a member!
From what Bazzy pointed out it looks like it is a resources issue. Why bother packaging for users what already exists elsewhere? Which is sort of obvious anyway and consistent with the aims of the project.
Someone said good things below are just a fantasy.
* * *
The good things would be:
1. It could exist as long as FSF exists.
2. It could become the main libre distribution, effectively obsoleting other libre distributions, which in turn unites the shattered community.
* * *
I remind that GNU project was a fantasy. Linux was a fantasy. Sage was fantasy. Everyone who believed in those fantasies has been blamed to be crazy and out of mind.
Someone said fragmentation is good because there is diversity like in cola drinks. That is, however, apples and oranges analogy. There is no bugs to fix in cola drinks and it is great if there is sustainable business for 1000 different cola drinks.
But a cola drink have nothing to do with OS. It is bad if there are 10 different sucky libre operating systems, each of them failing to fix bugs, when there could be one main libre system which could be developed to be superior derivate.
Why should there be 10 extreme minority libre distros, even if they would manage to fix bugs. Why to fix the same bugs in 10 different libre distros, when bugs could be fixed in one distro?
If something, then maintaining 10 different super minority libre distros is absurd waste of resources, and reinventing a wheel is genius when compared.
Distributions assemble software packages and there are no two identical distributions. Most of the packaging is done by the mainstream distribution and reused by the derivatives. Bugs in the applications are fixed upstream. In the end, there is little duplicate work.
Who said there is two exactly identical distributions? The point was that instead of actually improving the libre derivate distribution, people spend time in fixing freedom bugs from 10 different libre distribution. How is that not duplicate work? Another issue is different non freedom bugs, which must be fixed for a distro to become good. It doesn't matter are they similar between distros because they have to be fixed for each distro. So, actually, if they are different bugs like you said, then that just means more work because the solution to fix them does not work from one distro to another.
Now, lets go more abstract. I am sure you can handle this.
Let there be Y distributions with X bugs in each. Z people will be working on X * Y bugs, instead of Z people working on X bugs.
Understand?
What is funny in all this is that people say that it is not worth it to do another libre distro, yet they say it is good to have 10 libre distros.
Double standard much?
Could you chose how it is, or is it so that 10 distros is The Holy Amount of distros xD
Anyway, have had enough of arguing about this here. In the first place I just wanted to say my opinion without beginning to argue about it.
You are free to have your opinions, they can be different than mine, and we can still get along, I think.
I perfectly understand you... but I persist: you are wrong. The "freedom bugs" are usually solved by not including a package, what is pretty easy. Most of the work is about *identifying* the software not respecting the users' freedoms. That can be tricky but this work is "mutualized", not duplicated. See this page for instance: http://libreplanet.org/wiki/List_of_software_that_does_not_respect_the_Free_System_Distribution_Guidelines
It takes as a base Debian's split between free and unfree software and lists 89 adjustments with rationales.
If you are atually talking about the projects that fix freedom issues in upstream projects (instead of just not including the applications), they mainly happen upstream too: all free distribution use Linux-libre, all of them can use GNU Icecat. Those are the two only freedom-fixing projects I know of. You can answer that Trisquel does not use Icecat and you are right. It could though. It is a choice.
I have not seen anybody objecting to an eleventh free distribution. People object to the desire for "the one true" distribution. Such a distribution would not attend every need fulfilled by the current free distributions. For instance, you would not have ABrowser but only Icecat even if Ruben has valid arguments to prefer not to reuse Icecat (waste of memory, of space on the ISOs, etc.). Take Musix as another example. Do you really believe that the one true distribution would fulfil the needs of Musix's users (including a real-time kernel)?
I have been talking about /main/ libre distribution. Imagine it like Ubuntu, Debian, or Fedora, which have tons of derivatives.
I haven't said there should be only one libre distribution.
And you are wrong for saying there should not be one superior (main) libre distribution. It is like saying there should not be Fedora, Debian, or Ubuntu but instead a mess of distributions.
It is good to have a shining star leading the way.
Imagine it like this:
a) A mess of libre distributions.
b) One main libre distribution bettering a mess of libre distributions.
Questions:
1. Which of the above situations describes best the current situation?
2. Which of the above situations would be better?
I do not think a distribution can/should be started and declared a "main" free distribution. Such a thing may happen though. For instance someone who wants to create a free distribution for different needs (or simply having a different vision of what are the best desktop applications) can take Trisquel as a base. When it started, Ubuntu did not really aimed to be the base for derived distributions.
Looking at the current 100% free distributions, such a main distribution would not have decreased the amount of duplicate work (which I still believe is not a significant problem anyway): Trisquel is based on Ubuntu, Parabola is based on Arch, gNewSense is based on Debian, BLAG is based on Fedora and the remaining distributions do not seem to be based on any major distribution.
In other words, developers disagree on what is a good base. Just pick Debian (resp. Arch) as a base and try convincing Parabola's (resp. Trisquel's) developers to base their distribution on it! The same would happen with a 100% free Debian-like (resp. Arch-like) "main" distribution.
So, to answer your question:
1. yes, the current situation is a mess but...
2. ...no, developers prefer this mess to a common base that would be imposed to them (they would simply stop working on GNU/Linux distributions).
Well, there was a libre distro based on Slackware too. Talk about choise!
In my opinion it is very good to have libre derivatives of popular distros.
Maybe it takes another 5 years of evolution for one libre distro to get enough users and developers to be the leader of free distros, bettering the rest including the original.
I think 5 years ago there were no libre distros. Now we have 9 FSF approved. In my opinion it is progress, even though it is a mess.
Maybe the next step could be that one emerges above the rest. Or maybe the next step is that there will be three shining star libre distributions. The latter being hugely optimistical thinking.
If after 5 years the situation is the same it is now, then one maybe could argue that no progress has happened in 5 years.
GNU OS, in my opinion, should be released someday. Why it would not be, so to say, today?
-_-
One True Distribution would by definition fulfill whatever needs.
Edit: Unless you are paranoid and fear the perfection xD
The point that I think that is being overlooked, that it is not the number of libre distros but the number of package management systems (debian, arch, gentoo etc) with each having its own following, how would you produce a single distro and keep everyone happy/onboard? at least the current situation allows people to use the desktop/package management they are happy with while still supporting the 4 freedoms, perhaps if all libre distros used the herd kernel when it is ready would keep everyone happy?
just my 2 cents
I can see where systemovich is coming from on this idea. It is stated in many places that the purpose behind the GNU project was to build a free operating system. I find it a bit strange that they got everything done, then their kernel took longer than expected, but instead of building a system using the Linux kernel they left that up to other people to package it all up as distros.
I can understand what happened and why but we never have come to the actual GNU Operating system.(though it could be argued that GNU/Linux is technically the GNU Operating System).
At the point we're at now though, it almost seems counter productive to offer something straight from the GNU project. It would be competing with itself in terms of having its own distro and then a bunch of others it endorses. It would be a bit like the big auto makers owning so many brands that they're spending resources competing with themselves.
It seems that the GNU project's resources are better spent elsewhere. As long as good distros like Trisquel exist that is.
- Anmelden oder Registrieren um Kommentare zu schreiben