Talk:GNU/Linux naming controversy/Archive 11

Archive 5

←

Archive 9

Archive 10

Archive 11

Archive 12

Archive 13

Lead section edits

Hello there, 63.153.253.49! As I've just commented, removed sentence still doesn't make sense to me, sorry. For example, what's the "discrete area", where is it visible how much the GNU project has contributed to it, how it relates to the Linux kernel code, and where is the GNU ld mentioned – speaking about the backing reference? I've read it three times, and either nothing like that is mentioned, or I'm unable to see it. Please advise. Thank you. — Dsimic (talk | contribs) 04:03, 29 January 2014 (UTC)

The previous paragraph lists a primary motivator for "GNU" inclusion in GNU/Linux naming schemes, GNU's contributions to the free (and in general) software community. The proposed addition clarifies that GNU's contributions to the operating system* lie outside the Linux kernel and with GNU libc and ld, through which most modern software make system calls are made instead of directly to the kernel.

*Operating system: software that manages hardware and access to it, synonymous with "kernel" until a few years after RMS started asking everyone to call Linux GNU/Linux. Even Wikipedia articles that mention operating systems make only a short note of the definition of "operating system" as a kernel and packages, but have at least a paragraph describing only a kernel. 63.153.253.49 (talk) 05:47, 29 January 2014 (UTC)

I do agree with everything you've just described, but the trouble is with the actual addition to this article (quoted is only the second sentence):

The contributions GNU has made to this discrete area only make up less than 0.5% of the kernel's code and GNU ld (the GNU linker.)

Maybe I'm stupid or something, :) but this sentence has no clear meaning to me; it's pretty much hedged and thus undechiperable, sorry. I know what you're trying to say, but unfortunately these words are not carrying that message. :) Also, used reference simply isn't backing the sentence, and—as we know—Wikipedia is all about summing up reliable sources. For example, where did "0.5%" come from?

Hope it makes sense. Of course, I'm more than happy to discuss it further, so we end up with a good addition to this article. — Dsimic (talk | contribs) 06:37, 29 January 2014 (UTC)

GNU Project's answer to the question. --AVRS (talk) 11:40, 29 January 2014 (UTC)

Well, an operating system kernel itself isn't making up a whole operating system for sure, but also that same kernel isn't just one of the programs – it's much more. — Dsimic (talk | contribs) 01:20, 30 January 2014 (UTC)

GNU has never been, is not, and will never be a reliable source of information regarding the terminology. They are the ones who pushed for the separation of the two definitions. The fact remains that the two terms meant the same thing for a long time before RMS started campaigning for "GNU/Linux," and I've never seen mention of other influences trying to change the terms' meanings. Personally, I will continue to assert that they are the same thing like most books and Wikipedia articles do. (In particular, I note they use "Unix operating system" as an example- no one ever called Unix's kernel + programs an operating system, they called it a tape.) 63.153.253.49 (talk) 02:09, 30 January 2014 (UTC)

Just to nip this in the bud: the Wikipedia article is not the place for editors to insert their own arguments about naming. The only way to adhere to policy (not to mention sanity) here is to restrict the article to arguments made by prominent commentators and/or notable publications.— Steven G. Johnson (talk) 15:54, 30 January 2014 (UTC)

Not sure how allowing lobbyist organizations to change definitions of English words and phrases nips anything in the bud. Publications that meet RS, dictionaries, and Wikipedia articles all refer to an operating system in the same manner that they refer to a kernel. Nothing written by someone other than a GNU member differentiates between the two to the degree GNU does. — Preceding unsigned comment added by 63.153.253.49 (talk) 18:34, 30 January 2014 (UTC)

Well, just as I wrote above: "... and—as we know—Wikipedia is all about summing up reliable sources." Also, a kernel and an operating system aren't the same thing; that's why we have "Linux" and "Linux kernel". — Dsimic (talk | contribs) 18:38, 30 January 2014 (UTC)

Yes, the distinction between "operating system" and the kernel in common usage is widespread and well-sourced in Wikipedia; see also Microsoft Windows versus Architecture of Windows NT#Kernel or OS X versus Mach (kernel). Or the Unix article, or NeXTSTEP (vs. Mach), etcetera. — Steven G. Johnson (talk) 19:49, 30 January 2014 (UTC)

A kernel and an operating system are the same thing. Just because nowadays everyone calls malware viruses, doesn't mean the definition has ever been correct. Operating system is a marketing term at best, a politically fueled quest for advertisement rights at worst. 75.170.245.101 (talk) 01:28, 31 January 2014 (UTC)

Ok, then take just the Windows NT Kernel and try playing Solitaire... Does it work without applications on top of the kernel? By the way, I'm never calling today's malware viruses; those DOS-era viruses were true masterpieces of programming, when compared to the majority of today's malware. — Dsimic (talk | contribs) 02:02, 31 January 2014 (UTC)

Your argument is completely irrelevant, and has been hashed out several times by several people. You can't arbitrarily label userspace: it's a bazaar. Your belief that a machine isn't useful without X, Y, or Z does not make X, Y, or Z an integral part of the hardware management and access abstraction the operating system does. 75.170.245.101 (talk) 18:39, 31 January 2014 (UTC)

Ok, let's keep it like that for now; could you, please, clearly describe your arguments for making no clear distinctions between kernels and operating systems? — Dsimic (talk | contribs) 22:55, 31 January 2014 (UTC)

Before 2002, there was no difference. Only when GNU started going around trying to get everyone to call Linux distributions "GNU/Linux distributions" did that start to change, because GNU started telling everyone GNU/Linux was the operating system's proper name. They tried to convince Linux kernel developers to change kernel behavior to report itself as GNU/Linux- they said no. They tried to convince distribution maintainers to change uname behavior to report the system (-s option) as GNU/Linux- they said no. Finally GNU just added an option to uname (-o, operating system) which always reports GNU/Linux: no matter what. WP:OR as heck, but easily verifiable in an unbiased journalistic/encyclopedic way. What GNU calls an operating system is a "distribution" and was, until slightly before GNU became the uname vendor in about 2004.

That should be reason enough to see no clear distinction: there is none. It's entirely a GNU fabrication designed for petty, political reasons, and has absolutely nothing with technology. If it did, why did kernel developers and distribution maintainers not see the light and graciously accept the GNU definition of "operating system?" 75.170.245.101 (talk) 04:52, 1 February 2014 (UTC)

Non-WP:OR, | IEEE uname specification which has never included -o and calls -s (you guessed it) operating system. Every Linux distribution reports Linux as the operating system, not the distribution's name, and not GNU/Linux. 75.170.245.101 (talk) 05:01, 1 February 2014 (UTC)

Hey, wait a second; are you talking about "kernel" as "Linux kernel", and "operating system" as "Linux" – or about "kernel" and "operating system" as general terms (outside the Linux domain)?

I do agree that "GNU/Linux" is pretty much a nonsense, but "kernel" isn't the same as "operating system" when speaking about general terms. As a (redundant) proof for the early distinction, there are numerous books on operating systems design from the 1980s. — Dsimic (talk | contribs) 05:08, 1 February 2014 (UTC)

Yes, an operating system and a kernel are the EXACT same thing, in general. Uname#Examples shows this- Linux is the exception- not the rule (coincidence that uname on Linux distros is written by GNU?) 75.170.245.101 (talk) 05:14, 1 February 2014 (UTC)

Whatever. Anyway, please keep in mind that the world extends beyond the boundaries of an uname's output. — Dsimic (talk | contribs) 05:21, 1 February 2014 (UTC)

The world, yes. Use of the phrase operating system to mean distribution, however, does not. At least not outside regurgitation of "Operating system is all the software on your computer." 75.170.245.101 (talk) 06:00, 1 February 2014 (UTC)

Why then do we need more stuff beside what kernel.org directly provides (short of various hosted mirrors), in order to do something with Linux – to compile a Linux kernel, for example? If the kernel equals the operating system, shouldn't then a Linux kernel be capable of compiling itself? — Dsimic (talk | contribs) 06:34, 1 February 2014 (UTC)

An operating system manages and abstracts hardware resources. Continuing to use this strawman argument of "Well, my computer can't do X with just a kernel!" will get you nowhere. I can't drive to the store with just a car, that doesn't make roads part of my car. Even the FSF has gone on record stating that simply using GCC to compile Linux doesn't make your system GNU/Linux, so why would it make GCC part of the operating system? 75.170.245.101 (talk) 18:35, 1 February 2014 (UTC)

Continuing on your car/road example, if you can't use Gimp with just a Linux kernel, that clearly doesn't make X part of your kernel; then, what is that X is part of? "Linux distribution" would be a good answer. Once again, I'm against using "GNU/Linux", but still can't accept that a kernel equals an operating system. — Dsimic (talk | contribs) 03:51, 2 February 2014 (UTC)

For what reason can you not accept it? I've given ample technical reasoning for the definitions I use for the terms. Give me a legitimate reason why "operating system" and "distribution" should be the same thing when "distribution" is already clearly defined. If you just think "important" software should be part of the operating system, give some kind of remotely sane way of clearly defining "important." You can't, and that's why people can't "not accept" the traditional definitions (and frankly GNU should accept them too.) 75.170.245.101 (talk) 06:18, 2 February 2014 (UTC)

Regarding what's important software, let's say it's the shell for example, which is mandatory for the System V–style init process. What can you do without a shell (short of running an embedded system, of course)? Another question is how would you relate "microkernels" and operating systems? I've been flipping through Tanenbaum's books, for example, and "microkernel" is mentioned there many times, while—putting my hand on my heart—there isn't such a clear line between kernels and operating systems.

To me, equalizing a kernel and an operating system made sense perfectly in the MINIX era, for the educational purposes, as there wasn't that much of the freely available software to make a difference; commercial Unix world didn't care about the naming, I'd guess. Later, free software movement gained traction, and much more software became available, making the gap between kernels and operating systems bigger; the real trouble is that GNU persistently tries to capitalize on that. — Dsimic (talk | contribs) 06:55, 2 February 2014 (UTC)

That's a smart question to ask. Operating system and MINIX and microkernel. A microkernel is darn close to a kernel and system daemons. Of note, GNU's Hurd is a microkernel. However, I still think that's the only place where kernel+userspace confusion legitimately crops up. You've again referred to "usefulness." An operating system is not what makes your system useful, it's what manages the hardware and gives you the ability to write programs that work on your x86 system and your PPC system without two codebases- because you use OS (kernel) system calls and let Linux (in this case) developers do the low level stuff for you. 75.170.245.101 (talk) 08:36, 2 February 2014 (UTC)

Still, my opinion is that equalizing a kernel and an operating system made sense only in the MINIX era. Many of the operating system design textbooks were created at that time; influence of microkernel-based designs and lack of freely available software really mangled the difference between kernels and operating systems back at the time. Many years later, things are quite different.

I'd really love if other editors would jump in (with valuable discussion, of course, not just by referring to the Wikipedia's rules), as me and 75.170.245.101 are pretty much like Statler and Waldorf in this thread. :) — Dsimic (talk | contribs) 21:18, 2 February 2014 (UTC)

How does the "era" matter one bit? An operating system manages and abstracts access to the hardware on your machine. That hasn't changed one iota as time has gone on. Again, you have not addressed this and given a definition of operating system other than that. 75.170.245.101 (talk) 02:02, 3 February 2014 (UTC)

It matters a lot; just remember CP/M, for example, or even DOS. They're both operating systems by definition (or kernels?), but they're nowhere close even to what MINIX provided. Having that in mind, if the definition of an operating system stayed unchanged since its invention, we'd have no concept of multitasking, for example. — Dsimic (talk | contribs) 02:58, 3 February 2014 (UTC)

Neutrality means accurately reporting the arguments even of people and groups you disagree with, if notability can be established. (Or are you suggesting that the FSF is not notable?) — Steven G. Johnson (talk) 19:43, 30 January 2014 (UTC)

Operating system

I've reverted the edits that try to remove "operating system" from the article for a few reasons. First of all, trying to reword the article to remove "operating system" in this way creates inaccuracies such as the lede suggesting that the controversy only concerns the distros and nothing more; this is obviously not the case. Going further down, GNU was planned as an operating system, reliable sources are quite clear in this. This article is not distro-focused, so trying to remove any instance of "operating system" just because an editor doesn't agree with common usage of what that word means doesn't benefit the article. - Aoidh (talk) 20:05, 1 February 2014 (UTC)

To your first concern, yes, the naming controversy is related to distributions. GNU has never been so bold as to attempt to get people to call the Linux operating system (kernel) GNU/Linux. This is not an inaccuracy. To the second, there are instances of "operating system" in this article that refer to entire distros. For example, the GNU project was an attempt to create an entire Unix-like distribution of free software- not just userspace tools. Other instances of "operating system" clearly refer to the Linux (or Hurd) kernel. No one is trying to "remove operating system" from the article. Someone is trying to use consistent terminology to refer to consistently defined abstractions. 75.170.245.101 (talk) 20:13, 1 February 2014 (UTC)

It is not related solely to distros, that is inaccurate as shown by the reliable sources, so that's the first inaccuracy introduced by your edits. Also inaccurate is your summary of the GNU operating system, every reliable source I could find (including the FSF itself) describes it as an operating system. Therefore, your changes to the article are inaccurate and will be reverted unless you somehow manage change the entire scope of the article so that it's only referring to distros, which is unlikely since reliable sources refute your assertion that the naming issue only applies to distros; it doesn't. - Aoidh (talk) 01:20, 2 February 2014 (UTC)

There is controversy surrounding the question of whether or not a distribution that includes both Linux and GNU software should be called a GNU/Linux distribution. That is the ENTIRE subject of this article, and the controversy itself is largely borne of people's lack of knowledge of what exactly an operating system is. Linux people know what an operating system is- the kernel. GNU people do not and choose not to- they think it's everything they deem "important" to the usefulness of the machine. They even go on record saying that the fairest name is a very long one that includes every major component, but that would be too silly so they assert that GNU is the only one that simply cannot be omitted.

You need to give specific examples of your problems and cut out the extremely specific, yet simultaneously vacuous statements about article scope and reference context. Name one section in which the issue is something other than "People can't agree whether bundles of software that include Linux and GNU software should be called Linux or GNU/Linux."

As for whether or not GNU is an operating system, that's already been hashed out above, and you're free to add to that discussion with something other than opinion. How GNU and the FSF have chosen to change the definition to meet their political agenda is pretty obvious and historically documented. Continued regurgitation of "The operating system is everything important on my computer" will continue to be nothing more than factually incorrect. I can write an article about how a Gnu is actually a floating ox and get it published by mainstream media all I want, that doesn't make it relevant to Wikipedia until it becomes an infotainment outlet. 75.170.245.101 (talk) 06:30, 2 February 2014 (UTC)

Unfortunately you're arguing against reliable sources, not me. So long as reliable sources call GNU an operating system, it's an operating system. But what would the FSF know about their own system, when you obviously know better, right? No. What the FSF (or any reliable source, for that matter) has called GNU hasn't changed. The scope of this article hasn't changed, and isn't limited to distros, despite your protestations to the contrary. Given that reliable sources contradicts your edits, they will be reverted. If you feel they shouldn't, you're more than welcome to try to get a consensus for your changes. Otherwise, they will not remain in the article per Wikipedia policy. - Aoidh (talk) 07:30, 2 February 2014 (UTC)

I'm not holding my breath. Arguments during these discussions have followed a pattern: Well, I don't think that's right; Well so-and-so says that's not right; Well it's Wikipedia policy to report on things the way they are commonly interpreted. I'm fairly confident the next step in that chain is "Maybe we should be encyclopedic and aim for correctness, instead of infotainers aiming for maximum bubbling." 75.170.245.101 (talk) 08:39, 2 February 2014 (UTC)

So what you're saying is that you think you know better than reliable sources. Unfortunately, reliable sources trump your personal opinion, and when you use words like "infotainers" to describe sources like the FSF back in the early 80's when referring to their own operating system, it only hurts your position and weakens any argument you might have. When it's your opinion against reliable sources, the reliable sources will trump you every time, I'm sorry but that's how Wikipedia works. - Aoidh (talk) 09:25, 2 February 2014 (UTC)

No, I'm saying what I'm saying. And considering GNU and the FSF (and infotainment outlets with no interest in the technology) are the only ones who believe an operating system is something other than what manages hardware and gives programs access to it, yes I'd say they aren't a reliable source for this topic. They're the ones who started the controversial usage of "operating system" to mean "What GNU deems important to the usefulness of your machine," and they're the ones who have benefited from it. How's that for neutral and factual information sourcing. Again, policy policy policy. I'm not worried about policy at this point. 75.170.245.101 (talk) 21:06, 2 February 2014 (UTC)

So the FSF, who made GNU, has no idea what they're talking about and isn't a reliable source for GNU. Instead we should use your personal opinion and ignore core Wikipedia policies such as WP:RS, which reflect consensus? Unless you have any reliable sources to back up what you're saying, we use reliable sources to dictate content, not your rants about "infotainment". Unless there's a consensus for the change, WP:STATUSQUO needs to be kept until a consensus is established supporting your changes, which seems unlikely. Given that reliable sources contradict your changes, I've reverted them. Until there's a consensus (or at least reliable sources) supporting your change, it has no place in the article. - Aoidh (talk) 21:48, 2 February 2014 (UTC)

I'm not a wikilawyer. I'd love to, as you temporarily seemed to want, see someone uninvolved make a determination on the validity of my changes. I don't think you care about policy at all, I think you simply disagree and are looking for a reason to. 75.170.245.101 (talk) 02:00, 3 February 2014 (UTC)

Until then, per WP:STATUSQUO we keep the previous version, the one that used reliable sources and not an IP's rants, to decide what was written. If that's truly what you're interested in, get a consensus and then make the changes, but if you just continue to push your preferred version then you'll do nothing more than make a liar out of yourself and the article will be protected to prevent you from edit-warring, either way it's going to take a consensus to make your changes. - Aoidh (talk) 13:56, 3 February 2014 (UTC)

Recent Edit History

I would like an uninvolved editor to arbitrate an edit dispute on the article at this time. I've come to the conclusion an involved editor has no basis for his opinion other than that I'm slighting published sources by referring to historically accurate terminology as well as the definitions for terms that current standards bodies use. I don't think policy is the issue, I think the fact that it suits someone's opinion is. 75.170.245.101 (talk) 01:58, 3 February 2014 (UTC)

The fact that reliable sources and policy both support keeping the previous version should tell you something...on Wikipedia that's how things are decided, trying to recharacterize it as a "clash of opinions" does nothing more than to highlight the fact that you have no basis for your changes and are grasping at straws to push your POV. However, I've asked for a third-opinion at WP:3O to hopefully settle this. - Aoidh (talk) 14:05, 3 February 2014 (UTC)

Third opinion

You made a request on Wikipedia:Third opinion for a fresh pair of eyes, so here I am. I'm not going to “make a decision” for you as 75.170.245.101 suggested; that's not the point of a WP:3O. All I can do is state my opinion; you will still have to decide yourselves.

Your controversy is kind of tricky because, in a way, you are trying to resolve the very issue the page is about. That said, the wording proposed by 75.170.245.101 (“software distribution”) seems to me to be much more geared towards one viewpoint than the wording previously used (“operating system”). It is true that there are conflicting definitions of the term “operating system”, and I have heard it used synonymously to “kernel”. From my subjective experience, however, it is generally considered to include the parts of the system which the kernel relies on to be able to run. For example, when referring to Windows as an “operating system”, this includes all the userspace components which are shipped together with the kernel (see Operating system—this is much more than the equivalent of GNU). Additionally, it is less accurate because the term “distribution” has a special meaning in the Free Software context, and it is not the distribution (as in, Fedora (operating system)) which is called “Linux”, but the software which is distributed.

Therefore, I would suggest that you keep the previous version of the page. The proper use of the term “operating system” may be controversial; however, as far as it is not immediately related to the GNU/Linux controversy, it is probably out of scope of this article.

Yokokokon (talk) 21:00, 3 February 2014 (UTC)

I am a frequent volunteer at Third Opinion. I have requested a sockpuppet investigation in regard to the 3O given above, for the reasons stated on the SPI page. I would suggest that the editors in this dispute not rely upon it until that SPI has been resolved. Regards, TransporterMan (TALK) 22:18, 3 February 2014 (UTC)

The investigation has cleared both Yokokokon and Aoidh, for which I am glad. My apologies for the interruption and to both of them for my suspicions. Regards, TransporterMan (TALK) 14:18, 5 February 2014 (UTC)

Failed verification

The addition of the failed verification tag was reverted with the edit summary "the plain English meaning of "created" is equivalent to "created from scratch" ... when the Bible says "God created the Heavens and the Earth" it does not imply that God was probably just adding the last few mountain ranges". That is, however, untrue. The plain English meaning of created is not equivalent to "created from scratch", that is self evident in the wording (hence the "from scratch" disambiguation of what kind of creating) and in the dictionaries I consulted; none of them support the idea that to create something, it must be done "from scratch". That a source verifies that Torvalds created something does not verify the claim that he created it from scratch. References to the Biblical creation are irrelevant; the source given does not verify the text in the article, and implying a conclusion the source does not state is WP:OR. I've tried to reword it to make it at least closer to what the source stated, but if there's a disagreement about it I'd rather discuss it on the talk page than via edit summaries. - Aoidh (talk) 04:41, 2 May 2014 (UTC)