Talk:Linux/Referring to this article

This page is discussing the nature of link text in other articles that are referring to and linking to this article. In other words, it discussing how the operating systems made out of the Linux kernel and various combinations of other software such as the GNU system and/or X11, should be referred in other articles in Wikipedia. It is not about the name of the Linux article. Those conversations should be had on Talk:Linux directly until a better place can be found.

Summary of Arguments

Common ground and things everyone agrees on:

1. "Linux" is the most commonly-used name for the operating system although many other terms are also used including "GNU/Linux". While the terms are often used interchangeably, many experts insist on a particular term.
2. The term "Linux" originally referred only to the Linux kernel but is now commonly used to also refer to operating systems that use the kernel in addition to other purposes.
3. The name of the operating system is a subject of active and protracted debate in the free and open source software community.
4. Using only one term for the name of the operating system annoys some contributors and dissuades them from editing.

Arguments in favor of using Linux:

1. It is the most common name.
2. From a historical perspective, the scope of the kernel and those layers above it which are Linux-specific and of non-GNU origin exactly matches that traditionally defined as being an "operating system". Other systems with a similar scope, such as MINIX, which may also be distributed with GNU utilities and libraries, are not subject to the demanded naming convention.

Arguments in favor of using GNU/Linux:

1. "Linux" is inaccurate because Linux more correctly refers to the Linux kernel. Wikipedia should use of the name "GNU/Linux" because it is more accurate.
2. "Linux" is misleading because the media regularly prints that Linux was written by Linus Torvalds in 1991. When Wikipedia then tells people that "Linux" is an operating system, the reader is mislead.
3. Using the term GNU better reflects the history and authorship of the operating system in question.
4. "GNU/Linux" is completely unambiguous. "Linux", on the other hand, is notoriously ambiguous. Using just "Linux" for software other than the Linux kernel just leads to confusion and chaos that is contrary to the mission of organizing knowledge.
5. Expressions such as "FOSS" accomodate the two main stances in the free software community ("free" and "open"); likewise, "GNU/Linux" accomodates the same two, respectively. A label referring to software, the core of which includes GNU and Linux, should fairly represent both the free and open stances.Freed42 (talk) 23:23, 25 July 2008 (UTC)

Arguments in favor of enforcing consistency in reference (i.e., in favor of either Linux or GNU/Linux but not both):

1. Wikipedia should be internally consistent.
2. Wikipedia's articles should not be written from the point of view of their subjects. Therefore, even where an article's subject uses either particular name to refer to the operating system, Wikipedia should not necessarily preserve the use of this name as the general name of the the family of operating systems in such articles, except where it forms part of a proper name (such as Debian GNU/Linux).

Arguments in favor of not enforcing consistency in reference:

1. This issue is akin to the differences between British and American English or the AD/BC or Gdansk/Danzig controversies and should be subject to similar conclusions that do not enforce full consistency on either term. Articles should be internally consistent and written by neutral sources but should be described using whatever term is most relevant or appropriate given the nature of the article subject and third-party references to it.
2. The cost of continued debate on this issue outweighs the benefits of completely consistency within Wikipedia.
3. Any confusion from use of either term could be offset with short explanatory text applied as a style guideline.
4. The terms are very finely nuanced, the article authors choose their terms carefully, and the article authors should be trusted to use the best term, case by case, and do not need to be compelled to comply to some officially sanctioned, politically correct set of terminology in writing on technically complex matters.

Responses

Each response below corresponds to an argument above. Please try to preserve this structure or add to or rearrange if the structure breaks.

Response to arguments in favor of Linux:

1. (response to common name argument)
2. The scope of this article extends beyond this historical definition to indisputably use a more high-level definition of "operating system".

Response to arguments to use GNU/Linux:

1. Wikipedia is not a defender of the truth. The accuracy of such claims is sufficiently disputed that we cannot and should not present them as factual.
2. Our policies allow for the reporting of such opinions, but they also demand that we refrain from advocating minority positions in the absence of an agreed body which would lend them authority.

Response to arguments to enforce consistency:

1. Consistency throughout Wikipedia is eschewed in areas where there are multiple correct answers including British and US English, AD/BC and CE/BCE, Gdansk and Danzig, and others.
2. Some distributions like gNewSense are referred to almost universally as "GNU/Linux." A neutral observer could base a decision to use the term "GNU/Linux" in the article on gNewSense on evidence other than a desire to write an article from the point of view of the article subject.

Response to arguments to not enforce consistency:

1. The dual use of UK and US English is primarily a compromise to allow two very similar languages to share one domain, and as such is really a technical hack. Unlike that situation, which is a matter of two undisputed positions which both have authoritative bodies behind them, there is an obvious component of prescription in the use of the terms here.

Proposed remedies

The follow remedies and paths forward have been proposed and are being considered.

Consistency on Linux

Wikipedia should not use the construction "GNU/Linux" in any cases except where it is included in the proper name of the distribution (e.g., Debian GNU/Linux). In these cases, the text of the article should describe Debian as "Linux."

Consistency only within articles

Articles should be consistent in their terminology and should not be written from the point of view of their subject but may employ either term when neutral editors feel that it is appropriate given the subject of the article and a close relationship to the GNU project or important community identification of the project as a GNU/Linux distribution.

Employ explanatory text

In articles referring to products described as either "Linux" or "GNU/Linux," editors should employ agreed upon text to accurately describe the relationship between GNU/Linux and where relevant to point out the naming controversy. This can be either in the text of articles linking or at the source of links on pages including Linux distribution and Linux. This might include raising the profile of such discussion which is already present in both articles and linking to the GNU/Linux naming controversy article.

Add clarification to the top of relevant articles

Add a note at the top of the article linked to by GNU/Linux and Linux saying that in Wikipedia, the two terms (unless the latter is clarified to mean the kernel) mean the same operating system.

Note: there is a note currently at the top of the article, but it contains no mention of "GNU".

Consistency on GNU/Linux

Describing what is not part of the kernel as "Linux" creates confusion and is highly misleading. The term "GNU/Linux" should be used to refer to amalgamations of the GNU operating system and the Linux kernel.

First step of a remedy: Linux disambiguation page

When a user searches for "Linux", present the disambiguation page. As Disambiguation explains, this could reduce confusion. Freed42 (talk) 20:59, 25 July 2008 (UTC)

Paul Beardsell has already argued extensively for your solution; search this page for the word "disambiguation". There is no confusion. A clear and obvious majority of people searching for and linking the term Linux are meaning to go to the operating system, so we link there (as it is the primary topic) and add a hatnote. See WP:HATNOTE. Chris Cunningham (not at work) - talk 22:06, 25 July 2008 (UTC)

No prescribed solution

Editors should feel free to make use of whichever term they choose to refer to the OS.

Note: this has been the norm in Wikipedia articles until now.

Discussion

I've just put up a first draft of everything above. It's a first stab at trying to summarize the arguments made, the responses, and the concrete proposals. Please feel free to edit this. We can discuss controversial edits. —mako๛ 16:48, 22 May 2008 (UTC)

Authorities at odds: so the market decides

With proprietary products then the manufacturer (lawyers) maintain consistent product naming. With GNU or Linux, there are few authorities who could help us arbitrate. The FSF offers "GNU/Linux" and the FSF only considers the "Linux" part to be the kernel (which we would expect). As a product that uses the license mark "Linux", I'm curious if the FSF have sublicensed that wording anyway [1] but that's an aside. With the word "Linux" then the license holders (The Linux Mark in the Linux Foundation) clearly consider the "Linux" to be bigger than just-a-kernel if you look at their website. These two authorities are at odds with each other and so can't be relied on for guidance. Authority falls back to what the majority in the marketplace uses. We know the answer as it is easier to just say Linux rather than GNU Linux in any language. Ttiotsw (talk) 21:29, 22 May 2008 (UTC)

So if authorities and the marketplace for gNewSense, Debian Ututo seemed to imply that GNU/Linux was more appropriate and more widely used, would you support that naming in those articles? —mako๛ 21:37, 22 May 2008 (UTC)

If only it was that clear but it isn't that way. Google still shows that gNewSense with Linux outnumbers gNewSense with GNU/Linux. UTUTO does a lot better though GNU/Linux still outnumbered. Inside the articles then the wording is whatever the consensus but given they are very much Free without non-free and the distro sites themselves use GNU/Linux then GNU/Linux is a logical description even if it goes against the market view. We can't do the same for "Linux" article though. In the end we're arguing for "Linux" the product as it is promoted by all and sundry verses "GNU/Linux" the ideology as argued by FSF. Whooosh - that goes over most people's heads along with Microsoft EULA, the GPL/LGPL and any other flavour click licences. Ttiotsw (talk) 01:25, 23 May 2008 (UTC)

I am not trying to argue that "GNU/Linux" is more popular than Linux; it's not. I'm not trying to talk about the Linux article name; after all, that's why I created this page in the first place. I'm trying to understand the arguments being made and positions being taken. If authorities and the marketplace seemed to imply than GNU/Linux was more appropriate (we can argue later about whether this is the case in particular situations), would you support that naming in those articles? —mako๛ 02:51, 23 May 2008 (UTC)

Yes, I would have to else I would just be pushing a minority POV. Ttiotsw (talk) 11:35, 23 May 2008 (UTC)

Good. Chris (and possibly Gronky) seems to have argued that consistency throughout Wikipedia on this issue is important enough to trump this. I'm just trying to find out where other people sit. —mako๛ 18:56, 23 May 2008 (UTC)

I don't think consistency trumps allowing "GNU/". I think consistently using "GNU/Linux" is the optimal outcome, but unless there is clear consensus for consistently using either term (which I think is very unlikely, this year), then a "consistency" policy, no matter which term is chosen, would not be appropriate. My reason for this is that even if there emerges a slight preference for the use of "GNU/Linux", I don't think it would be good for Wikipedia to make a hostile policy towards the large "Linux" minority. I disagree with the "Linux" people, but I wouldn't want them to feel less welcome. Wikipedia thrives on having a massive community of editors, and that's more important than consistency (even if it was consistency for "GNU/Linux"). --Gronky (talk) 08:06, 27 May 2008 (UTC)

I'm touched by your hypothetical magnanimity. Hypothetically, I would be more than happy to accept "GNU/Linux everywhere" as the conclusion, were it not for all the reasons why this is the wrong call. Chris Cunningham (not at work) - talk 09:45, 27 May 2008 (UTC)

What I said was that even if the side I support gets the most support, if the margin is not big enough, I still wouldn't support implementing a always-GNU/Linux-everywhere policy. If you think this concession is empty, feel free to make the same concession. --Gronky (talk) 11:26, 27 May 2008 (UTC)

The market decides?? Then we would name things according to the wishes of the company that puts the most money to advertise its "trademark". Linux is only the kernel. Period. That some companies choose to use the name linux for the entire OS, does not change the fact that linux is only the kernel. Because some companies say so wikipedia has to bend to their marketing wishes and follow their naming rules? The use of the name GNU/Linux to describe the OS that uses the GNU system the linux kernel cannot be forbidden or removed from articles as has been done and it is being attempted by some users to push their own point of view which is in fact, as clearly said above, the point of view of companies that sell the OS as "linux".--Grandscribe (talk) 22:05, 25 May 2008 (UTC)

If some are so keen on letting the market decide why are some so keen on *enforcing* consistency here at WP? Surely, then, the "market" of editors here will arrive at some position with which all will be happy? Perhaps, but then we must guard against the brute force mass "market" manipulation of WP (which in the financial and commodity free markets around the the world is illegal) where someone just decides WTF let's change all occurrences of X to Y. Paul Beardsell (talk) 00:10, 26 May 2008 (UTC)

And! The authorities are not really quite so much at odds, day to day, as we portray. The various terms in use (Linux distribution, GNU/Linux, Linux kernel, and even the bareword "Linux") are all understood perfectly well by all the authorities and any disambiguation necessary is done naturally (either explicitly or by context) by the authorities when they communicate amongst themselves or when they deign to communicate with us. Where the problem lies is in the readership of this WP. Our purpose here at WP it is *not* to promote some FSF or anti-FSF agenda: It is to represent the state of knowledge. Not of the man in the street (because their state of knowledge on these issues is poor) but of the authorities. Anyone who *insists* on consistency where none is possible because these terms are all nuanced is really dumbing down WP. The woman in the street (or at home, if those plebs have houses) needs to be able to find out the knowledge we have set ourselves up as guardians of. Just how arrogant is it for any *one* of us here to decide on imposing consistency accross the whole of WP in the midst of controversy? Paul Beardsell (talk) 00:10, 26 May 2008 (UTC)

So: Let the article *authors* decide whether they want to write "Linux" or "GNU/Linux" or "Linux kernel". They will automatically be disambiguating explicitly or by context. Should they fail to do so then edit the article and provide the disambiguation. It is not useful to mass-edit "GNU/Linux" to "Linux" because without carefully reading the context you cannot be sure of providing disambiguation which may now be required. To change "GNU/Linux" to "Linux o/s" or "Linux distribution" without careful thought is also not appropriate - the author may be referring specifically to a GNU aspect. Also the reverse is not correct - as not all Linux operating systems (i.e. those that contain the kernel called Linux) have much GNU in them. Paul Beardsell (talk) 00:10, 26 May 2008 (UTC)

Those insisting on imposing consistency are reducing the quality of WP. Paul Beardsell (talk) 00:10, 26 May 2008 (UTC)

Having decided that, the next step is to make the article Linux a disambiguation page. Paul Beardsell (talk) 00:10, 26 May 2008 (UTC)

I agree with Paul Beardsell. Linux should be a disambiguation page. I also think we should stop immediately to redirect GNU/linux to "linux". GNU/linux should direct to an article about the GNU system when it uses the linux kernel. Or it should direct just to GNU because an article with such a title would mainly describe the GNU system. --Grandscribe (talk) 05:20, 26 May 2008 (UTC)

I added some more suggestions to the proposed remedies--Grandscribe (talk) 05:53, 26 May 2008 (UTC)

It's political not ontological

The name of this subpage and the opening paragraph I do not find helpful. Worse, they are at odds! They serve merely to continue to muddy the water.

How does one refer to anything? By using its name. When we "refer to this article" we use its name! You've got to decide: Are you referring to this article or to the various things which have the word Linux in their name?

The question as to how we refer to the things with Linux in their name or which contain Linux (the operating system kernel) is political not ontological. We need to reflect the controversy here at WP, it is not for us to gloss over it (for someone's misplaced sense of neatness) or to resolve it (to satisfy someone's RMS-bias). That is why we need to leave the article authors alone and also to make the article Linux a disambiguation page.

Paul Beardsell (talk) 00:24, 26 May 2008 (UTC)

The waters are very muddy already. This page is trying to to help move toward consensus on one part of the complicated issue: What term or terms should Wikipedia articles other than Linux use when referring to the operating system composed of GNU, a Linux Kernel, and other necessary pieces? That's what I tried to convey in the title and first paragraph. If that's unclear, please edit it or rename it for clarity. For what's it is worth, I think I agree with both of your suggestions. However, I think the former is more likely than the latter and I would like to handle each issue separately. —mako๛ 01:23, 26 May 2008 (UTC)

There is one thing we should do soon. We should stop the redirection of GNU/linux to Linux. GNU/linux should direct to a page about the GNU system that uses the kernel linux. Linux should be a page only about the kernel because that is what it is: a kernel. To do otherwise is to create and maintain confusion.--Grandscribe (talk) 05:10, 26 May 2008 (UTC)

I too would like to have Linux being the article about the kernel *but* the WP guidelines are plain: That slot in the namespace needs to be a disambiguation page. Paul Beardsell (talk) 12:25, 26 May 2008 (UTC)

Even Gronky doesn't agree that GNU/Linux should point to a different article from Linux, which makes this a very unpopular proposal indeed. Chris Cunningham (not at work) - talk 17:14, 26 May 2008 (UTC)

Popularity isn't everything. With three votes you've considered yourself entitled to brute force edit the whole encyclopedia, destroying meaning in the process. Paul Beardsell (talk) 07:31, 27 May 2008 (UTC)

I didn't go by "votes", I went by the consensus generated by discussion. It's also plainly contradictory to argue that I can't edit things because they only got a few "votes" despite a firm rhetorical backing, yet your position's arguments are strong enough that it should be implemented despite its undisputed minority status. Chris Cunningham (not at work) - talk 08:16, 27 May 2008 (UTC)

And who judges that this position is clearly rhetorically firm? When I look at the many wikipedians who disagree with you, I don't see worthless contributors to be ignored. This isn't a vote, but one does have to question the strength of your arguments that have convinced so few wikipedians. --Gronky (talk) 08:30, 27 May 2008 (UTC)

Chris, this is the same mistake, made again. The arguments I am forwarding do not have minority status - they are plain WP policy. That you say this would afford a "minority" term with WP:UNDUE weight is your argument, that I say this consideration is not a trump card and that there are plenty(!) of strongly worded caveats and riders quoted throughout other guidelines (and that same guideline) against what you say, is my argument. What we should both want is a better encyclopedia. What my view on the "correct" use of the term is, and what your view is, should be irrelevant. Paul Beardsell (talk) 08:36, 27 May 2008 (UTC)

Actually, I'd support making "Linux" point to the kernel and making "GNU/Linux" the name of the operating system article. To Paul's suggestion that "Linux" has to be a disambig: couldn't it be the kernel article and just put a disambig sentence at the top? --Gronky (talk) 08:18, 27 May 2008 (UTC)

To explicitly make your point: Chris wouldn't like that as it imposes a POV in *exactly* the same way he has come close to succeeding to impose his POV. But that would make Chris as unhappy as you are now. For some strange reason (I seem to be wavering in my resolve on this point) I am keen Chris is able to live with what we do here. Paul Beardsell (talk) 08:41, 27 May 2008 (UTC)

Keeping the "Linux" people involved in Wikipedia is a worthy goal - I defend this too in my comment above (which starts with "I don't think consistency trumps"). Building a massive (diverse) contributor base is very important for Wikipedia's quality in the long term. --Gronky (talk) 09:30, 27 May 2008 (UTC)

Actually, no, that's not what I would object to: Putting the kernel at Linux is not imposing a point of view. What it is doing is inconveniencing a majority of readers for the sake of a rather questionable purity. Having Linux be a disambig page is also, in itself, not imposition of POV - but again, it's disruptive. What would be POV would be putting the OS article at GNU/Linux, which an explicit attempt to have WP endorse the FSF's terminology. Chris Cunningham (not at work) - talk 09:58, 27 May 2008 (UTC)

So, you're saying that Wikipedia should adopt your policy because it's already been implemented (pre-emptively by you), and because fixing masses of links is too much work (despite you proving this wrong last Summer). --Gronky (talk) 10:21, 27 May 2008 (UTC)

Please be honest here. At the time that I started changing links, I believed (based on previous discussion) that the issue was settled. I did not pre-empt discussion by changing the links. Furthermore, when challenged on this change it was me who initiated formal discussion on it. This characterisation of my actions as underhanded is pretty ridiculous. As for the quantity of changes, as I've pointed out there was at least an order of magnitude's difference in the number of links to each article at the time; making those changes was definitely the path of least resistance, whereas doing it the other way around would have been a great deal more work. Chris Cunningham (not at work) - talk 10:34, 27 May 2008 (UTC)

Chris, I don't see how you get all that out of what Gronky wrote! But most of the Linux links to which you refer would have been disambiguated by context. At least some of the GNU/Linux links you changed to Linux would have relied on the GNU to provide the context. That is why some of your edits would have been destructive. The article authors need to be trusted. We should neither gloss over the controversy for political reasons, nor should we do so for reasons of specious neatness. However you sit, politically or consistently, I don't like it. Paul Beardsell (talk) 10:54, 27 May 2008 (UTC)

But they really don't. I really think you should examine some of the situations where this was changed. In any case where there was ambiguity (for instance, basically any trivial application which can be compiled with GCC will work across multiple platforms), I changed the wording of any inclusion to say "cross-platform" or "on a variety of free software platforms" or (in the case of one user who insisted on the GNU component being noted "on any glibc-compatible platform". I didn't just robotically change "GNU/Linux" to "Linux" across every article without looking. There is an argument that platforms such as gNewSense are "intrinsically GNU" and so need the mention, but this is (as far as WP is concerned) merely the opinion of the subject. Take a look at the wording at the start of Debian. It makes clear that the project has a wider scope than Linux, but still definitively states that "Debian GNU/Linux is a Linux distribution" because "Linux distribution" is the idiom used to describe that thing that Debian GNU/Linux is. I don't have a problem with "trusting" article authors where they aren't editing for the sake of personal opinion. The original rewording of the gNewSense into to reflect the more pro-FSF position ("gNewSense is a distribution of the GNU operating system on the Linux kernel" or such) was done by the FSF's webmaster. Chris Cunningham (not at work) - talk 11:24, 27 May 2008 (UTC)

You campaign looks political, not ontological. All you need to do to counter this is to identify the edits where you changed "Linux" to "GNU/Linux" (or you changed "Linux" to "Linux operating system") to aid understanding and to improve the encyclopedia. Paul Beardsell (talk) 01:34, 28 May 2008 (UTC)

Examining your "GNU/" purge is tedious because your edit summaries didn't mention what you were doing, but I've just gone and picked a sample. It happens to be 14 edits in a 2 hour period from September 7th 2008: [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15].

In each of those edits, you replaced a term for the OS with a term which sometimes means the OS, sometimes means the kernel, and I don't see where care was taken to lessen this introduced ambiguity. --Gronky (talk) 17:34, 27 May 2008 (UTC)

Let's see. In [2], the examples chosen are arbitrary. In [3] I replace "the GNU operating system and its GNU/Linux variants" with "GNU and Linux", which is uncontroversial. Your point is valid with [4] and [5] (and I'd be happy to reword them). With [6], the term refers to a specific Linux distro and isn't ambiguous, as with [7], where the only supported free software platform is Fedora (this is a common reason for simplifying to "Linux"; if the only free platform an app is supported or available for is a Linux distro, mentioning GNU confuses things because it implies portability which may not be present). [8] also refers to Linux distributions as an example. In [9] I removed platform information completely. [10] again offers a list of example ports; for a complex app like OpenOffice, "GNU" is most certainly not one port and thus it doesn't make sense to mention it. In [11] I changed the family Edubuntu is in to Ubuntu, which is the correct call. [12] already used "Linux" in the infobox. [13], [14] and [15] were all over the place, and weren't even internally consistent. By my count that's two cases out of fifteen where there's a reasonable argument that GNU should be reintroduced for accuracy and disambiguity purposes, and in neither do I believe it's necessary to do so using the loaded term "GNU/Linux". Chris Cunningham (not at work) - talk 18:21, 27 May 2008 (UTC)

Chris has been participating actively and in good faith in this conversation. What's done is done, let's focus on how we are going to solve this one. —mako๛ 23:41, 29 May 2008 (UTC)

(Outdent, sorry.) 2 out of 15 of the edits were bad ones! That is a much, much worse result than I expected. I am very upset. It is so very difficult to fix bad edits when time has elapsed and other edits have been done since. And all because you think "GNU/Linux" is the computer science equivalent of a 4-letter word. For goodness sake! Let reason prevail. Paul Beardsell (talk) 01:59, 28 May 2008 (UTC)

Two out of fifteen, extrapolated to the 200-250 edits made, results in roughly 30 articles where the change made the situation potentially more confusing to the reader. This is not an unacceptable level of fallout for a large project. Making high drama out of it seems a little silly in that case, doesn't it? Chris Cunningham (not at work) - talk 07:06, 28 May 2008 (UTC)

Primum non nocere. Paul Beardsell (talk) 09:42, 28 May 2008 (UTC)

Please. I won't believe that the issue here is the n edits that Chris made that might have made things more confusing. In any case, it's over and done with. Let's focus on what we're going to do in the future and stop pointing fingers. —mako๛ 23:37, 29 May 2008 (UTC)

OK. Paul Beardsell (talk) 00:38, 30 May 2008 (UTC)

---

The question as to how we refer to the things with Linux in their name or which contain Linux (the operating system kernel) is political not ontological. We need to reflect the controversy here at WP, it is not for us to gloss over it (for someone's misplaced sense of neatness) or to resolve it (to satisfy someone's RMS-bias). That is why we need to leave the article authors alone and also to make the article Linux a disambiguation page. Paul Beardsell (talk) 08:42, 27 May 2008 (UTC)

Paul you're doing good efforts to try to make this move on. Just a couple of questions. Shouldn't there be an article titled the "GNU/linux" operating system not just a redirection to this page? As Gronky said indeed this article is about the GNU/linux OS. What he says makes sense, to rename this to GNU/linux. Doesn't the proposed disambiguation point back just to linux biased pages?--Grandscribe (talk) 09:51, 27 May 2008 (UTC)

Yes, why not? But not to the exclusion of an article named "Linux operating system", or similar. Paul Beardsell (talk) 10:10, 27 May 2008 (UTC)

The best way to reflect the controversy is to make not of it. It is not to disrupt the project by forcing primary topics to live in disambiguated pages. Chris Cunningham (not at work) - talk 09:58, 27 May 2008 (UTC)

No, but the approach you propose is not taken universally at WP and there are guidelines as to when to choose one over the other. WP:UNDUE has been quoted on "your" side, and a number of caveats and riders, some strongly worded, are quoted by "my" side. Paul Beardsell (talk) 10:10, 27 May 2008 (UTC)

Thanks for the answer Paul. As you said a solution that does not exclude linux but neither GNU has to be found. The apparent "solution" that has been given by some to just add a few lines of text about GNU as a naming controversy(with dismissive remarks) at the beginning or end of an article using the name linux for GNU software (which is not part of the kernel) would not be acceptable. Wikipedia originally had an article titled GNU/Linux. Then someone redirected it to this page. The GNU/Linux page would have to be restored and one part of the solution would be not to force editors to link to the linux page when he/she wants to link to the GNU/Linux page where the focus would be description of the GNU system. The moment he/she would like to give detailed information about the kernel then he/she could point to the appropriate linux kernel page.--Grandscribe (talk) 11:02, 27 May 2008 (UTC)

But there's an entire GNU article, why duplicate info? If you want page with focus on GNU that's the right one. -- man with one red shoe (talk) 02:40, 28 May 2008 (UTC)

There's more than one thing called "GNU". But that's a separate if related issue and a discussion for another time. Or, at least, another section. Paul Beardsell (talk) 03:10, 28 May 2008 (UTC)

Then one idea could be this: when someone searches for GNU/linux then he/she could be directed to the GNU article. If a user simply types linux then he/shes could be led to the linux article. I understand Paul that is a separate but very related issue for later discussion. The redirection of GNU/linux to linux has to be stopped and changed to direct to GNU or to a page titled GNU/linux. This can help find a solution.--Grandscribe (talk) 08:39, 28 May 2008 (UTC)

I would bet that virtually nobody searches for "GNU/Linux" so for what I care that can be redirected to anything, including GNU, the problem is that zealots will make an effort to have "GNU/Linux" in all the texts that talk about the unnamed operating system and that's obviously not right to be redirected to GNU. Not even FSF clam that what they call GNU/Linux is GNU, that would be factually incorect. My guess if that you make here outlandish proposals just to make a WP:POINT, from now on I will just ignore the hot air... -- man with one red shoe (talk) 01:06, 30 May 2008 (UTC)

You have this the wrong way around. Doing as you describe would make one a zealot. Arguing in favour of the unthinking removal of the term would also make one that. Let's judge people by their actions, not by a label we've assigned them. I assure you, if you find anyone acting as you fear then tell me and I'll be in there arguing against that too, and hitting the undo button. I'm not aware of behaviour such as you describe, however. Paul Beardsell (talk) 01:45, 30 May 2008 (UTC)

Let me get this straight, do you consider that this proposal "when someone searches for GNU/linux then he/she could be directed to the GNU article" is serious, and even more it deserves our attention? To me anyone who makes such a proposal loses the latest shread of credibility. Now, if you consider that position correct and mine incorrect I will stop arguing here because it's clear we have different ways to process information. -- man with one red shoe (talk) 01:51, 30 May 2008 (UTC)

I took the time to check through many articles and their respective history and talk pages going as far as a few years. It is recorded that many wikipedia users were indeed searching for GNU. They even tried to contribute by editing and adding references or links to GNU. But their edits were quickly reverted by some editors such who claim to have the right to decide for them that is "linux" what users are looking for. That is not acceptable.
As you yourself have said users should be pointed to a disambigution page. They should have the option to decide for themselves what they want to read about. If nobody was searching for GNU/linux then WHY bother to redirect that page to linux??? If you do not like a proposition to point a search with the word linux in it to a GNU page so you should not agree to force users who make searches with the word GNU in it to be pointed only to the linux page. If some enters GNU/linux then he should be directed to a disambiguation page and NOT to "linux".--Grandscribe (talk) 10:09, 30 May 2008 (UTC)

Many is not equal to Majority or being right. Wikipedia's policy of NPOV does not mean equal time, and your comment that being reverted is unacceptable, shows that you have failed to understand how wikipedia works. (ie. by consensus). You can find many people who insert religious comments into Big Bang, or many people who try to insert strange political comments into Global warming (which is about the science) etc. etc. They all get reverted, because of WP:NPOV. --Kim D. Petersen (talk) 12:12, 30 May 2008 (UTC)

(outdent)It is unfair of you to characterise your opponents together with those who prefer Creation to Big Bang, or as Global Warming dissenters. We all agree (say if you don't) that properly and formally Linux refers to the kernel. We all agree (say if you don't) that popularly and informally Linux refers to a family of operating systems, sometimes all or a particular stream of FOSS, a community and even a philosophy. All we are disagreeing about is how the name "Linux" should be used at WP and about the near banning of the term GNU/Linux. It isn't Big Bang vs Genesis, it isn't Global Warming vs its deniers, we are *not* having a fundamental disagreement about fact. We are having a argument about the meaning and use of a word. Paul Beardsell (talk) 00:35, 3 June 2008 (UTC)

I did not characterize my "opponents" as anything. I gave some examples that "many" did not correspond to correct, consensus or being right, and i gave those from articles that i've edited, and examples that i hope everyone could see where wrong. I'm deeply sorry if it could be misunderstood in the way you apparently see it - but as said, that was nowhere the intention. But please take the gist of it to heart - an argumentation of "many" is a fallacy, unless you can demonstrate that its actually a consensus opinion. --Kim D. Petersen (talk) 02:20, 3 June 2008 (UTC)

I'm getting lost - shall we work on the summary?

I'm getting lost in these threads. Like most people, I can't work on Wikipedia every day, and a lot can be written and asked during a 48hr gap.

I've made some small changes to the Talk:Linux/Referring_to_this_article#Summary_of_Arguments section. Maybe we should all divert some of our editing time for this issue to improving that shared section.

Meanwhile, if anyone's waiting on a response to a question, please let me know. --Gronky (talk) 12:21, 30 May 2008 (UTC)

KDP revert. Discussion to avoid edit war.

Gronky wrote "GNU/Linux" is the name used by the primary authors of the operating system, the GNU project. and this was removed by KDP on the grounds that [he says] it is untrue. Why is that untrue, KDP? Paul Beardsell (talk) 15:03, 2 June 2008 (UTC)

Because GNU never had anything to do with the Linux project. The Linux projects used GNU utilities, since they were free, and could be used without getting into copyright problems. The GPL was the protection from this. If you take the argument that because GNU tools where used, then it must be a GNU project, then you are arguing against the GPL. Which specifically states that you can use it "as you please" as long as you retain the GPL and make the source available.

To state that GNU was the primary authors of the operating system, is a counter-factual representation of history. The real (and valid argumentation) albeit (imho) weak, is provided just above:

Using the term GNU better reflects the history and authorship of the operating system in question.

That is why i'm removing it. --Kim D. Petersen (talk) 15:15, 2 June 2008 (UTC)

I have never seen anyone claim or imply that GNU were ever involved in the Linux kernel--seems nothing more than a tired, old straw man. Suppose we could show that the LOC for GNU software is roughly an order of magnitude as many as the LOC for the kernel Linux; software that includes core components such as the C library. What else would you want to support "primary"? Freed42 (talk) 16:00, 25 July 2008 (UTC)

This has been discussed several times. There are several arguments against the (imho) rather silly LOC count.

1. In text-books about operating system theory the operating system is defined as the kernel. See for instance Tannenbaum.
2. The GPL specifically disallows this usage of required naming. You can use GPL'd software as you please - with no other conditions attached than that it stays open source and has a GPL licence attached. You may argue that its "fair" - but thats quite another argument.
3. If #2 had been correct - would Linux have based so much on GNU? (i don't think so - then we'd have a Linux with BSD tools instead)
4. There is no requirements for Linux to run with GNU software. Linux functions quite well without (and does in many appliances and dedicated purposes).
5. The GNU C library is not a requirement, its a convenience. Linux has been distributed with its own C library before.

....etc. So it really does not describe the "primary authors". --Kim D. Petersen (talk) 23:41, 25 July 2008 (UTC)

Rebuttals

Note: the software being discussed is any distro based on the Linux kernel and GNU utilities.

1. You cannot have it both ways. If you insist on this, for starters, Wikipedia's definition of OS---and whatever else in Wikipedia that depends on it--must be corrected.

   Why would we change the Wikipedia page on Operating system? It very much is in correspondence. --Kim D. Petersen (talk) 01:04, 26 July 2008 (UTC)

1. This is another tired straw man. Who has ever claimed that the GPL requires some kind of name? Where's the proof? That's about as silly as the GNU-worked-on-Linux-kernel meme.

   You just did - you claimed that after someone has built an OS - it should be named after the components - despite that they had no activity in creating the OS. Either GNU has no requirements for naming here (per GPL) or you are contradicting yourself. --Kim D. Petersen (talk) 01:04, 26 July 2008 (UTC)

2. Moot and useless.

   Except of course that it goes right to the heart of both the open-source movement and RMS's philosophy. --Kim D. Petersen (talk) 01:04, 26 July 2008 (UTC)

3. Irrelevent since what you describe--the Linux kernel in some system without GNU--is not even the object being discussed.

   This page is about Linux (and references to it) in all its shades. Embedded, Servers, Desktops etc. And quite a few servers and almost all embedded systems run without GNU. Or are you saying that my WRT router or my Synology disk isn't a Linux system? --Kim D. Petersen (talk) 01:04, 26 July 2008 (UTC)

4. Irrelevent since, again, what you describe is not even what is being discussed.

   Why did you bring it up then? Read your own text. (LOC argument) --Kim D. Petersen (talk) 01:05, 26 July 2008 (UTC)

In case it is not clear, non-GNU systems based on the Linux kernel may be the cat's meow, but it's also utterly not the kind of system whose label is being discussed here. Now, you mentioned "etc". Can you introduce a viable argument against LOC this time? Freed42 (talk) 00:27, 26 July 2008 (UTC)

Of course non-GNU systems are also being discussed, as said above this is about Linux in all its shades. Embedded systems are one of the largest deployment areas of Linux - and the one place where Linux is doing better than every other OS at the moment.

What systems do you want to limit the discussions to then? Only desktops? Only servers? Only .... - of course it would be convenient if such systems (non-GNU) didn't exist - or weren't in so damn widespread distribution. But the inconvenient fact is that they are. --Kim D. Petersen (talk) 01:02, 26 July 2008 (UTC)

Let's get more precise, shall we? You can consider whatever systems you want. Anything that I say, however, involves only a certain kind of software, the "Subset" of software that is based on GNU software and the Linux kernel. The primary authorship that I dispute with you is of the Subset. The LOC argument that I advanced is of the Subset. Is that clear enough?

Then you are arguing another case than what this discussion is about. This discussion is about Linux, and the article Linux. Not some subset of Linux that you personally define. --Kim D. Petersen (talk) 13:46, 26 July 2008 (UTC)

You agree that primary authorship is the particular matter here, introduced by Gronky, right? Primary authorship of what? Of systems based on GNU software and the Linux kernel. Did Gronky or I suggest any other kind of system? Freed42 (talk) 17:48, 26 July 2008 (UTC)

We are talking about Linux systems here - not whatever subset of these that you seem to prefer. --Kim D. Petersen (talk) 17:55, 26 July 2008 (UTC)

I am addressing the class of software that Gronky brought up. Why don't you state precisely this class that Gronky is referring to? If this class includes some system other than that based on GNU software and the Linux kernel, then I bet Gronky would differ with you. I.e., his claim does not refer to your class. Freed42 (talk) 19:09, 26 July 2008 (UTC)

How do you know what class of software Gronky referred to? The reasonable assumption to draw is that Gronky was referring to what is being discussed here, not some abstract subset of Linux but Linux in general. Especially since Gronky had been involved in the discussions that preceeded this page, where the OS part had already been discussed. --Kim D. Petersen (talk) 20:13, 26 July 2008 (UTC)

I will give you direct quotes. First, the very top of this whole article, defines "what is being discussed here": "This page is discussing the nature of link text in other articles that are referring to and linking to this article. In other words, it discussing how the operating systems made out of the combined GNU system and Linux kernel should be referred in other articles in Wikipedia."

That text should be changed. Since it now is apparent that it confuses you. This page is about what we should call links to the Linux article - which most definitively doesn't only talk about "operating systems made out of the combined GNU system and Linux kernel" taken in a very strict pedantic sense. Now you could of course argue on the Linux talk page, that the article should only contain information about Linux with GNU inside - but frankly i think its a lost cause. --Kim D. Petersen (talk) 23:58, 26 July 2008 (UTC)

Strictly for purposes of discussion on this talk page, I am not sure but it seems that you possibly believe that non-GNU systems based on the Linux kernel have mattered a great deal in this discussion. I also see the label "GNU/Linux" in many places on this page. Therefore, it should be trivial for you to indicate one statement by a participant in this discussion which implies that "GNU/Linux" should be applied to non-GNU systems based on the Linux kernel. It seems unlikely that an informed person would make such a fundamental mistake. Freed42 (talk) 02:39, 27 July 2008 (UTC)

This is a discussion on what we should name links that point to the Linux page. Since the Linux page isn't limited to systems that contain GNU its a very important distinction. And you are going circular now. Limit the discussion please to the subject that the Linux article is about. --Kim D. Petersen (talk) 03:58, 27 July 2008 (UTC)

I agree with your first two statements. It seems that no one has implied that "GNU/Linux" should be applied to non-GNU systems based on the Linux kernel, so that should simplify the discussion. Freed42 (talk) 04:37, 27 July 2008 (UTC)

You can see for yourself, that "what is being discussed here" is not Linux in general, since that quoted text limits it to only GNU Linux kernel. Now, this whole thread began when you reverted Gronky's text, which states, "'GNU/Linux' is the name used by the primary authors of the operating system, the GNU project." Gronky's text, in particular, the "operating system" could apply to many things. Knowing that this text belongs to a summary of "what is being discussed here" and assuming good faith on the part of Gronky as we are supposed to do, we are forced to assume that the class of software to which Gronky refers is the same class as defined in "what is being discussed here". I.e., "operating systems made out of the combined GNU system and Linux kernel". There is a chance that Gronky meant something else--we cannot read his mind, after all. However, I find it very unlikely that he meant anything else, and, in any case, I am content to discuss here only those systems mentioned in "what is being discussed here". Is that clear? Freed42 (talk) 22:24, 26 July 2008 (UTC)

Gronky has been around, and had been in the discussion for long enough - not to make that silly mistake. The Linux article does not limit itself in such a way. --Kim D. Petersen (talk) 23:58, 26 July 2008 (UTC)

It seems that you imply that Gronky refers to software systems, including non-GNU ones based on the Linux kernel. We also see that his reverted comment applied the label "GNU/Linux" to the software he refered to. Therefore, although you say he did not make one kind of silly mistake, he surely made a different silly mistake by labelling as "GNU/Linux" non-GNU systems based on the Linux kernel. Agree? Freed42 (talk) 03:26, 27 July 2008 (UTC)

You assert that OS = kernel, but if this were true, WP will need overhauling. E.g., from Windows XP, "Windows XP is the successor to both Windows 2000 Professional and Windows Me, and is the first consumer-oriented operating system produced by Microsoft to be built on the Windows NT kernel." The text states that Windows XP is built on the Windows NT kernel, thus "Windows XP" != "Windows NT kernel". Now, OS = "Windows XP", kernel = "Windows NT kernel", but "Windows XP" != "Windows NT kernel", so OS != kernel in this case, contradicting your assertion. Thus, either WP needs overhauling (you don't seriously believe this would be an isolated mistake, do you?) or your assertion is false.

Why are you mixing windows in here? Windows!=Linux and i frankly do not care if the Windows article is correct or not. (you "logic" btw fails by assuming that successors need to be identical on the kernel level, and your "=" should be subset operators). And frankly WP:OTHERSTUFFEXISTS is not a good argument. --Kim D. Petersen (talk) 13:46, 26 July 2008 (UTC)

You did, after all, assert that OS = kernel. MS Windows is an OS, so it is fair game. I just as well could have used the FreeBSD OS and you would still be wrong, just see freebsd.org. Freed42 (talk) 17:48, 26 July 2008 (UTC)

We are talking in the context of Linux here. And in the context of Linux (or for that matter FreeBSD) the kernel is the Operating system, as defined by OS theory, and the experts that work in that context (Tanenbaum etc). Read the Operating system article for an overview. (mem-management, hardware interface, device drivers, proces-handling ....) --Kim D. Petersen (talk) 17:59, 26 July 2008 (UTC)

What you refuse to acknowledge is that something like "FreeBSD OS" does not always refer to a kernel. It may refer to a proper superset of this, i.e., more than this. If you think that such references are simply wrong, just say so. If I take your word for this, and your word is correct, then I am justified in, say, making the many needed corrections to articles in WP and elsewhere. Freed42 (talk) 19:09, 26 July 2008 (UTC)

In that case you are even more wrong - since then both the Kernel and GNU LOC contributions disappear in the plethora of contributions such as X11, OpenOffice, KDE, etc. So now it would be GNU/MIT/Sun/Trolltech/... /Linux. --Kim D. Petersen (talk) 20:17, 26 July 2008 (UTC)

You are lucky that this and other objections are rebutted in one convenient place.[16] If you wish to counterargue any rebuttal, I am ready. Freed42 (talk) 22:24, 26 July 2008 (UTC)

The "rebuttal" there is based on feelings - not rational argumentation. And frankly, citing an GNU advocacy document is not really NPOV is it? --Kim D. Petersen (talk) 23:54, 26 July 2008 (UTC)

The rebuttal is based on fairness and giving credit, two values that do not imply anything about requiring "feelings". The document in this context is superfluous anyway, however, since the argument is straightforward: give credit where credit is due and be fair. Such considerations underlie activities associated strongly with rationality: such as scientific research (publishing, citing, cooperating, etc.). While in principle, there may be absurd cases, "A/B/C/...", in practice, it is obvious that reasonable compromises can be reached. Since the "Linux" side depends on citing google search results for suggesting popularity of the term, it is fair to consider an example of how established the "GNU/Linux" compromise already is for describing GNU distributions based on the Linux kernel: one need only go to www.google.com, search for GNU/Linux distribution[17], search for Linux distribution[18], and then note that the hit counts are within a mere 5% of each other. Such a simple search is a counterexample to the notion that "GNU/Linux" is used by some insignificant minority. It also casts doubt on the argument that "one should use 'Linux' because 'everybody else does'." Not that that "argument" has any validity anyway (see Argumentum ad populum). Freed42 (talk) 04:12, 27 July 2008 (UTC)

Your google search has a problem (the "/").. "Linux distribution" [19] "GNU/Linux distribution" [20]. --Kim D. Petersen (talk) 08:24, 27 July 2008 (UTC)

I don't see the problem. After all, the very top hit in the results of GNU/Linux distribution ([21]) points to www.debian.org, with the text: "Debian GNU/Linux is a free distribution of the GNU/Linux operating system.". How often are these terms associated with each other? To properly accomodate the flexibility of language--obviously demonstrated by hits such as that very first, it is only natural to plainly search for these terms and not artificially restrict the search with quotes. That does not mean I think my search is free of problems. Consider that "GNU/Linux" has a precise meaning in contrast to the notorious ambiguity of "Linux" (hence the disambiguation page for it). Thus, the Linux results will always be systematically overcounted in comparison. That's one of the problems with vague language. Of course, there is also a systematic overcount of "Linux" being counted as a substring of "GNU/Linux". A whole set of problems involve pages that merely contain the two terms used in an unassociated way, but, of course, that affects both sets of results. Note, however, that I merely presented counterexamples and never tried to make a positive claim, in contrast to the "Linux" argument as featured in GNU/Linux naming controversy. The "Linux" argument depends on citing inherently unreliable statistics fraught with difficulties to make a positive claim. Freed42 (talk) 15:31, 27 July 2008 (UTC)

Since the first result page from your search[22] contains several results[23][24][25] that doesn't even contain the word "GNU" - it should be obvious that something went wrong. For some reason just applying "" around `GNU/Linux distribution' doesn't search for the specific text. On the other hand searching for the specific phrase "GNU/Linux distribution". (in advanced search) - you get correct results. Which was what my searches showed.

In all likelihood the / acts as a search operator (while not documented anywhere i could find) for other than numbers. [26] --Kim D. Petersen (talk) 20:44, 27 July 2008 (UTC)

Thanks for pointing out that data anomaly, but it turns out to be a moot point because correcting for it does not change the conclusion, i.e., that "Linux" and "GNU/Linux" are similarly associated with "distribution". There are any number of characters, undocumented or not, that allow us to bypass "/". E.g., substituting ".", " ", "?", or " " for "/" all give results similar to "/". To find out why terms do not show up in a page, see the Google Help Center[27]: "Google looks not only at the content of the page itself, but also at the anchor text of links that point to the page." Note that this criticism applies at least as much to "Linux" as it does to "GNU/Linux" if we assume that "Linux" has at least as many hits. Your claim of "correct results" is false, since your search would not contain pages such as the top "Debian" hit that I pointed to, with the possible exception of help from "anchor text" (which you fail to acknowledge). If any of these findings inspire less confidence in trying to show "use", then great, because that was my larger point anyway. Indeed, there are more serious defects with the "Linux" citation in the article. We have already seen how much context matters for the meaning of something, so I will have to study "context" and then later show additional flaws in a citation that has been given a free ride. Freed42 (talk) 16:19, 28 July 2008 (UTC)

Correction: my mention of " " above is with the understanding that "GNU Linux" would have to be put in quotes. Freed42 (talk) 16:43, 28 July 2008 (UTC)

Its very simple - a search for "GNU Linux distribution" (with exact phrase) vs. an (exact phrase) search of "Linux distribution" gives us 588,000 [28] vs. 3,320,000 [29]. But lets try to unbias the search - Linux distribution w. mention of GNU vs. Linux distribution wo. mention of GNU: 977,000 [30] vs. 1,720,000 [31].

No matter how you slice or dice the search Linux is the preferred wording. (btw. take a look at User:KimDabelsteinPetersen/LinuxWeight, where i've already done some searches in various contexts, although i'm probably going to have to redo these based on the / - but that will only make the GNU results lower...) --Kim D. Petersen (talk) 21:58, 29 July 2008 (UTC)

Thanks for the search data. I do not disagree with your post, although things like unbiasing are extremely tricky I suppose. My criticisms about the statistics are about the ones that are currently used in the article (google *search volume* data instead of *search result volume* data), systematic "linux" (substring, ambiguity, etc.) overcounts, and people taken for granted that "GNU/Linux" is used by a fringe minority when there are easy (e.g., just type two terms) and clear counterexamples to the contrary (i.e., results that show significant if not majority usage). Freed42 (talk) 22:33, 29 July 2008 (UTC)

What you appear to get wrong is context here. I'm not saying OS=Kernel in all cases. For most Unixen and for Linux in particular, the blocks that are defined as Operating system is the kernel. This may or may not be the case for other operating systems. In particular it is not the case for operating systems that are based on a micro-kernel (such as Minix), but is the case for most OS's based on monolithic kernels (such Linux). --Kim D. Petersen (talk) 13:52, 26 July 2008 (UTC)

In such a controversial article, if you do not mean in all cases, then you should take some time and carefully qualify yourself. I am not asking for 100% precision on everything. Please show what is so special about the Linux kernel, that it by itself is enough to be called an OS, in all contexts. That needs to be shown, that is what is being disputed. Again, plenty of other kernels are not so special. E.g., FreeBSD is not like this--one can refer to the FreeBSD OS and not mean only a kernel. Freed42 (talk) 17:48, 26 July 2008 (UTC)

Try to fix in your head: We are talking about Linux here, and what it should be called. That is the context. If you have no background about what operating systems are, or what defines a microkernel vs. a monolithic kernel, then perhaps you should read up on it - before starting to argue about these? --Kim D. Petersen (talk) 18:03, 26 July 2008 (UTC)

You have not even proven one error in what I have stated and yet suggest that I have no background. This thread is based on what Gronky posted. I assume he meant only systems based on GNU and Linux. Simple question: do you think that Gronky meant something else? Yes or no? Freed42 (talk) 19:09, 26 July 2008 (UTC)

arbitrary section break

Since your comment about Windows is 100% irrelevant in this context - i don't see any reasons at all to point out the mistakes in your faulty logic.

And i assume that Gronky was referring to the thing that we are discussing here: Linux in general. Not some invented subset of that. --Kim D. Petersen (talk) 20:20, 26 July 2008 (UTC)

I like to learn from any mistakes pointed out to me, is that not good enough reason? :) Your assumption about what Gronky meant may very well be wrong--see a previous comment above where I address that issue. Freed42 (talk) 22:24, 26 July 2008 (UTC)

Now, you said, "You just did - you claimed that after someone has built an OS - it should be named after the components - despite that they had no activity in creating the OS. Either GNU has no requirements for naming here (per GPL) or you are contradicting yourself."

First, you are putting words in my mouth here. Please quote my words that you disagree with here, and I will respond. Second, if you still insist that "OS" = "kernel", then we have bigger issues, like, uh, overhauling WP.

Your other remarks simply do not relate to why LOC is not a valid way to show primary authorship in the Subset. Moreover, do not labor under the impression that I have ever used primary authorship of the Subset as the sole reason for how to label the Subset. Those other reasons are, after all, summarized on this very page. Freed42 (talk) 02:43, 26 July 2008 (UTC)

In case you needed another example of how WP conflicts with your assertion that OS = kernel, note that the leading sentence of Kernel (computer science) is "In computer science, the kernel is the central component of most computer operating systems (OS)." That implies that kernel < OS, contradicting your claim that OS = kernel. So, again, either your assertion is false, or we must correct this WP page like so many others. Freed42 (talk) 04:24, 26 July 2008 (UTC)

Do you know that you can't bring Wikipedia articles as references or arguments in a debate about content in another Wiki article? First of all Wikipedia is not a reliable source and second of all articles on Wikipedia are changeable. Also is not polite to ask people to change other articles, we discuss here about THIS article not about OTHERS. -- man with one red shoe (talk) 04:43, 26 July 2008 (UTC)

First, where did I appeal to WP as an authority? Instead I posited that if WP conflicts with KDP's assertion, then either KDP or WP is wrong. Do you understand? In any case, is trying to maintain consistency within WP forbidden? Second, where did I ask for KDP to change an article? Third, a quick glance shows that I am not the first user to have referred to other WP articles in the discussion for the current article. You yourself did the same, repeatedly. Are you assuming things about me again? Freed42 (talk) 05:31, 26 July 2008 (UTC)

Interesting where did I refer here to other WP articles? You've been on Wikipedia for such a little amount of time and you already had time to study my contributions... But in any case if I did do that, I made a mistake, there's no reason for you or somebody else to make the same mistake, right? Again, we are not here to discuss the validity of "Operating Systems" article or the changes needed in that article as such I don't think we need to discuss things based on the definitions in a specific article in Wikipedia, and Kim D. Petersen brought a clear example where kernel is treated as an OS: Tannenbaum, you say you don't appeal to WP as an authority, but that's exactly what you do when you oppose Wikipedia's content to a Computer Science renowned professor. Granted there's a debate where you draw the line of demarcation between OS and the utilities, it's not a fixed thing in Computer Science, but Petersen brought a relevant argument that's supported by a reliable source, your counter-argument used Wikipedia. -- man with one red shoe (talk) 06:00, 26 July 2008 (UTC)

You referred to other articles from this very article itself. Is there something wrong about me studying the article that I am editing? Anyway, if I assert (1) "X is true because ABC says so", then we agree that this is an appeal to an authority, right? Appealing to an authority must involve asserting some statement whose credibility depends on the credibility of an authority. However, if I assert (2) "if X is true according to DEF, then either DEF is correct about X (i.e., X is true), or X is false", in no way did I appeal to DEF as an authority for (2); I have merely stated something that is impossible to dispute; whether DEF is an authority at all or indeed whether DEF is correct is irrelevant to the truth of (2). The upshot of (2) on an article discussion is immediate, hence my repeatedly stated wishes to correct WP (including this article) if KDP's assertion is indeed true.

You may mention a flexible line of demarcation, but we cannot suppose this based on KDP's assertion. The following, however, suggests that Tannenbaum's line is flexible. On the one hand, KDP asserts that OS = kernel. On the other hand, Tannenbaum's very own current OS project[32] notes that, "MINIX 3 is a new open-source operating system." Let's assume that Tannenbaum's official MINIX 3 home page reflects his opinion about his own OS. On this very same page is a list entitled "MINIX 3 features", i.e., features of the MINIX 3 OS. Again, that's what "MINIX 3" is defined as on this official home page, as an OS, right? This list contains the item "Languages: cc, gcc, g , perl, python, etc." Thus, these languages are features of the MINIX 3 OS, i.e., they are a part of it. If KDP is right, that OS = kernel, then these languages are included in the kernel for this project. I think that even KDP will agree this is highly doubtful and therefore will have to admit that "An OS is defined as a kernel.", in the unqualified form that has been asserted by KDP, is not always true, i.e., it's false. Of course, there is the possibility that this MINIX 3 home page simply makes a big mistake here. Freed42 (talk) 09:56, 26 July 2008 (UTC)

You need to read some books on Operating System theory, more specifically what defines an operating system, and the difference between micro-kernel and monolithic kernel approaches. And you need to try to learn to separate absolute statements, from statements said in Context. The statement kernel=OS is true in context, false in general. It is true for Linux and most Unixen that the kernel is the OS. It is false for microkernel based architechtures such as Minix 3, where amongst other things device-drivers and memory-management (and lots of other things) have been moved out of the kernel. --Kim D. Petersen (talk) 14:09, 26 July 2008 (UTC)

I will quote your own words to you. You said that "GNU was the primary authors of the operating system" is counterfactual, i.e., it's a false statement. I proposed LOC as a way to determine primary authorship, and your first argument against this was, "In text-books about operating system theory the operating system is defined as the kernel". Here, you are clearly trying to take a statement from one context and apply it to our context, hoping that it will stick. All I needed to do was simply present you with counterexamples to cast doubt on your attempt, and indeed that is what I have done. The microkernel versus monolithic kernel information is all well and good, but your knowledge here cannot help you point to any error in anything I have stated. There are a ton of examples like the "Windows XP"/"Windows NT kernel" that I mentioned. I just mentioned "MINIX 3" not because of anything about a microkernel but because Tannenbaum is the very same authority that you appealed to. Need a Unix-like system example? E.g., would you demand that the FreeBSD OS must always refer to a kernel? If you recognize the importance of contexts, then you should be willing to make a qualified assertion to support your reasoning. Unless you can explicitly qualify your assertion to suit our discussion, I can only conclude that your case is built on fallacies, such as hasty generalization[33].

In short, you cannot have it both ways. You cannot expect to say, (1) kernel = "Linux", (2) OS = kernel, therefore (3) OS = "Linux". The problem is (2). If it does not always work for other OSes, microkernel, monolithic kernel, whatever, then what makes "Linux" so special? Freed42 (talk) 16:48, 26 July 2008 (UTC)

I'm sorry if you misunderstood my point about operating systems, i was operating under the assumption that you had background for understanding my argument. So let me summarize: The Linux kernel is a monolithic kernel, that in and by itself encompasses all requirements and definitions for an Operating system, (with very few exceptions - (bootloader and init). Therefore to use an argument that GNU project is or or could be the primary authors of the operating system (Linux) is a counterfactual rewrite of history. The GNU toolkit is not an operating system - Linux (specifically the kernel) is. --Kim D. Petersen (talk) 18:11, 26 July 2008 (UTC)

Let's put into a simple, logical form what you are doing here. What you are trying to show is, (1) "Linux OS = Linux kernel". You note that other things beside the Linux kernel, init and a bootloader, are needed for an OS, so you are implying (2) "Linux OS = Linux kernel init bootloader". Let's assume that (1) is correct. Well, (1) and (2) are simple logical identities, right? Therefore, by (1) and (2), we see that "Linux kernel" = "Linux kernel init bootloader", which is nonsense. Clearly, (1) and (2) cannot both be right. So is (1) correct or instead is (2) correct? If you mean something else, then please state it precisely enough to withstand logical scrutiny. Freed42 (talk) 21:10, 26 July 2008 (UTC)

I'm so sorry. I said both (1) and (2) - yes. That was by your dissection of what i said wrong - but if you had any sense of proportion, you'd know that the bootloader and init are components so small that they dwindle away in the big picture. In all but a pedantic nitpicking sense - the Linux kernel comprises all aspects of an Operating system. And GNU is not part of it. --Kim D. Petersen (talk) 23:14, 26 July 2008 (UTC)

In a discussion involving terminology, which is the focus of this talk page, it is entirely natural to make fine distinctions and precise points. Although you downplay the importance of elements such as init and the bootloader, they are essential. In any case, we have acknowledged so far that an OS based on the Linux kernel requires at least the Linux kernel and other essential elements. We have also seen that different "contexts" are necessarily a part of interpreting terms such as "OS". How can we be sure that our notion of an OS meets the minimum necessary elements for a given context? After all, the contexts may be numerous. You might mention XYZ OS as having the minimum necessary elements (say, kernel, init, bootloader) for an OS in one context, and yet still see "XYZ OS" in Tannenbaum, in which the pedagogical context is the XYZ kernel. How can we be so certain that there is not a context in which someone thinks "XYZ OS" as being the combination "the minimum necessary elements PLUS one extra utility"? Do we preemptively forbid that possibility, ruling out a multitude of contexts, e.g., does adding a kernel debugger to XYZ OS immediately invalidate the name "XYZ OS" to stand for the combination? If the combo is not called "XYZ OS" or even an OS at all, what is it called then? Freed42 (talk) 01:50, 27 July 2008 (UTC)

Why don't you sit down, and read the Wikipedia article on Operating system, perhaps even frequent your local library for some of the reference text? That way you would have a background for arguing what the "contexts" are. But its actually very simple an Operating system is the interface between user-programs and hardware, its the resource manager (processes, memory, interrupts,...). Once you've found a coherent argumentation for stating that GNU must be part of the Linux operating system, we can continue this discussion. --Kim D. Petersen (talk) 02:04, 27 July 2008 (UTC)

Ah, but I never implied such a thing, that GNU must be a part of an OS based on the Linux kernel. Moreover, I have already repeatedly referred to non-GNU systems based on the Linux kernel. Since you chose not to answer my simple question about "XYZ OS" and based on your suggestions that I do not know what an OS is, can I assume then that you assert the impossibility of labelling as an OS the combo (XYZ OS kernel debugger), as I presented? I.e., no, the combo cannot be called an OS under any circumstances? Freed42 (talk) 03:01, 27 July 2008 (UTC)

And i've already given you examples of Linuxen without GNU. Most uClinux based system are non-GNU (my home contains at least 4 of these - and they are not home-made but commercial products). As for your "XYZ OS" example - i'm referring you to the one location where you can find the requirements for an OS. (our article, as well as text-books). A kernel debugger is a system application - and not a required part of any OS that i know of. --Kim D. Petersen (talk) 03:09, 27 July 2008 (UTC)

I will take that as a "NO", that you believe that under no circumstances, can the combination "kernel init bootloader kernel debugger" ever be called an "OS". Freed42 (talk) 03:21, 27 July 2008 (UTC)

Then you would be deliberately putting words into my mouth that i neither support - nor in any ways have indicated. Perhaps if you tried to actually understand what is being written, you'd fare better? Especially since such a thing both exists and has an extremely useful purpose in embedded development. - they are used for debugging hardware/kernel issues. --Kim D. Petersen (talk) 03:49, 27 July 2008 (UTC)

OK, I'm sorry--I will take the words out of your mouth, then. Instead of me asking for an answer that you seem to be unwilling to give, suppose I try to read more of what was written and then came up with a "yes" or "no" answer for myself. If I later gave that answer to you, would you tell me if I was correct? Freed42 (talk) 04:25, 27 July 2008 (UTC)

Maybe the two arguments can indeed be merged. I'll give that some thought. But since it was in the "pro-GNU/Linux" section rather than the "everyone agrees" section, I'm not sure if your disagreement is a good reason for removing the point. If that's how we were doing it, I'd just clear all the "pro-Linux" arguments since I disagree with them :-) --Gronky (talk) 22:32, 2 June 2008 (UTC)

Gronky, its a collaborative project. That means that we remove things that are demonstratively incorrect. If you think there are any incorrect statements in the pro-linux (whatever that means) or any other section. Then you should remove them - and discuss it, if challenged. --Kim D. Petersen (talk) 22:41, 2 June 2008 (UTC)

Components of Linux-based and GNU-based OSes

As with most naming controversies, I believe this one can be resolved by providing much more detailed information, and focusing on informing, rather than dictating to, the reader.

What are the 10 most used modern Linux-based operating systems? What are the major component projects/library-sets of each? Split the answer into major components of GNU/Linux and other systems.

Repeat for the top 10 OSes that use GNU tools. What are the major components of each? Do they all include the full set of 319 GNU packages? different subsets? Explain the background of the differences; how are these decisions made?

sj 15:05, 20 July 2008 (UTC)

I don't get you. Insofar as the argument is "should Wikipedia refer to the OS in the general sense as Linux or GNU/Linux", either you're arguing for including an additional sentence-worth of material to every such reference including more details on the dispute, or you're siding with the argument that the term used should be context-sensitive depending on how much GNU code each particular subject makes use of. Either that or you're addressing a different argument. Chris Cunningham (not at work) - talk 15:21, 20 July 2008 (UTC)

Socks should not get a vote on this

I find it funny that very recent users that were obviously banned before on Wikipedia and which are SPAs (they edit only "GNU/Linux" controversy articles) try to influence this issue, I propose we ignore them. Thanks. I also ask admins who watch this discussion to pay attention to stuffing of ballot kind of behavior, using a sock for double voting or supporting an opinion is forbidden. -- man with one red shoe (talk) 17:19, 25 July 2008 (UTC)

Personally, I prefer to stick to the subject at hand. BTW, I have been using WP very recently, and I invite anyone to find whatever they want about me, since I have nothing to hide, such as the fact that I was banned for running a Tor exit node (I was unaware that this automatically banned me from WP). Now that everyone knows that I was banned, it seems that I don't pass your legitimacy test. You mentioned "users", so how *many* other users besides me are not measuring up? Is it a positive integer? —Preceding unsigned comment added by Freed42 (talk • contribs) 18:51, 25 July 2008 (UTC)

I used plural because I talked in general. That's one principle of law, it should apply to everybody. So, you actually admit that you have been banned on Wikipedia and created a new account? Just want to make this clear. I hope an admin will take the appropriate measures. -- man with one red shoe (talk) 21:50, 25 July 2008 (UTC)

No, he admitted that he was blocked for running Tor on his current account - something which he's never hidden. Please stop accusing people of being sockpuppets when you don't have a shred of evidence - there's enough bad faith around here without people looking for trouble. And this isn't a vote. Chris Cunningham (not at work) - talk 22:02, 25 July 2008 (UTC)

OK, my mistake, but I tend to assume that when people with new accounts who don't edit other subjects repeat the same tired (and refuted) arguments made by some banned people. -- man with one red shoe (talk) 22:27, 25 July 2008 (UTC)

A thought which some may find helpful

One of the oldest sayings that we've had about American versus British English is "Please relax". (Although, strangely, I find no evidence of this in google, which may be an indicator of how unrelaxed Wikipedia has become, I'm sorry to speculate.)

It seems to me to be quite problematic for people to go around enforcing one POV or the other when this is a complex editorial matter. Relaxation and harmony, and a firm refusal to allow people to campaign for one or the other, strikes me as a reasonable thing to do. --Jimbo Wales (talk) 19:32, 3 March 2009 (UTC)

There are lots of people on this bus

As Jim Gettys, the founder of the X windowing system, put it, "There are lots of people on this bus; I don't hear a clamor of support that GNU is more essential than many of the other components; can't take a wheel away, and end up with a functional vehicle, or an engine, or the seats. I recommend you be happy we have a bus."

If You think the Linux family of operating systems should be called GNU/Linux, you really should insist on calling it Apache/GNU/Busybox/GNOME/GNU/Java/KDE/LILO/LLVM/Mozilla/Perl/Slackware/TeX/Minix/Xorg/Xfree386/Linux. It is disrespectful to attempt to put GNU's contributions to the Linux family of operating systems (yes, Linux is a family of operating systems, not just a kernel) above the work of all of the other contributors.

Also, please read How much GNU is there in GNU/Linux?:

"Figure 1 shows the total LOC in Ubuntu natty split by the major projects that produce it. By this metric GNU software is about 8%. I didn’t include GNOME in the GNU category because it seems to now be effectively run outside GNU but including that the total for GNU would be around 13%."

"I found two things to be really surprising in this chart. The first is that the kernel is actually comparable in size to all the GNU software1. The second is that small projects actually dominate the total amount. It seems that at least for what Ubuntu packages, the origin of the software is highly dispersed."

On the subject of how much GNU contributed to the Linux family of operating systems, on of the most popular versions of Linux is simply called "Android", and nobody insists on calling it "Linux/Android". Care to guess what percentage of Android is GNU? Hint: Apple has a strong anti-GPLv3 policy that has lead them to replace every GNU component with non-GPLv3 alternatives.

Tizen (and its predecessor MeeGo) is another operating system that is based upon the Linux kernel but has no GNU, and the Yocto Project has an option to exclude all GPLv3 licensed software. Far from GNU being the indispensable component that the GNU/Linux advocates insist it is, it turns out that making a Linux OS without GNU is possible, and has been done multiple times. GNU software is convenient -- it is usually far easier to simply use the GNU component than looking for an alternative -- but in no way is it essential.

I highly recommend reading all three chapters of...

• Labyrinth of Software Freedom: (BSD vs GPL and social aspects of free licensing debate)

...by Dr. Nikolai Bezroukov. --Guy Macon (talk) 05:05, 7 June 2020 (UTC)

"One so-called freedom that we do not advocate is the 'freedom to choose any license you want for software you write'." --Richard M. Stallman[34][35][36][37]

--Guy Macon (talk) 06:26, 7 June 2020 (UTC)