Talk:Linux/Archive 34

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 33 Archive 34 Archive 35

GNU project is a main contributor in "Linux" operating system

An average "Linux" Distro has more GNU software than any other pool of software under a single project. Alaukikyo (talk) 16:55, 17 September 2011 (UTC)

If I'm right, however, GNU is not the writer of the core OS. Jasper Deng (talk) 17:11, 17 September 2011 (UTC)
We seem to be tediously revisiting Richard Stallman's agenda that the word "Linux" should never appear unless "GNU" is somehow pasted onto it, preferably always as the prefix (and better still, in larger type, I suppose.) It's an unnecessary ownership claim that long been settled by consensus here on Wikipedia. Stallman (and anyone else) is free to proclaim that the whole thing was his idea but we're not here to speak in his voice. As an open source project with many contributors, to the extent anyone is singled out for credit, that simply has to be Torvalds for his initial authorship of the Linux OS kernel, which is at the core of what this article is about. There are a lot of Linux boxes out there being used by people who have no idea it comes with gcc, bash and gawk. Pasting GNU in front of Torvalds's name in the credits for the OS is just silly. Msnicki (talk) 18:20, 17 September 2011 (UTC)
A kernel is useless without the rest of the software that makes a computer useful. The first Linux distros were Linus' kernel and GNU software for the rest of it. Linus wrote his kernel, because it was the missing piece of a GNU complete system - the only piece that the GNU project was still working on. The consensus here on Wikipedia that Msnicki claims to exist does not. Please argue based on facts and history, and not your unfounded personal beliefs. Stallman argues for the joint name for the whole system, not the OS kernel. He always credits Linus for that. Part of the problem here is that the phrase operating system is murky - does it mean the kernel or the whole computing system? Perhaps what is needed is clearing that distinction up, and stopping the use of a murky term. Core OS is another murky term. In this case is it's just Linus' kernel, or does it also included the C, C++, etc. system libraries that the GNU Project wrote? The kernel is useless without the libraries. Lentower (talk) 18:42, 17 September 2011 (UTC)
If you think the term operating system is murky, maybe you should read the article. The OS is the part that runs below the application layer. Most of the GNU stuff runs at the application layer. Msnicki (talk) 20:56, 17 September 2011 (UTC)
The question is what fraction of the code below the application layer is GNU code from system libraries, etc. The term operating system could be further disambiguated in this article and elsewhere on WP, probably being replaced by clearer language. It's yet another case, where specialists are writing for themselves instead of working towards the best way to explain things to the everyone on the planet. Lentower (talk) 23:19, 17 September 2011 (UTC)
The answer is, very little. OS internal code tends to manipulate only OS-internal objects using OS-internal routines. You won't see a lot of calls to printf. Even C runtime library routines like malloc are usually replaced by OS-internal versions inside an OS. Msnicki (talk) 23:38, 17 September 2011 (UTC)
Sad that Msnicki has so little experience, assuming he is a system designer or programmer at all. I've been involved with the design of both Unix and non-Unix systems, both real-time and general purpose, for over 40 years. What he's said may be true in his experience, but it's not true in general. Lentower (talk) 23:53, 17 September 2011 (UTC)
The joint system of GNU/Linux is not the topic of the article, and is a subset of the distributions covered here. Note that the GNU Project is heavily covered in History of Linux (and the summary history section here) as they should be.
Most distributions use a large amount of code written for/by the GNU Project - but the only common factor between all of them is the kernel so that is what is mentioned in the infobox. KDE, Qt, Firefox, Pango, Python, Perl, Freetype, FFmpeg, etc. are also key parts of many distributions but are also not common to all of them so they are not mentioned in the infobox. If you want to try to divide up effort by lines of code, hours of work, monetary investment or some metric like that, the Linux kernel is sure to dwarf everything else. I wouldn't be surprised if Red Hat and Intel have written more code than those working under the banner of the GNU Project but I don't see any reason why we should replace "Many" with our own POV views of who deserves credit for the accomplishments of the entire community.
See Linux#Embedded_devices for some examples of operating systems which do not use glibc (they may use uClibc, and Android uses Bionic), and use BusyBox in place of the GNU userland. There are also "desktop" distributions like stali which don't use much GNU project code but instead use for plan9/bsd userlands.
This page used to be a Good Article and we should concentrate on getting it back to that quality instead of wasting time bikeshedding. I'd like to point out that the person who started this topic has resorted to vandalism to push their POV on this topic.[1][2][3] -- strcat (talk) 20:26, 17 September 2011 (UTC)
strcat no, his edits were NOT vandalism, but a result of the lack of consensus among editors on this topic. Making charges like you do here, does not help improve consensus, or improve the article. Lentower (talk) 23:19, 17 September 2011 (UTC)
I gave 3 links to what I consider to be vandalism: removal of all content from two templates used on both Linux kernel and this page. The templates are purely technical information (they just contain the current kernel version number). We will just have to agree to differ on whether the changes were made in good faith. strcat (talk) 00:57, 18 September 2011 (UTC)
strcat writes "If you want to try to divide up effort by lines of code, hours of work, monetary investment or some metric like that, the Linux kernel is sure to dwarf everything else." strcat is wrong. In system design, a major decision is how much functionality is placed in the kernel, and how much is in the system libraries. The decision when Unix was designed be Dennis and team at Bell Labs, was to minimize the amount in the kernel, and maximize the amount in the system libraries. All Unix derivatives that I am aware of continue this split. Application level programmers on Unix systems seldom directly use a kernel call, instead using system library calls. The code is hard to write and maintain regardless of which side of the line it's on. And most variants of Linux, and the bulk of Linus systems deployed, use system libraries written by the GNU Project. Lentower (talk) 23:33, 17 September 2011 (UTC)

If this article is just about the Linux kernel, the attribution in the Infobox, is fine. If it is about the kernel and the system libraries, the GNU Project needs to be listed there. If it is about those and "userland", BSD, and perhaps others, need to be added as well. Otherwise, this article is not describing the subject as it is, but according to the personal biases of a set of editors. Lentower (talk) 23:44, 17 September 2011 (UTC)

Len, on your user page, you identify yourself as "one of the founders as well as a founding member of the Board of Directors of the Free Software Foundation (FSF) in 1985." Don't you think this conflict of interest could be getting in the way of a neutral point of view? I really don't know what other reason you might have for your earlier ad hominem argument, questioning whether I'm really a programmer. Who cares what I am. For all you know, I write compilers and diff utilities, too. But I'm anonymous and anyway, personal experience is irrelevant. Msnicki (talk) 00:15, 18 September 2011 (UTC)
Msnicki: I raised two issues, you raised a third. Len-1) You made a technical statement of fact that is wrong. Len-2) I speculated on why you may have written a falsehood. You're right, that speculation isn't all that relevant. You have my apology for the fact I wasn't clearer that I was speculating about why you were writing falshoods. It could be based on bad writing here on WP, or almost anything. Comment But that you are making false statements, in the discussion about an article, is what's important. Please stick to the facts. Msnicki-1} I'm careful about COI issues. Have you made any effort here to WP:AGF? I'm not the only person here, or wider, who sees that the way this article is written mis-represents the subject. We owe our readers better. Lentower (talk) 00:42, 18 September 2011 (UTC)
I think I'm right. From Malloc#In-kernel: "Operating system kernels need to allocate memory just as application programs do. The implementation of malloc within a kernel often differs significantly from the implementations used by C libraries, however. ... In Linux and Unix kmalloc and kfree provide the functionality of malloc and free in the kernel." Msnicki (talk) 01:52, 18 September 2011 (UTC)
Guess the best we can do is agree to disagree. Other editors will form an opinion or not. It really takes several graduate level computer science courses, or equivalent work with experienced designers. And sigh, the Malloc article you link to has WP:OR issues, poor citations, the need for better and more citations, too much dependence on code examples, etc. And touches on a tiny corner of the issue without even beginning to understand it. Just read Malloc, and did some low-level wiki-gnoming on the way. Lentower (talk) 02:28, 18 September 2011 (UTC)
The infobox lists the developers as "Many" and the userland as "Various". Can you please elaborate on why that is biased? Linus is just listed as holding the trademark on the term "Linux". Alaukikyo removed that text from the article without an edit summary or any discussion, so I reverted it. That seems to be the cause of this whole kerfuffle, so I think we should talk about the real issue. Alaukikyo also made this edit, which was also reverted because there are far too many major contributors to list them all, and not all distributions use glibc and the GNU userland, making it untrue for some of them. What do you propose we do with the section on Embedded Linux (which includes Android and webOS)? strcat (talk) 02:46, 18 September 2011 (UTC)
strcat The last three paragraphs include the more direct answers to your questions.
The central question is what's best for all readers of Wikipedia. How to get the core of Linux across to all readers, and leave the details to be found elsewhere (with citations, externals links, (both carefully chosen) and search engines we aren't leaving our readers hanging). This is hard for most editors, because most of us contribute to articles we know a lot about, and topics we care about. And too many editors want to be cheerleaders and supporters of their topics first, instead of providing what the reader deserves from Wikipedia.
One source of guidance on this specific question, is what is done with this Infobox in other articles. I just looked at a dozen or so, and found that the OS Infobox is either not used at all, or the parameters being discussed here aren't included at all. The later points out, that just because a template has a parameter, doesn't mean it has to be filled in. The choice of which parameters to leave empty is editing. This avoids bloat, as well as increasing the quality and the information density per word. Another point is the more wider an OS is used, the fewer parameters need to be included in the Infobox. Quality editing and writing involve careful choice and is hard.
As a reader of Wikipedia, what I want is an article, that is tightly written and concentrates on the big picture. This includes how the topic connects to the rest of what is known, and how the topic fits into history and the world.
This article does too little of that, and too much of trying to track the current state, and the tiny twists and turns of recent development.
This is why, if this articles lists developers at all, the originals developers of Linus Torvalds and the GNU project are notable. I wish the OS Infobox has an "|Original developer" parameter. Linus Torvalds would not have written the kernel, if the GNU Project hadn't had the rest of the Core OS ready. How Linux happened and why is of broad interest. (The trademark doesn't need to be in the Infobox at all - I'm about to put it in the Table of Contents.)
But what best for the general reader is to delete: "|developer=", "|marketing target=", "|ui=", "|userland=", "|prog language=", and probably "|programmed in=". And possibly others. Too much detail for the general reader. Along with similar tightening in the article's text. I'll let this change wait until there is an opportunity for research and discussion by others.
And as I wrote earlier, very careful use of murky terms and phases with too many meanings. Lentower (talk) 19:47, 19 September 2011 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── A few things to clear up:

  1. "The joint system of GNU/Linux" is precisely what this article is about. Most people call that system Linux. The Free Software Foundation and a relatively small part of the extended fanbase (along with anyone who wishes to cooperate with the FSF, as they refuse to cooperate with those who don't use their terminology) uses the FSF's ex-post-facto rebranding, but in the interests of NPOV we are obliged not to treat this minority opinion as factual.
  2. How much of a given distribution's code is © FSF is irrelevant. We use names based on how common they are rather than how accurate they are, so even if one were to take the FSF's position on "accurate naming" there's still no ground to refer to the subject as "GNU/Linux" here.
  3. The next time Alaukikyo (talk · contribs) flogs this particular WP:DEADHORSE he'll be reported for edit warring. This lame POV-pushing is ineffective. Stop it.

Chris Cunningham (user:thumperward) - talk 14:04, 3 October 2011 (UTC)

Edit request: remove kernel version number

{{edit semi-protected}} It seems a bit hypocritical, if you don't mind me saying, to claim that "Linux" refers to any operating system environment using the Linux Kernel, but then to claim that "version" of any Linux operating system environment is the version of the Kernel itself. That belongs in the Linux Kernel article, not here; the two ideas are mutually exclusive. Zbuhman (talk) 23:21, 21 October 2011 (UTC)

I don't think they are "mutually exclusive" but I would remove such detail that's not needed here if I know how, hope other editor will jump in... man with one red shoe 00:14, 22 October 2011 (UTC)
I'm not sure what edit is being requested; could we get more specific about that? Also, I'm thinking that Zbuhman might be able to edit this article by now. – Luna Santin (talk) 00:50, 22 October 2011 (UTC)
This template may only be used when followed by a specific description of the request, that is, specify what text should be removed and a verbatim copy of the text that should replace it. "Please change X" is not acceptable and will be rejected; the request must be of the form "please change X to Y". - so, please re-state, and re-request (if applicable). Thanks,  Chzz  ►  07:25, 22 October 2011 (UTC)
Let me do the request then, remove the Linux kernel version from infobox. Clear enough? man with one red shoe 14:06, 22 October 2011 (UTC)
I apologize, I'm obviously not very experienced with this sort of thing. Zbuhman (talk) 08:04, 29 October 2011 (UTC)
  • Oppose. The Linux kernel is the thing that makes it Linux, not just a bunch GNU tools running on Windows or something else. We commonly identify the latest release in articles describing software products and the kernel version is the most meaningful way of doing that here. The infobox makes clear that this applies only to the kernel, not anything else. Msnicki (talk) 16:13, 22 October 2011 (UTC)
    • Reading the Wikipedia article on what an "operating system" actually is, is in fact being defined there as simply a "set of programs"; how is that any different from "a bunch of GNU tools". And again, why do we have an infobox talking about "The Kernel" in an article that is, as you said, not about "The Kernel". — Preceding unsigned comment added by Zbuhman (talkcontribs) 08:04, 29 October 2011 (UTC)
That's pretty clearly not what I said. If you'd like to argue with my points, it'd be helpful if it was more apparent you understood them. Msnicki (talk) 14:36, 29 October 2011 (UTC)
  • Support It's too much detail that's not needed in this article, what does the kernel version have to do with the article? The article is about the generic Linux OS, it needs no version number since there's no version number for the OS, various distributions that do make the full OS have their specific versions Ubuntu 11.10, Fedora 14, and so on. Since this is about the generic OS a specific version of the kernel or other parts of the system are not needed to be specified in the Infobox, especially that no distribution runs on the most recent kernel. I would also claim that more relevant for users is the version of the desktop environment, whether they run GNOME 3 or 2.x or KDE 3.5.x or KDE 4.x than the kernel version. -- man with one red shoe 22:51, 22 October 2011 (UTC)
The Linux operating system has version number. Latest is 3.1.4. You can find OS version numbers from kernel.org or using a uname program to give it to you with parameter -r. But the GNU uname has falsefied the IEEE standard to falsify the truth of Linux operating system (and same time falsefied technical side of other OS's as well). Check more from here (what has been countered later back to GNU propaganda: http://en.wikipedia.org/w/index.php?title=Uname&oldid=277002719 The operating system version is the kernel version when it is about monolithic operating system like what Linux is. If the operating system is Server-Client by architecture, then it has separate version number than what the microkernel or the servers has. There are two OS architectures and it is enough to cause misunderstanding for most of the tech people who do not understand operating system technology. Golftheman (talk) 15:00, 5 December 2011 (UTC)
  • Oppose* Any Linux operating system is primarily defined by its kernel, and I don't see any particular reason to not include the kernel version no matter how insignificant the version has become for most aspects of Linux operations or programming. In any case, the kernel version number is a noteworthy milestone identifier in Linux history. Kbrose (talk) 23:33, 22 October 2011 (UTC)
    • "Primarily"? To the contrary, a Linux operating system is in fact primarily defined by the version number assigned to it by (what is usually) the distribution packagers. Zbuhman (talk) 08:04, 29 October 2011 (UTC)
  • Oppose We settled this ages ago by prefixing the LSR/LPR subtemplate output with Kernel. Readers actually expect to find the Linux kernel version number in the Linux article as the name Linux is considered to be synonymous with the Linux kernel itself. The Linux kernel is also obviously an integral part of a Linux distribution and this article fills the gaps between the Linux kernel and Linux distribution articles.

    I could be off the mark, but I also find this request by User:Zbuhman just a little suspicious. The account has a creation date of 7 March 2011 with the account's first edit only occurring 7 months later on 15 October 2011 and this edit request being made on 21 October 2011. I think some of this is a little too "coincidental" given that I just created this subpage in my userspace on 20 October 2011, which happens to link to some of the May 2009 discussions... Given the background, perhaps a SPI is in order? --Tothwolf (talk) 17:12, 23 October 2011 (UTC)

    • That discussion was never resolved. Why is a template clearly talking about "The Kernel" being placed in an article about "Linux"? As for the rest of your "argument"; I have no idea where that came from, or, more importantly, why that's relevant. Zbuhman (talk) 08:04, 29 October 2011 (UTC)
Your question asking why the kernel version number is useful has been answered several times and I think there's a consensus. I don't think the problem is that you didn't get a good answer. I think the problem is that you just don't like it. Msnicki (talk) 14:42, 29 October 2011 (UTC)
But it's not useful; one could take an ancient userland and toss the absolute latest kernel in there; there is not necessarily any relationship between the version of the kernel and the version of the rest of the software present in the operating system. Zbuhman (talk) 16:55, 29 October 2011 (UTC)

Proposal: new definition what is Linux

Hello all

I've been thinking about the term "Linux" and its definition provided by Wikipedia. I see the current definition is very limited and it pays too much attention on details like for instance "Unix-like computer operating system" or "free and open source software development and distribution". Also statements referring to operation system kernel in the beginning of the article are perhaps little confusing. Somehow it would be better for Linux to understand it as a technological enabler rather than only operating system.

Therefore I would like to connect more general point of view than the current technologically orientated definition. The aim is increase the popularity of Linux, one aim to reach this is provide less technical definition for broader audience.

Based on those reasons, I suggest following for defining Linux:


Linux is a free and publicly available technology which can be used on digital devices. Linux-technology based on collaborative development model which guarantees that it is not dominated by any single author like commercially licensed technologies. The immaterial rights in the Linux technology allows to install, use, edit, copy and distribute the technology without any commercial terms or restrictions.

It would like to stress the words publicly available technology since it is more than just the kernel (kernel is the form of technology). — Preceding unsigned comment added by Ryytuo (talkcontribs) 22:09, 30 October 2011 (UTC)

  • Oppose. It's an operating system, not some kind of vague "technology". What's there represents a fought-over consensus. Msnicki (talk) 22:30, 30 October 2011 (UTC)
  • Oppose I agree with Msnicki, its an operating system and you dont mention that in the definition. — Preceding unsigned comment added by Beefcake6412 (talkcontribs) 22:36, 30 October 2011 (UTC)
  • Oppose technology it's way too vague (plumbing is technology too...) man with one red shoe 14:06, 9 November 2011 (UTC)

According to proposed definition, it is not intended to remove mention "kernel". Linux is not operating system, for instance Ubuntu, Debian or Solaris are operating systems. Operating system is larger entity than kernel, which is restricted comparing to operating system. For instance, Windows is product based proprietary technology, where the product is license and the role of technology is Windows and its variants. Talking about Linux, we can see it as free technology where Linux itself is not a product. Linux-products are products which uses technology called Linux. The current definition is not wrong, but it is very limited. Therefore I suggest to expand the current definition by using technological point of view. — Preceding unsigned comment added by Ryytuo (talkcontribs) 15:49, 1 November 2011‎

Linux is exactly a kernel that requires exactly a computer to run. Rursus dixit. (mbork3!) 14:49, 17 November 2011 (UTC)
If "Linux" referred to a kernel and only a kernel, this article would not exist, since we already have Linux kernel. That being said, I oppose the proposed "Linux is a technology" definition, which seems inexcusably vague to me. — Steven G. Johnson (talk) 18:57, 17 November 2011 (UTC)
  • Oppose The starting point for the opinion is that Ubuntu, Fedora and other distributions are the operating systems. That is wrong conclusion and is based to marketing, not computer science. The Linux is Linux kernel. Nothing else is Linux or belongs to Linux. When someone say "Linux" it should mean only Linux what is available from kernel.org site. Linux is not microkernel, it is a monolithic kernel. A monolithic kernel is original architecture for operating system and still used today, as even Linux existing proofs it. Linux has never meant a whole distribution. People who not know what Linux is, use it that way but they are victims of marketing and wrong conclusions. Linux kernel is the operating system. Nothing else belongs to that operating system, not from GNU or any other party. This article is about distributions. A distribution is software system what is distributed. The term "distribution" comes from history where first people to get a open source operating system (operating systems and software were at start open and free, until they were closed) and they were distributed on magnetic and paper tapes. A distribution was a single tape what had operating system and/or programs on it. And later a with Linux, like with BSD's, the user otherwise should have download all the source codes, download a compiler. Compile the software and then install it. Usually it meant, they needed a another computer with working software system and development platform to do so to get their new computer working. So idea to distributions came too that user should not need to hunt every software by themselfs, but they can just get a package what has all softwares in it. And those different software systems were distributed. Linux operating system is wanted software. But no one does anything just with the operating system as the programs are the ones what does the job for the user, but programs needs operating system to work. So they are distributed together. Different people need different programs. So there are different distributions from what to choose. Some distributions can be almost ready as that, some needs some tweaks and customisation but it is about the user, is she or he a user what distributors are targeting for. The Linux distribution article should only touch about distributions what use Linux as its operating system. If distributions would not exist, we would not need this article. And even that we have Linux operating system, it does not make distribution article obsolete because distribution and operating system are two different things. Operating system is technology and distribution is more like habit and idea how to collect other softwares what are ported to that operating system and how they have made work together and how that all has finished as final product under own brand. Golftheman (talk) 13:53, 24 November 2011 (UTC)
  • Comment I agree with something here, particularly Stevenj's comment, "If "Linux" referred to a kernel and only a kernel, this article would not exist, since we already have Linux kernel." So what is Linux? Ryytuo's proposed text might be good if we were writing the marketing brochure, but we are not. Most people, including many users of Linux, have no idea what it is technically, yet they use the word and it has meaning for them. As encyclopedia writers, I think we have two options: (a) we can write an article that explains what the word refers to purely technically (i.e a kernel) and so disabuse them of their 'confusion', or (b) we can be post modern about it and try to say that Linux is whatever most people think it is. In case (b), we might say it is a poster child for FLOSS, it is an alternative to Microsoft and Apple, it is about collaboration and freedom, it is owned by everybody... if we stick strictly to (a) we might as well delete this article as Stevenj points out. (Hint: Usually when there are two options, the answer is, do both) --Nigelj (talk) 16:50, 5 December 2011 (UTC)