Talk:Linux/Archive 1

From Wikipedia, the free encyclopedia

 This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
Archive 1 Archive 2 →

Contents

GNU/Linux

Please remove this long GNU/linux discussion. Nobody is intrested in RMS 's dogmatic egocentrism. It take to much space of the article.

Please open up a new GNU/Linux site.

Just for the sake of correctness (not political correctness though) shouldn't Linux be a link to GNU/Linux and the body of the article go to GNU/Linux? --Damas

I scrupulously call it "GNU/Linux", or even simply "GNU", as the kernel does not matter that much to the final user, but Wikipedia tries to intelligently report on the current state of things. Therefore, while it is true that "Linux" is (at least originally) the kernel, more than enough people call the OS "Linux" to warrant a section on it under this heading. Let's use GNU/Linux as the place where we discuss the reasons why the current usage is proper or not (and I believe it is not proper).

Changed GNU/Linux to Linux for legibility (how do you read that?). Cleaned up the explanation of the use of GNU/Linux in the last paragraph, with a link to the naming debate, and a short description of the distinct responsibilities of the Kernel and the system libraries. -- ksmathers

The article as it is is heavily slanted towards enforcing the "correct" terminology of calling Linux just the kernel, not the OS. This is obviously wrong and biased; calling the OS "Linux" is enormously widespread and cannot be dismissed as wrong usage in an encyclopaedia article. Even the article itself refers to the "OS" being adopted by manufacturers, while the "OS" is nowhere mentioned before that in the article.

The bias needs to be corrected. -- AV

Wow. I see Asa has decided to throw down the GNU/Linux vs Linux gauntlet. Even though I basically support the use of GNU/Linux, it seems that the entries are better served by simply using the popular term "Linux" with the good explanation of use in the Linux and GNU/Linux entries. --The Cunctator

heh. i didn't read this page until after i did it  :). maybe i should have! still haven't read it much. maybe tomorrow night  :) Well lets see if anyone disagrees with me enough to go thru all of them. The problem with doing that is that many references to the OS in Wikipedia were already GNU/Linux anyway! I think i have been reasonable with the terms on which i have made the changes. See my personal page for details. -- Asa

I agree that making the "Linux" page a pointer with actual explanatory text is better than the simple redirect. But let's not put any actual descriptive content on the Linux page that really belongs on one of the pages to which it points, otherwise we risk duplicate or out-of-sync data. --LDC

Don't forget, Linus was originally going to name the project "Freix" (pronounced like "freaks") but a friend talked him out of that idea and suggested "Linux" instead.

I removed the following from the article, as it is not actually an argument, but just another instance of a naming issue like linux vs. gnu/linux:

-- one could certainly replace the vast majority of Solaris's tools with GNU equivalents, for instance, but very few people would then call the result GNU/Solaris.

And can someone give some examples of distributions that use the linux kernel but no GNU tools (ie. not compiled with GCC, no bash, etc.)?

there are a number of Linux distributions (particularly those for embedded or dedicated single purposes) that include no or very few components of the GNU project, opting to replace them with BSD equivalents or specialized rewrites.

--snoyes 03:57, 1 Jan 2004 (UTC)

Yes. LRP is one (linux Router project). Uses uclibc (not glibc) and busybox for 99% of the tools (including the shell). Check the vast majority of embedded Linux systems. They cannot afford the resources of the GNU toolchain, and almost always never glibc.

As for "not being compiled with GCC"; that's a strawman. Is it GNU/Apache, GNU/X11 and GNU/Postgresql?

I don't agree with the GNU/Linux naming scheme (as you probably guessed), and while the arguments for it are fairly reasonable, I felt some reasonable argument against were missing. I had not realized the 'pedia was a political forum, and added the information that was missing to present a (more) balanced view. Silly me, I have been careful to leave the opposing point of view intact.

--coren

WP is not a political forum. We write encyclopedia articles ... they should just be as neutral as possible. The arguments you added are still there, as you can see. --snoyes 05:08, 1 Jan 2004 (UTC)
My apologies. Kneejerk. Shame on me.  :-)
I /still/ think the GNU/Solaris project illustrates my point well, though. Tought it was too inflamatory? (And, incidentally, that example didn't come out of thin air-- I admin such boxen).  :-)

--coren


The most confusing part of this entry is the following:

Slightly over half of all its code (counting by line) was licensed under the GPL. The Linux kernel contained 2.4 million lines of code, or 8% of the total, showing that the vast majority of a Linux operating system is not contained in the Linux kernel.

Is only just over half of the source code GPL'ed? If so, what is the other license (or is it a mix of BSD licenses etc)? Or is it just untrue? As for the second sentence - that's a paradox which cannot be solved - how can the Linux kernel be 8% of the Linux OS? They are one and the same. Kabads

Most of the rest of the source code is BSD liscensed. The rest is primarily BSD-like liscences (such as Artistic (Perl) liscence, Apache source liscence, and relatives). Some distributions include non-free liscencese (e.g. SuSE supplies an installer, yast, that is not free software). There is no one standard for Linux, thus it is not possible to be more specific, and still be generally relevent. The figure of 50% is highly arguable, and depends on the counting method.


The Linux kernel is not the totatlity for the OS, expect in some extreme definitions of OS. For most definitions, including the practical the kernel requires a set of programs external to itself in order to do anything useful. In Operating System this is reffered to as the shell, although in common useage the term 'user space' or 'userland' is more common, with shell having a more specific meaning. For example, function calls detailed in section 3 of the manual (and required to be POSIX complient) are _not_ provided by the kernel, but by the c library. Also, the Single Unix Specification requires the presence of certain runnable utilites, also not part of the kernel. Syntax 03:05, 23 Jan 2004 (UTC)

I separated out the long, long GNU/Linux debate and put in a summary paragraph that is hopefully fair (could do with a point for "Linux", not just one for "GNU/Linux"). The reason is (again) thinking in terms of this article being read by the casual non-geek reader who may or may not be able to work a computer, but would like to know what this "Linux" thing is. - David Gerard 13:24, Mar 3, 2004 (UTC)

From my talk page, a discussion on the subject between me and User:Gnu:

I have exactly same things to say. You seem to mix up FSF, GNU and Richard Stallman. Stallman is not FSF. Please understand that. And just because the majority calls the system "Linux" doesn't mean the system as a whole is "Linux". It's just the way some people call it, and some other people call it GNU/Linux. In the strict sense "Linux" is the kernel, and you are spreading false information through the Linux page and paint Stallman is bad picture through GNU/Linux. If you don't know what you are talking, leave it to others who know what they are talking. -regards Gnu

I basically agree. It's still POV and advocacy even if I agree. Note it says Richard Stallman and the FSF, by the way. - David Gerard 09:51, May 28, 2004 (UTC)
I agree, but calling it "linux" is also "POV". We need to have a balanced view in there and present both the views.
That's why there's a section which leads to a lengthy sub-article.
Roping in Stallman's views, kind of, says that only Stallman wants to have it that way. That's clearly not the case.
Richard Stallman is a primary advocate of the position and is well known as one.
Most people who call it "Linux" instead of "GNU/Linux" do so because of ignorance, and this page is supposed to educate them, IMHO. --Gnu 10:20, 28 May 2004 (UTC)
Nope, It's supposed to be descriptive (NPOV), not didactic (POV). It's not a lecture.
Of course, yes. But, unfortunately, calling it Linux *is* also POV. Otherwise GNU/Linux page shouldn't have specifically pointed to the naming controversy. It should have been a page in itself!
There's a knotty history of pages being redirected, etc. I think (going by the talk) the current situation is because "Linux" is far and away the most common name, and after everyone wanted every possible controversy about Linux mentioned with all options in the first paragraph, they were (sensibly) cleared away to form a readable article in clear English for someone who knows nothing about this "Linux" they've heard so much about. Both names are mentioned in the intro, representing the other POV - David Gerard 13:37, May 28, 2004 (UTC)
As I said, I strongly urge you to read the past discussion in Talk:Linux and Talk:GNU/Linux naming controversy - these have been gone over at length. You (and I) may think it's incorrect to call the OS "Linux" rather than "GNU/Linux", but that's conventional usage; pretending it isn't would not be appropriate - David Gerard 10:59, May 28, 2004 (UTC)
Yes, I am reading it. Thanks.

Can I second the opinion that if the article makes reference to "GNU/Linux" then there should be an article at GNU/Linux (linked to from here) explaining how GNU/Linux refers to a GNU system with the kernel, Linux (and the debate surrounding that).

Or else (and even better IMO opinion) replace references to "GNU/Linux" (a confusing term which is awkward to say) in this article with something like "Linux-based GNU systems" which explains what you mean much better . (Then there does not really need to be an explanation though you could still have discussion of the issue and what GNU/Linux is at GNU/Linux and/or at Linux-based_GNU_Systems.)

--Joe Llywelyn Griffith Blakesley 12:49, 2004 Jun 18 (UTC)

Minix-like

I am changing Linux to not be a Minix system but a Unix like system.


Partisanship/Advocacy

"At the present time, Windows and Mac users can easily migrate to Linux if they use desktop managers like KDE or Gnome, and many free software packages offer the functionality of programs available on the other operating systems." is blatant and misleading propaganda. I like Linux, but that kind of OS partisanship belongs elsewhere. -- The Cunctator

"blatant and misleading"?

  1. Migration is easy as
    1. many Linux GUIs (Gnome or KDE offer the same experience as Windows/Mac
    2. many Linux programs allow easy import of (for example) Word documents
  2. It is a truism that many packages offer the functionality of those available on other OSes ... witness Gimp for example ... not "partisanship" as Microsoft (and Adobe et al) would not be half as worried if it were not true.

-John Lynch

"Thus, as many corporations and public offices are finding out, migration to a Linux-based system is thus not the chore that competing vendors would like one to believe." is better, but I don't think it has any place in the Linux page. If you'd like, you should make an OS advocacy page to hash this out. "Truism" doesn't mean what you seem to want it to mean. Gimp is a great program, but it's not identical to Photoshop. The assertion that Gnome or KDE offer the same experience as Windows/Mac is as laughable as saying that Windows offers the same experience as the Mac. They are all different systems, and have different strengths. I just think we should keep the advocacy outside of the descriptions, unless expressly indicated as such. A history of OS advocacy would be a great resource.

-The Cunctator

It's not an issue of advocacy ... the statement that "many corporations and public offices are finding out, [that] migration to a Linux-based system is thus not the chore that competing vendors would like one to believe", is true no matter what way you parse it.

Likewise, your reading of my statement that "it is a truism that many packages offer the functionality of those available on other OSes" is flawed. Gimp (for e.g.) offers the functionality of Photoshop. I was not arguing that they were 'identical' (as per your reading).

Lastly, you're assuming I'm a linux advocate.

"Lastly, you're assuming I'm a linux advocate." I never did, nor did I even imply as such. Your language ("It is a truism", "is true no matter what..") indicates a strongly held position; without supporting evidence, I must assume the assertions are based to some degree on faith or emotion. Perhaps such phraseology is simply a rhetorical tic; a "truism" is a self-evident truth. Water is wet; that's a truism. And it's not even a perfect truism, because of the dependence on temperature and pressure ranges. The equivalence of functionality between Gimp and Photoshop is certainly not self-evident simply by saying so. That kind of language, and the introduction of the issue of marketing propaganda and public opinion into a paragraph about the nature of Linux, is why I refered to such a sentence as "advocacy" and "partisanship".

--The Cunctator

Hello Guys I think with these kinds of issues It is probably best that we agree to differ and include both opinions in the article. I am really keen on linux and have introduced it to several organisation I work for. But, I agree that the sentence "many corporations and public offices are finding out, [that] migration to a Linux-based system is thus not the chore that competing vendors would like one to believe" is slightly biased and may be a bit misleading. It seems to say that you can replace _Any OS_ and _Application_ with a linux based alternative that is _just as good_. This is obviously not true. Firstly most users who use the non linux os and applications will find that linux alternativs to their applications will be well below par. On the other hand there are some specialised applications on linux which are WAY Better than the alterntives. At the moment for a certain small number of applications the Linux "alternative" is the best solution. However for Most Applications for the Majority of people, Linux is definately Not the best alternative.



Portability

Linus didn't develop Linux from the start with the goal of portability in mind, that came later. At the time, his goal was to learn about the 386. Hence, I'm changing that bit.

--Robert Merkel


GIMP vs Photoshop

Just stepping in with my two cents: Gimp does not offer the functionality of Photoshop; it offers most of the functionality of Photoshop. Photoshop allows a user to select and then cut and paste part of an image on any layer; Gimp does not. This was the chief, and (for me) most vexing difference between the two. --Koyaanis Qatsi


Linux wikis

Aside from the german linux wiki does any know of another? -dgd

Answer: We are working on doing it international / multi-lingual (therefore it is http://LinuxWiki.org now). We are just evaluating different possibilities HOW to do that technically the best way. If you want to contribute, ideas are welcome! -- ThomasWaldmann

A2: Very soon now we are going to start a Dutch one at nedlinux.nl. Yngwin 19:07 18 May 2003 (UTC)


Legal issues section

The legal issues paragraph is a horrible mess - and probably largely redundant with SCO v. IBM. Anyone feeling strong enough for a serious hack'n'slash job on it? - David Gerard 11:14, Jan 28, 2004 (UTC)

David,

I did so feel, and undertook it. It was more work than I'd first expected. Stevenj didn't agree and removed almost all of it. I replied to his removal and am awaiting a response (see UserTalk for Stevenj for the discussion).

Briefly, I don't think there is a clean and brief way to cover the issue(s), even for a Linux newcomer (I am taking this as an article of first reference). Nevertheless, all the buzz suggests that SCOG's claims are a major problem for Linux, and so for Linux users. I have spoken to folks who foresee non-licensed Linux users being hit with $100000 lawsuits rather like the RIAA slapped on all those people (including the 14 year olds here and there) recently. Thus, I think we're stuck with covering it here somehow. My edit was (not cleanly and briefly, to be sure) an attempt at illuminating the dim legal and historical recesses of the mess and at lifting as much as possible of the 'legal magic' veil covering it all.

Comments?

I agree that the litigation should be covered, but I think more than a paragraph about it belongs in a separate article. Moreover, a separate article already exists, and covers every issue that was in the deleted material (not only SCO v. IBM, but the Redhat lawsuit, Novell's claims, etcetera). If you feel that the other article is missing some material, you should merge in the previous material from Linux as appropriate. Steven G. Johnson 18:45, 5 Feb 2004 (UTC)
I think we agree on most of this. Brief coverage (the problem is how brief), more extensive coverage elsewhere (we agree that SCO v IBM is probably the right place), it was a long few paragraphs, legal stuff like this is a painful annoying mess which in a better world wouldn't exist, etc. We disagree that 'every issue' in the removed material is covered in SCO v. Certainly by following the links one might eventually..., but the point of a summary is to save just such chasing down. We also disagree on whether the summary material removed from here should be included at SCO v. It's too summary for that article, I think -- too much background, too much perspective, elides much legal detail, ... We also disagree about whether more than a pointer para is sufficient in a first resort article such as this one.
Accordingly, I still think the removed material (or something similar) belongs here -- in addition to more detail in SCO v.
Perhaps others might chime in here?
ww
The current article is better than restoring the previous version, which was mainly a huge unreadable paragraph. Also, "all the buzz suggests that SCOG's claims are a major problem for Linux, and so for Linux users" is itself a matter of hot dispute - it's not all the buzz by a long shot.
The way it's now done here - a "Main article" link then a para of summary - is the way it's done in most articles on an extensive subject of this nature. I suspect the Linux article should in fact mostly be along these lines, rather than keeping it in the main article. If there's a reasonably standalone subarticle, Wikipedia practice appears to be to put stuff in that article as far as is reasonable. Hence the similar move on the 'Distributions' section. (Of course, if you take it too far you end up with an article that isn't in fact readable in itself as an overview.) - David Gerard 13:30, Feb 6, 2004 (UTC)
I've extended the para with a mention of the threats to sue individuals. The SCO v. IBM article is also torturously ill-structured and needs significant work, but nevertheless I still think we shouldn't have more than a summary in this article. - David Gerard 13:48, Feb 6, 2004 (UTC)
David,
I agree that the SCO v article is torturously ill-structured but I suspect it cannot be otherwise. This is ongoing litigation with the significance of much not even vaguely clear to anyone, including the participants. Since it's an article about the lawsuit, it necessarily is messy and over-detailed and picky and complex and tortuous. If I (or indeed anyone) revised it for some kind of better stucture, an honest article would still be ill-structured. It reflects the underlying reality.
As for '...all the buzz suggesting...', indeed nearly all of the general purpose journalism I've seen on the issue has so suggested. Your basic journalist is trying to make a point or perhaps two in a story. From the coverage, it seems that 'big problem for Linux and maybe for its current and future users too' is a point they can understand, and think their audience may get as well, so it's the one they make. Nuance and underdetermined reality get lost; for those who think physics, it's a superposition situation in legal space. Perhaps editors see that as the point to make given the dramatic possibiltiies, and steer in that direction?
Anyway, the specialized press (ie, the computer folk) have done somewhat better, though they miss legal issues (and mischaracterize them when they don't miss) with some regularity. But it's the general press and its readers I had in mind, not any specialist press and its much smaller readership. Presumably they don't need an overview. The rest of us do, in my view.
As for unreadable, that is easily remedied in principle; rework it for more readability. Wholesale replacement by a pointer does not address readability, and hasn't in this case. If the problem in re readability is the content as opposed to phrasing, than I suggest that's not a deletion issue. The litigation (and connected issues) exist regardless of wording, and they are together a dog's breakfast, and they are a prominent issue in the minds of unspecialized observers aware of the lawsuit(s). That's why I put considerable work into revising that section for the non legally specialist reader looking for a first level reference (not detailed coverage!), and worked hard to leave out details without distorting the 'possible outcomes' picture. What resulted, however much the content is annoying and repellent, reflected a serious attempt from an informed observer at NPOV general coverage of the issue(s) for a general reader.
It was because I appreciate both now and then that an overview article has some obligation not to be merely a collection of pointers that I felt the litigation chunk needed to be revised. My revision WAS a summary, did not duplicate entirely the material over in SCO v (certainly not in perspective or structure), and was longer than any such thing ought in a reasonable world to be. Regrettably, at this stage of things, it was not possible (without adopting some sort of POV on the legalities and so dismissing this or that by not mentioning it) to shorten/save folks the pain of a balanced account of annoying legalistic stuff any further. But an overview (my term in this discussion has been 'an article of first reference') OUGHT to include, in this instance, more than pointers to a legally oriented (with all the consequent confusing detail for the non legals) article. An overview reader can reasonably expect such in an Encyclopedia article, and certainly in a Wikipedia article. Readers are not well served, on this issue, by the current state of this article.
On the issue of length of coverage devoted to the litigation in this article, I might direct your attention to the article's current (extensive) discussion of 'GNU/Linux' versus 'Linux'. In my personal opinion, this is more than a minor issue deserves. I suspect it was the content which offended, more than positions on article structure or readability. I agree and am also offended. It's my judgement, after the research that I did, that SCOG is balancing on no legs at all, that its past acts have removed even those, and that it is wasting all our time. But judges/juries have surprised me before....
ww
I think the essential problem is that the Linux article needs a top-to-bottom rewrite. It appears to have originally been cobbled together from pieces of Linux advocacy, so we end up with tortured sentences and paragraphs consisting almost entirely of a thicket of subclauses and parentheses to placate every conceivable viewpoint right there in the same sentence - rather than anything that flows readably. Various people are coming through and attempting to fix the writing piece by piece, though, which may achieve the result.
I expect we could go to a second paragraph on SCO v. IBM :-) It does sound like the main article on the lawsuit(s) could do with an overview of legal implications, though getting that NPOV will be more than a little work!
The GNU/Linux naming debate is also something that should be in a separate article, if anyone feels strong enough to write it. Again, getting it NPOV will be interesting. I may give it a go later. Unless there is such an article already, in which case almost everything on the subject here should be refactored into that and a summary and pointer placed here. - David Gerard 17:56, Feb 8, 2004 (UTC)
Please don't do a top-to-bottom rewrite - that's not how wikipedia works. The current version of this page is the work of ~100 contributors. Personally I'd like to delete all references to SCO since they should be on the Linux kernel page, but I'll let people that are interested in the SCO issue work that out. Right now I'm reviewing the "GNU/Linux" section, hopefully I can shorten it. I believe it's the most important issue on the page, but I think the explanation could be optimised.
If the layout of a page really bothers you, it might be time to turn off the computer and go for a walk outside - or maybe direct your energy to a page that doesn't have so many contributors. - Anonymous
I doubt I shall be bothering with the mooted top-to-bottom rewrite :-) But a paragraph whose structure has come together with the unspoken aim of placating each of a group of advocate special interests is quite different to something that flows readably. The piecemeal work and refactoring will get there okay, I suspect. - David Gerard 21:23, Feb 8, 2004 (UTC)
A top-to-bottom "rewrite", more appropriately it might be called a "rework", does not disregard the work of other contributors. A new, cohesive article "from top-to-bottom" may be what is necessary to make an article cohesive, internally sane, and sensible, but it is not to say that such a "rewrite" would not incorporate the information currently included in the article by other contributors, nor is it even to say that the entire text is different from the current text. While it proceeds from top to bottom, the information and the text already included in the article is incorporated into the new, cohesive whole. The data provided by contributors remains, and the syntax and diction themselves would remain the same in many areas. Indeed, entire paragraphs and sections might would go unmodified. We should not let stand some inferior work just because "it was worked". The article will be changed to remedy problems, and the work of other contributors will be overwritten by those changes. When this is done piecemeal, it lacks the coherence of a total reworking, despite the fact that the contributors' work will be overwritten. The contributors will be overwritten similarly, the difference will just be in the time of it. Indeed, piecemeal reworking may be worse in nullifying the effort of contributors, as it does not have the benefit of a sound base that will facilitate prevention of redundance.
If the sentiment of your statement, "direct your energy to a page that doesn't have so many contributors", is that one's time is better spent on other, less complete, articles, I remind you that this is not a zero-sum game. For less active contributors, an aversion from a certain article might be an aversion from spending that time on Wikipedia altogether. For others, the spark of interest of a particular day may only be in a certain article.
- Centrx 02:06, 13 Jun 2004 (UTC)

Intro as summary

"The term "Linux" is now even applied to whole Linux distributions, which typically bundle large quantities of software along with the operating system, such as web servers like Apache, graphical environments like GNOME and KDE, office suites like OpenOffice.org, and the X Window System."

I took this out because it was largely redundant with the ==Linux distributions== section, and now it's completely redundant. I thought it was sufficiently covered in the first para, evidently you disagree ... There must be a much shorter way to put this. - David Gerard 20:41, Feb 10, 2004 (UTC)

The introduction should also summarize the most important information, and a summary is necessarily redundant with the detailed text. So, I don't think it's a proper goal to remove all redundancy from the introduction. Note also that it wasn't covered by the first paragraph, because it is important to note that the word "Linux" is used for three things: the kernel, the Unix-like operating system (a.k.a. GNU/Linux), and the Unix-like operating system plus gazillions of free-software packages (a.k.a. distros). The distinction between these three is one of the major confusions among the general public about Linux, so it has to be covered clearly up-front. Steven G. Johnson 21:21, 10 Feb 2004 (UTC)
Hmm, true. I'll give it another go now. - David Gerard 23:40, Feb 10, 2004 (UTC)
Here's another go - thoughts welcome. - David Gerard 23:57, Feb 10, 2004 (UTC)
The name Linux is commonly used for three related concepts:
1. The Linux kernel, the project started by Linus Torvalds and led by him to the present day.
2. The entire free Unix-like computer operating system (also called GNU/Linux) that is formed by combining the Linux kernel with the libraries and tools from the GNU project.
(The kernel is not officially affiliated with GNU, but is independently developed and distributed under the terms of the GNU General Public License.)
3. Linux distributions: the kernel, the GNU tools and large quantities of additional open source and free software, such as web servers, the GNOME and KDE graphical environments, office suites like OpenOffice.org and other software.
Linux, in the form of its distributions, has experienced rapid growth in popularity, overtaking many proprietary versions of Unix and even challenging the dominance of Microsoft Windows in some areas. Linux distributions are available supporting a wide variety of computer hardware and have been deployed in applications ranging from personal computers up to supercomputers and down to embedded systems such as mobile phones.

I'm not sure I see it as an improvement to use numbered lists, like a PowerPoint presentation. What's wrong with sentences and paragraphs? Steven G. Johnson

The present paragraphs demonstrably don't make it clear that we're talking about three related but different things - they're a murky, badly-written mess. What is done elsewhere on WP in these situations? - David Gerard 08:53, Feb 11, 2004 (UTC)

The current first paragraph is intended to give the quickest possible intro to the casual non-geek reader who doesn't care yet: what it is, and why it's important. Complication should start with the second para, if it really can't be put off until a later section. - David Gerard 01:10, Mar 3, 2004 (UTC)

Just done some more on the intro paragraphs. The new second paragraph needs a clearer way of putting across what's in a Linux distribution - David Gerard 10:08, May 17, 2004 (UTC)

Tux image

I've linked to the large Tux rather than the small one ... not sure about the effect that transparent background gives, though.

Also, I still think "famous" is better than "recognised" - in the case of Linux, I really think "famous" is the right word, and "most recognised" strikes me as a clumsy construction. - David Gerard 23:43, Mar 2, 2004 (UTC)


Too much discussion on irrelevant matters

The introduction seems more fiendly to anyone, but still as mentioned, the 2nd paragraph is a kind of an "elevated" discussion not for the introduction. When some one says "Windows", it is understood that it's the whole thing. If the windows kernel is Dos or is a patchwork, its never left for an introduction, but for detailed sections. --Energiza 15:27, 3 Mar 2004 (UTC)

Brief is not Incomplete

Think what Windows gives you: Environment (you see it but can't exploit it), the monitoring-licensing method , few unimportant tools and that's it. Now Linux: IT GIVES YOU ALL: As server (web, ftp, telnet, database.., Server Programming) and As Workstation (Open Office, Drawing, X-Window and Programming tools TO EXPLOIT THE GRAPHICAL ENVIRONMENT) all at low cost, security and reliability, the lack of which keeps Microsoft in scandals all the time. Few sites say WHY LINUX IS SO IMPORTANT. If the introduction has to be keep brief, THAT'S OK, but has to have HARD FACTS. The kernel thing...most people think a kernel is a grain of some cereal, and they are right, so kernel discussions have to go further down or in an article on KERNEL. The same applies to the mention of GNU/Linux or plainly Linux. It's not something for the introduction, but for History perhaps.--Energiza 15:47, 3 Mar 2004 (UTC)

That's why the 'why it's important' thing reads "the best-known example of free software and open source development." Can you put it in a manner more comprehensible to a non-geek?
Also, you need to reread Wikipedia:Neutral point of view. This is an encyclopedia, not an advocacy site. The Linux article must, by NPOV, be something that someone who hates Linux could not reasonably disagree with. - David Gerard 16:12, Mar 3, 2004 (UTC)

David, I don't think this is POV dispute. There are reasons why people choose Linux over Windows and the intro part of the article mentions nothing about this. Maybe Enrgiza or my wording is problematic, then why do you give a shoot? Why do you think Linux has become an important operating system? -- Taku 19:04, Mar 3, 2004 (UTC)

Mostly because it's low-price and so is the software. But it would help if the paragraph didn't read quite so POV. The replacement is much more sedate and answers the "why is this important" question quite well, actually :-) - David Gerard 19:59, Mar 3, 2004 (UTC)

I've moved the detailed paragraph on the quality of HURD to GNU/Linux naming controversy, where it seems to fit better than in "History of Linux." Unless the history of HURD has greater relevance to the history of Linux than the fact that it wasn't ready in 1991 - David Gerard 17:31, Mar 4, 2004 (UTC)


Desktop share quotes

Removed this:

In 2004, the question of when Linux would experience significant market share on desktops was debated. During the Linux.Conf.au conference at the University of Adelaide in January 2004 (Linux Australia, n.d.), Torvalds made a statement of his own opinion on the issue (Gedda, 2004): "This year there will be a lot of desktop users," but later that month he said, "I mean it's going to take literally five to ten years before 'normal users' start seeing [the] Linux desktop" (Mackenzie, 2004). (It should be noted, however, that Torvalds has no direct involvement in user-level software for Linux-based operating systems.)

- there's no point in (a) putting in two contradictory quotes, then (b) explaining why the quotes are meaningless! All that's left is that it was debated in 2004, which is no different to it having been debated every year since 1995. - David Gerard 13:33, Mar 5, 2004 (UTC)


Linux vs Unix

There can be a section on how Linux deviates from Unix. Or if there is this information in some other article I can create a section by taking the contents. Jay 11:55, 6 Mar 2004 (UTC)

If you're sure there's a point - "Unix" deviates from "itself" considerably, SUS notwithstanding. There's Unix-like, for what it's worth. - David Gerard 12:44, Mar 6, 2004 (UTC)
This is confusing, now what do we compare Linux with ! Ok can we have a section on the unique features of Linux, that are not found in other Unix-like systems. Jay 13:06, 6 Mar 2004 (UTC)
Yes, it is confusing :-) Considered as a Unix, GNU/Linux is an odd one. But the others diverge considerably amongst themselves - I suspect it's not so easy to meaningfully say "Linux is like this, all other Unix are like this." Then there's the BSDs. A comparative chart might be useful (and there are many such charts in existence), but Linux may not be the right article to put it in. Perhaps working out a comparison between varieties, then deciding how much of it should go here? - David Gerard 13:40, Mar 6, 2004 (UTC)


Best-known distributions

If the number of search results for a distribution name at Google Linux is directly proportional to the number of users who know the distribution, a statistic can be easily extracted:

Search Results Distribution
2,580,000 Debian
2,360,000 SuSE
2,090,000 Red Hat
1,460,000 Mandrake
922,000 Slackware
825,000 Gentoo

(Every distro shouldn't be listed in that sentence; it would be better to remove it altogether than make it a platform for an advocacy battle. I use FreeBSD myself ...) - David Gerard 13:37, Mar 23, 2004 (UTC)

"Windows is more popular than linux" does not constitute advocacy of Windows. Similarly, it is merely a factual statement to list the best known distros. I don't think there's much doubt that RH, SuSE and Debian are the top 3 mindshare-wise. So the sentence is fine. Arvindn 13:46, 23 Mar 2004 (UTC)
By "advocacy battle" I mean it being randomly changed (without an edit summary) to Slackware. Or Mandrake or TurboLinux or whatever ... - David Gerard 13:47, Mar 23, 2004 (UTC)
As has just happened again - cheers to MikaelSorlin! I've removed the sentence in question - David Gerard 23:13, Mar 23, 2004 (UTC)

> What would be a suitable measure?

I'm not sure if it is suitable or not. But googling ""red hat" linux gives 4.2 million hits. "suse linux" gives 3.4 million hits. "debian linux" gives 3.7 million. But "slackware linux" gives only 1.1 million. So from that point of view slackware has definitely less 'google mindshare'.

Sander123 10:38, 24 Mar 2004 (UTC)

Therefore, David, the solution is Dispute resolution and not removal of the contentious sentence. -- Arvindn 13:09, 24 Mar 2004 (UTC)
If you feel you can be bothered taking it through that, go for it. Have a look through MikaelSorlin's other edits and see what you think. - David Gerard 14:49, Mar 24, 2004 (UTC)


Therefore, "the best-known examples are Debian, Red Hat and SuSE." Can Slackware really be held to be more popular or better-known than Debian? What would be a suitable measure? - 81.223.147.9

--suitable measure. -- How about a list of linux distribuitions that are known by non- linux / unix software programmers or amateurs? Originally, the list was meant as a guide for amateurs and beginners. (Change the title to Recommended Distributions for Beginners?) I bought a Linux book years ago. Caldera, Redhat were recommended. I am a technical journalist based in Japan and the guy who enterred the list. Montevista is popular in the news (Used in Motorola'sA760 phone, controversial because Motorola was supposed to be using Symbian )and for the TRON (well known in Japan)cpu architechture. Turbo Linux is popular here in Japan too. Suse is the Java Desktop OS of SUN. For myself, enter Slackware and Gentoo if you feel they are popular(no , I don't know much about them). Or should the list be limited to Debian Red Hat or Suse. ? Still what is a suitable measure? my .02 $ Jondel

---Not only the U.S.--- Hi! Please consider other countries like India, Japan, Europe ,etc..when determining what distros are popular and not only the US. Here in Japan, Turbo Linux is popular. I think Montavista is also popular. Google seems to concentrate on the US (?) . --Jondel 04:07, 18 Jun 2004 (UTC)

I tried looking at DistroWatch. They have a page headed "The Top Ten Distributions", but do not say what they are measured on anywhere I can find. If it's only measured on hits on DistroWatch pages for the distro in question, I'd question its utility as a measure. Also, I wouldn't list FreeBSD here just because DistroWatch list it ... - David Gerard 08:36, 18 Jun 2004 (UTC)
According to the edit summary, those distros aren't even in the top 100 of Distrowatch, which is one of the big online names in the Linux community - ask a question about which distro to use and you'll likely be sent to Distrowatch. And SuSe is a popular European distro. If you want something different than Fedora, Suse, Mandrake, etc., better to list something with a bit more of a high-profile like Lorma Linux. Johnleemk | Talk 09:51, 18 Jun 2004 (UTC)

What Linux is

First of all, Mr. Gerard, I did not change a single mention of "Linux" to "GNU/Linux" or use the term at all. I do not consider it a particularly useful term as it is a superset of the same "Linux" upon which distributions are based and is not significantly more appropriate than "GNU/BSD/Linux" or "KDE/GNU/Linux", as in the case of certain distributions which disregard all but a single desktop environment. Its use is a matter of specificity, not accuracy. My edit was also much more than a resolution of the current ambiguity and inaccuracy of stating that "Linux" IS all of the myriad software included in various Linux distributions. A lazy, wholesale reversion of this edit elided much useful information and clear syntax and diction. Because of the totality of the reversion and the inapplicability of embedded comment and the above Talk comments, it is not even clear what your objection to my edits is. Is it your goal for the article to remain rather static for all time? - Centrx 02:22, 13 Jun 2004 (UTC)

Please read GNU/Linux naming controversy - David Gerard 08:43, 13 Jun 2004 (UTC)
As I said, I am not asserting that the Linux base system should be named one way or the other in the article, whether "Linux" or "GNU/Linux". If you think it inappropriate that I removed the phrase, "also known as GNU/Linux", which was in earnest for smooth, clean text rather than a desire to rid the term from the article, then you could have very simply added it right back in without reverting the entire edit. I would not have objected at all and doing so would have made it much clearer what your dispute with the edit was, without the exertion of this analysis and response. Otherwise, I don't see what the naming controversy has to do with anything else in the edit. Please be more specific if there is any further objection. - Centrx 09:00, 13 Jun 2004 (UTC)
It appeared to be unnecessarily convoluted grammar and bad structuring to no gain. The revised first paragraph is the sort of thing one's eyes slide off. Imagine someone who's barely heard of this "Leenux" thing reading the article seeking enlightenment. You can start with simple sentences, or you can try to put every nuance in the first sentence and confuse them utterly.
(While the ridiculously large vocabulary available in English allows an amazing degree of precision in a single well-chosen word, "evinces" as a substitute for "demonstrates the effectiveness of"? That's showing off, not writing clearly.)
Your misuse of a factual accuracy dispute header when someone doesn't like your wording does you no credit either. Since you have specified no facts you have a problem with, I'm removing it - David Gerard 14:36, 13 Jun 2004 (UTC)
First of all, it seemed to me that you were disputing the factual accuracy of my edits, and I do dispute the factual accuracy of some of the text as it stands now. By no means is Linux the complete suite of tools found in a Linux distribution. When used in this sense, the word is always used as a descriptor, as in the case of "Linux distribution" where it is an adjective. When used so, it means that the distribution is based on Linux, but it does not mean the the compilation of software IS Linux.
Where is this unnecessary convolution and poor structure? As it stands now, it seems to me to be unnecessarily convoluted with the wordiness an impediment to the flow of reading. For instance, the entire second paragraph is a single sentence with many clauses that should be broken up. Generally though, I retained the same wording and the same general structure and order.
It is quite the insulting assumption that I was trying to "show off". This is simply the way I naturally write and it is not due to an attempt to impress others. Most people who read the page are not going to know that I wrote it. Only those who currently have the page on their watchlist might, if interested, find that I wrote it but thereafter I think it unlikely that anyone would be perusing the ancient history diff's of the article. "Evince" is the only word that may be impenetrably sophisticated and I have changed it in the revision below. However, "demonstrate" is simply not the correct word. Only an agent may demonstrate, and he usually does so by experiment or by deduction whereas "evince", inter alia, means "to be an indication or evidence of". This is an excellent example of how simplification can be misleading or inaccurate. There are no synonyms; every word has a distinct valence. There is a wealth of language available to us; this is not the Simple English Wikipedia. The only other word that might be considered more sophisticated than the previous version is "descriptor" and it is, if not common, obviously and penetrably similar to "description" with a normal postfix in the language (actor, orator, demonstrator). A new revision is below; please indicate what you find objectionable, preferably with specific edits. - Centrx 18:51, 13 Jun 2004 (UTC)

In the field of computing, Linux is a free operating system kernel produced through the worldwide collaboration of developers. It is likely the most famous example of free software and of open-source development.

As a descriptor, the name also refers to software compilations, called Linux distributions or GNU/Linux distributions, with operating software that is based on the Linux kernel and combined with libraries and tools from the GNU Project. Such distributions typically contain many software packages, such as software development tools, databases, web servers like Apache, desktop environments like GNOME and KDE, and office suites like OpenOffice.org.

Originally developed for 386 microprocessors, Linux now supports a variety of computer architectures and its use ranges from personal computers to mainframe computers and embedded systems like mobile phones and personal video recorders.

Initially developed and used mostly by individual enthusiasts, Linux has since gained the support of giants in the field like IBM and Hewlett-Packard, overtaking many proprietary versions of Unix and challenging the dominance of Microsoft Windows in some areas. Many analysts attribute this success to its vendor-independence, its low cost of implementation, and its high configurability and modularity. Its success and technical utility evidence the effectiveness of the open source development model.

Quite independent of the word choices, I'm not sure I can agree with the meaning of what you propose. The word "Linux" is commonly used for three things: the kernel (not the subject of this article), the "operating system" (the subject of the article), and distributions. "Operating system" is something of a nebulous concept, but in the ordinary understanding it would include the kernel plus at least the userland software necessary to create a Unix-like system, and the popular usage would also include graphical infrastructure like X11 and GNOME. On the other, hand, the "OS" does not include software such as OpenOffice, which typically is bundled in a distribution.

Your version does not make clear these three usages, nor does it explain that the point of the article at hand is the "OS" usage (as there are separate pages for the kernel and for distributions). Indeed, your introduction would lead me to believe that the article were mainly about the kernel. Your phrasing also seems to imply that the fourth paragraph is about the kernel, but that interpretation would make that paragraph arguably untrue (the kernel, alone, challenges neither Unix nor Windows).

—Steven G. Johnson 19:43, Jun 13, 2004 (UTC)

First of all, the Linux kernel is an operating system. It satisfies the criteria, which does not include the need for a user interface. It is used without a user interface in embedded systems. Just because it is combined with the GNU and BSD tools to make the core of the system in Linux distributions does mean that it is not an operating system. It schedules processes, it allocates hardware resources, and it interfaces with peripherals. If one insists on including a user interface in the definition of operating system, one can be had without using the GNU tools or any of the other business usually in a Linux distribution. If the kernel and the GNU tools constitute a unique operating system, it is distinct and sometimes called "GNU/Linux". When distributions with different toolsets, ones which change the character of the system, become more popular (as is likely with the open-source development model and with a continued increase in Linux usage), it will be more obviously false to call such systems in their entirety, "Linux".
It is simply false to say that the operating system includes the graphical environment and least of all a desktop environment, especially with regard to Linux as they are userland programs that do not even run while the system is booting up. Even if you disagree with this, the many desktop environments available for X are sufficiently different that it is incorrect and misleading to deal with them as one in a single article. A hypothetical: if a Linux-based system were set up in such a way that one could only do word-processing tasks in OpenOffice, the OpenOffice application would not be part of the operating system. It's more nebulous with Windows because Microsoft does not enforce strict separation of the pieces of their system, and parts of Internet Explorer could be considered part of their operating system in this vein. In this and in the following, we cannot follow popular usage.
This is an encyclopedia of fact, and technical matters are ones that the population is least likely to get accurate. In popular usage, the Windows "operating system" is the only operating system in personal computers ("PC" is on the same level as, and distinct from "Linux") and it is nearly indistinguishable from the hardware. Combined, they are "the computer". The desktop, that is the case with the motherboard, PCI cards, storage drives, etc., is the "CPU". There are many more examples of this, but suffice it to say that popular usage is not a sufficient reason to make a false statement in the article, though dealing with it may be appropriate.
If the article is about this "Linux-kernel-based operating system that includes the GNU tools", then it would be appropriate to have the title be "GNU/Linux" or whatever general term is most appropriate. If not about the kernel or about both the kernel and the entire system, it seems to me that it is more appropriate for this article to be about the entire "genre" or "movement", which is quite nebulous and may not be possible to write about. Why should this article not be about "everything" with links to appropriate subordinate articles.
As for the last paragraph, it is about everything Linux. They have employees whose jobs at the company go from developing the kernel all the way up through distributing software and marketing. So, what should be said there? Just saying "the Linux kernel" isn't quite right and just saying "Linux-based systems" isn't quite right. For this and the third paragraph as well, which I have now changed, it is an accurate statement regardless of one's conception of what "Linux" means. This points out an important point about describing Linux as the entire system: it is tautological. For everything that can be said about the uses, compatibility, features, etc. of the kernel, the same can be said about the "operating system" that is based on it. It is tautological. The superset that some suppose should be the subject of this article is any Linux-based system. Nothing unique can be said about it that cannot be said about the capacity of the kernel itself. Any corporate support for Linux-based systems is at once support for the kernel and vice-versa. Architectures that the kernel supports are at once supported by Linux-based systems. This does not mean that they are the same, though, only a result of the universal flexibility of the term "Linux operating system", because it is not a term that has any definition in itself. - Centrx 02:17, 14 Jun 2004 (UTC)
Sorry, but you are completely at odds with virtually all usages of "operating system" in standard English. For instance, you'd be hard-pressed to find people, including technically knowledgable people, who wouldn't call MacOS X and Windows XP examples of operating systems ("Oh no, MacOS X isn't an operating system...it's just a bunch of tools on top of Mach"). Similarly, even technical people who talk about the "Linux operating system" are not talking about just the kernel in nearly all cases — at the very least, you also include enough userland to make it vaguely Unix-like. As for including graphical environments, I explicitly said that this is not necessary in the case of Unix, but the popular understanding of "OS" does include the GUI when a GUI is available. You, personally, may not like this, but Wikipedia should be descriptive and not prescriptive when it comes to usage.
Regarding embedded systems, they are the exception that proves the rule: if someone talks about the "Linux operating system" in a context where you don't include any userland, you generally have to distinguish this explicitly or to specifically say "Linux kernel", because otherwise people will misunderstand you.
We're talking terminology here, about language and not technical details of a computer's operation. The meaning of a word is determined by usage and not by fiat. "Operating system" in general, and "Linux operating system" in particular, is a case where both the popular and most technical usages coincide in including more than just the kernel. —Steven G. Johnson 03:32, Jun 14, 2004 (UTC)
Both Windows XP and Mac OS X are made by a single company and sold as operating systems. They have always been a single unit. For each, the components of the entire system cannot be interchanged with the components of others systems, for instance as the GNU tools are used on BSD. In the case of MacOS X, while Darwin is available independently and can be interchanged, most of the rest of the system cannot. One cannot make Carbon/Linux and there is only one possible "Mac OS X". The case of Windows XP is especially unique because its components are even less interchangeable than MacOS X. Indeed, the graphical system, web browser, and all manner of things are intertwined with the Windows XP core system. There is no clean break between what is the kernel and what is the userland, like there is with Linux systems. When someone is talking about the "Linux operating system", they are talking about a Linux-based operating system, because there is no "Linux operating system" even if you think that it includes all sorts of software above the kernel. There are several different compilations of this software, but they are substantially different "operating systems" when using such a wide scope of definition for "operating system". This is not the case with Windows XP or Mac OS X. Whether they are properly "operating systems" or not does not make any difference when they are referred to, because there is only one possible thing that someone could be referring to when they refer to the "Windows XP operating system".
It does not matter what the popular conception of what an operating system is. An error does not become truth by reason of multiplied propagation. We don't write articles about, say, centrifugal force or relativity in physics based on what the popular conception of it is. It is appropriate to describe the popular conception of it, but the assertions of the article should not be based on what the popular conception of it is. This is a science, and it should be treated as such.
The reason the GUI is part of the popular conception of what an operating system is, is because of Windows for the reasons mentioned above. Still, though, people understand the DOS is an operating system and they know it doesn't have a GUI. As for the specific case of Linux, the GUI cannot be considered part of the operating system, and least of all the desktop environment atop the graphical system. They are quite clearly applications that are run in the userland. If the "Linux operating system" includes X, then how can the console without the GUI be considered a complete "Linux operating system"? Yet, it is, and one of the main uses of Linux is in server systems with no GUI installed. It is not possible for a subset of a set be identical to that set; this is not the concept of infinity. What about when one is using the Berlin windowing system? or a more common example would be the KDE or GNOME desktop environments. If KDE is a part of the "Linux operating system", does that mean I am not running the Linux operating system because I am using a completely different and simple window manager? If you consider these akin to "configuration options", does that mean that certain distributions are not complete "Linux operating systems" because they don't include the alternate configuration? Why is OpenOffice.org not part of the "operating system"? After all, isn't it part of my operating system if it is the only application I use and, indeed, if I have set up my system for simplicity it may be automatically run when the computer starts and and the system may be restricted so that word processing is the only function I can use the computer for. This expansion of the meaning of "operating system", beyond that of a foundation, is such that it makes the term meaningless. A natural language is dynamic, but once a term has been expanded such that it really has no meaning, the term becomes useless and does not do what language is supposed to do. As such, there is no point in using the term in that way; it is empty.
Nope, it's not accurate if your conception of "Linux" is just the "Linux kernel", and your suggested first paragraph would lead many to believe that this is the sense intended by the article. The Linux kernel, by itself, is not even a competitor to Windows or Unix, much less challenging or overtaking them. You require stuff on top of the kernel to be a comparable product, and it's nonsensical to suggest that one cannot write about Linux-based systems as they are normally used: with a substantial userland. —Steven G. Johnson 04:05, Jun 14, 2004 (UTC)
And Lindows is not a competitor to commercial Unix and Peanut Linux is not a competitor to Windows. So, if an operating system consists also of the userland, then the various distributions are unique operating systems, and it is void to talk about them as a single operating system in a single article. They do look similar to the end user, but so does FreeBSD and that is not the same operating system, so appearance cannot be a deciding factor. Indeed, at the level of "operating system" you're talking about, the high level programs are nearly identical to the versions used on the BSDs.
(PS. I noticed your edit changing "testify to" to "evidence", with a sardonic comment about solemn declarations. I'm surprised that you're not aware of the figurative use of "testify" to apply even to inanimate objects in the sense of "serve as evidence of" — the OED cites continual usage in this sense dating back six hundred years through the present. I don't mind "evidence", but I thought you might enjoy expanding your vocabulary a bit. —Steven G. Johnson 03:29, Jun 14, 2004 (UTC))
Yes, I don't know if the OED definition is incorrectly general or specific but the ambiguity is misleading. Nearly all of the testification in the quotations are from anthropomorphized objects, or are records of a testament of a person. They all convey some will: "thi writyng testifieth", "brickes are aliue...to testifie", "The manuscript volumes...remain to testify", "the fact is testified by my present consciousness", "His...epigrams addressed to this lady...testify". The remaining ones are similar although not quite as clear-cut. (1644) The testification of ruins is like a testament of the anthropormorphized ages. (1794) Shells testify of the long ago ages of a "communication" between "contending" elements of fire and water, and if the shells are the kind that you hear things from then they are clearly "speaking". The only one I cannot quite penetrate without looking at the original text is the 1596 one from Taming of the Shrew, but even there it is speaking of a "note"; writing is penned by someone.
None of these are as mundane and material as low cost and software reliability. The definition states that it means "to serve as evidence of", but this is colored and on an equal level as "to constitute proof or testimony of". The evidence is one of witnesses and personal conviction conveyed, not of scientific data, points of fact, or macroscopic trends.
I hope it isn't a problem that I have moved your responses to the bottom of the "parent" comment so everything makes sense with proper attribution. - Centrx 05:25, 15 Jun 2004 (UTC)

I'd rather remove the GNU part and talk about what Linux is: a kernel. If you want to talk about operating systems, that is either GNU or GNU/Linux. The current state of the article is a shame. There is no excuse to propagate a common misconception instead of explaining it. -- jbc 14:05, 17 Jun 2004 (UTC)

It's a "misconception" according to some supporters of the position. That's called a point of view. POV-pushing isn't suitable.
By the way, don't casually remove others' comments from the page. If you're annoyed they're unsigned, go through the history, find who added them and attribute the comments. Else it's best to leave them the hell alone. - David Gerard 14:30, 17 Jun 2004 (UTC)
Don't just whip out the formless NPOV bat. Calling the common combination of Linux and GNU tools an operating system is no more of a "point of view" than calling the amorphous "Linux operating system" an operating system. Please, then, define what is the Linux operating system, because there is no such thing, and tell us why it is not appropriate to call Linux what it is. I agree that this page should be here as a general overview that points to other information about the kernel, distributions, the GNU/Linux system, etc., but that is no excuse for perverting the truth by defining "Linux" as what is in a distribution. The programs which you include in the "Linux operating system", like the GNU tools and the graphical programs, are common across BSD and other Unix systems, which I hope you don't include in the "Linux operating system". How is it useful to define this combination of systems as one? You haven't defended your position adequately or responded to my points of fact. - Centrx 19:10, 21 Jun 2004 (UTC)
It's POV to say that Linux is only the kernel because this is prescriptive, not descriptive—it does not reflect actual usage. Most references to "Linux", among both the technical and non-technical community, refer to the kernel + a certain amount of userland, usually enough to make at least a POSIX system. Yes, this definition is somewhat amorphous...so is the usage of "Linux" in real life. Real-world usage of language sucks for those who want absolute precision in all things; tough luck. You may not think this usage is useful, but other people, even those who know what's going on technically, apparently like this nebulous usage. (From a personal perspective, I agree with you...it would be much less confusing if we only used "Linux" for the kernel, called kernel + GNU/POSIX userland "GNU/Linux", etcetera...although even this can be ambiguous. But we can't change the English language in Wikipedia.) We have a separate article on the kernel, and we point out that "Linux" is used for many different things, and that's about as much as we can do. —Steven G. Johnson 21:47, Jun 23, 2004 (UTC)

For what it's worth, as a computer user, this is very irritating. Someone coming here looking up Linux is going to be looking for information about the collection of code that runs their computer, excluding the apps. It's fine to point out that there are devastating and wide ranging flame wars about the definition of these things, and point the perverts who are interested in that to those debates, and to mention that almost every commonly held understanding is disputed by a cadre of argumentative geeks who will disagree with any commonly held understanding, but please, this article is torturous and unhelpful. $.02. Intrigue 21:27, 23 Jun 2004 (UTC)

In the article, I can't see much space taken up by flame wars, it seems rather plain. As for the Talk page, that's what it's here for: discussion. What exactly is problematic with the article? - Centrx 21:51, 23 Jun 2004 (UTC)

Usability, market share and moving from Windows

I tried to untangle this section but didn't manage very much. Severe clarification welcomed. I also added a para on running Windows apps on Linux, that being a large part of the "usability" and "moving" question - David Gerard 15:20, 17 Jun 2004 (UTC)

Factual accuracy dispute

I'm going to remove this notice, as Centrx has only pointed out that he doesn't like the common usage of "Linux" to refer to more than just the kernel; he doesn't apparently dispute that this usage is, indeed, common. Language, unlike an equation, can be neither correct nor incorrect, it can only be standard or nonstandard — it is not our place to prescribe one usage as "correct", but only to describe how terms are commonly used. We can point out when certain usages are ambiguous or subject to confusion, but I think the article already does explain that the word "Linux" is used to mean many different things. —Steven G. Johnson 21:47, Jun 23, 2004 (UTC)

The facts are still disputed and with basis, so the tag should remain. The article contains information which I consider incorrect, and with basis that I have presented on this discussion page and which has not yet been refuted. I do not dispute that there should be a statement about common use, the common elision that leaves simply "Linux" in place, but it is an entirely different matter to say that "Linux is the name of a...operating system and its kernel". It simply isn't. There simply doesn't exist a system that you can point to as "Linux" that is unique from the Linux kernel. If the article is to talk about the combination of the kernel with the GNU tools that is common to many Linux distributions, then it must not conflate that "operating system" with the many other Linux-based systems out there. - Centrx 22:53, 23 Jun 2004 (UTC)
Your point seems to be that, if one uses the term "Linux" to mean the kernel plus other stuff, then this usage does not uniquely refer to a specific set of software. I have no disagreement with that. However, there are lots of terms in the English language, even in English for technical fields, that have similarly variable usages. We can't simply ignore this usage, especially when it is the most common one, nor does variable imply meaningless (language, and apparently our brains, functions by loose associations of ideas). The concept of "operating system" is itself extremely nebulous in where the boundaries lie between OS and applications. Your solution is apparently to state, prescriptively, that "Linux" is the kernel, period, and this is unacceptable. All we can do is to outline the general boundaries of the "kernel+stuff" usages of "Linux", giving examples and making exceptions where necessary, and to comment on the common issues surrounding most Linux-based operating systems. I don't believe the article is perfect, but you haven't given any constructive suggestions. —Steven G. Johnson 23:38, Jun 23, 2004 (UTC)
Please read my original edit or the draft I posted above. I have done much more than merely assert that Linux refers only to the kernel. The article doesn't explain much of anything about the various usage of the term "Linux", it has several different descriptions of examples of various uses but the version I posted seems to be much more descriptive and avoids nebulous definitions. I've given several constructive suggestions. Aside from significantly elaborating this concept, I wrote a draft that was very much constructive and constructed upon the current article, and requested suggestions yet I only received some incidental ones from Mr. Gerard, in fact the only other constructive suggestions presented here according to a definition that excludes the extensive discussion here. - Centrx 23:18, 24 Jun 2004 (UTC)
I would like to see some examples of this supposedly "common" usage. If there is not a definable thing to which someone is referring, you cannot say that it is a novel usage that is an appropriate part of the language. These same people would refer to FreeBSD as "Linux" simply because under certain configurations it will look like the same thing. Indeed, in most user aspects it is the same thing. In other words, when someone who is not knowledgeable in these matters refers to "Linux", that which they are referring to is the same thing as the user aspects of FreeBSD, for instance. If they were presented with a similarly configured FreeBSD system, they would know it to be the same Linux to which they were referring before. They are referring to the interface of the system and their definition is not even specific as to what kernel and other basic tools are there. Their definition of the word "Linux" does not include a Linux-specific kernel and it may even refer to distribution-specific tools. The items that make Linux systems unique in the Unix-like world are not part of their "Linux" definition. At its loosest, the term is even used to mean "OS other than Windows or Mac" or "Unix-like operating system". So, if the current definition remains, it must be fleshed out with information that is more general about Unix-like systems, because that is what Linux commonly refers to, but there already is an article for that information, Unix-like, in its proper place. In other cases, it is a matter of elision, that is, someone is referring to their Linux distribution as "Linux" eliding information about the distribution, or about it being "GNU/Linux" or about it being based on the Linux kernel. That does not mean that they are referring to a unique thing that warrants definition in an encyclopedia. - Centrx 22:53, 23 Jun 2004 (UTC)
In the vein of your argument, it is as if to say that the term "Windows" means all of the many and distinct Windows operating systems together as one, despite the substantial differences between Windows 3.11, Windows 9x, and Windows NT. When someone says "I use Windows", they either do not know what it is they are referring to, or they are eliding further information, just leaving out the "XP" for instance. As to the former, the person may know the parameters of what they are referring to, but they don't know that the name that fits those parameters is, say, "Windows 2000". But that does mean that "Windows" is an operating system that is "Windows 2000", and if it did it would paradoxically mean the same for the others, that "Windows" is an operating system that has several self-conflicting, internally inconsistent properties. Of course, it is useful to have a "Linux" article to point people in the specific direction where they want to go, and to have information that is very general (although this is much more useful when you have a single product line made by a single company with similar features, like Windows), but that is entirely different from saying that "Linux" is a whole operating system, one with internally inconsistent properties. This is scientific language, not fair natural language. We do not define things to be something they are not, and cannot be, simply because of ignorant usage. We may acknowledge the usage, but we do not plainly state falsities as truths. This one, in particular, is so anomalous as to be a useless definition, and such words leave common usage quickly in the scheme of things. Anyway, I would like to see this common usage, why it is not due to elision or wholescale ignorance, and some evidence that it is from a reliable source. Re: reliable source: that is, reductio ad absurdem, if the word were uttered by animal, say a repeated mangling by a parrot, or by a mentally slow person, it would not be evidence of common usage, so there exists some line. In the world of dictionaries and encyclopedias, that line is publishing. Novel usages of normal words are not accepted into the language except after a critical amount of usage in publishing, they meet certain criteria that qualify it as a permanent use by those in the know. To bring in a more apt reference, no original research: where is the evidence that Linux is what the article currently says it is? - Centrx 22:53, 23 Jun 2004 (UTC)
"I would like to see some examples of this supposedly "common" usage." If you really don't believe it's the common usage, count press mentions of the topic that say "Linux" to mean the operating system and ones that say "GNU/Linux" or make the point of stressing Linux is only the kernel. You can even keep separate categories for mainstream press and technical press. You've made a truly remarkable assertion, that "Linux" for the operating system is not common usage - now back it up, please. With non-partisan references.
"This is scientific language, not fair natural language." Well, no, it's not, despite your blank assertion. It's what a thing is called by people. It's also what is useful in the context of an encyclopedia: when people come looking for information on the operating system known as Linux, what name will they most commonly look for? - David Gerard 23:22, 23 Jun 2004 (UTC)
It's really not possible to prove such a negative. What you've asked is impossible, whereas one need only find a few credible references to prove the positive. However, the first page of Google results and the first page of Google News results (both for "Linux") all have references that refer to Linux in the way that I have outlined. My assertion is not that remarkable, I am only saying that you have misunderstood what the common usage is actually referring to. Whether they be elisions in the vein I demonstrated above, or a naming that would apply to several things that we would all agree are not Linux, these common usages do not say that Linux is the operating system that the authors of this article have made it out to be. - Centrx 23:51, 23 Jun 2004 (UTC)
Bull. The first article on Google News that comes up for a search on "Linux" refers nowhere to the kernel, and says "ISVs and system integrators will receive help porting, optimizing and certifying their products on Linux" — a statement that implies "Linux" including a userland (the kernel is irrelevant for most porting efforts). In fact, none of the articles that appear on the first page of this news search mention "kernel" anywhere. Almost all of them clearly refer to "kernel+userland" understandings of Linux (talking about building, porting, and running code in an Linux environment, for example). There are a few that aren't clear (for example, a Wired article on the SCO stuff, which technically targets the kernel, but the article doesn't mention the kernel and talks about Linux challenging Microsoft, etc., which most likely points to an understanding of "Linux" to be more than the kernel...just as SCO has been accused of deliberately exploiting the kernel vs. "OS" confusion in "Linux" usage.) I couldn't find any that clearly refer to just the kernel. —Steven G. Johnson
A source need not explicitly state that it refers to the kernel in order for the statement "Linux is the name of an operating system" to be disqualified or for the statement "The name refers to a host of software compilations with operating software that is based on the Linux kernel" to be qualified. Personally, I think that latter statement, which was a part of my edit that started this dispute, is a rather tame modification and has the added advantage of being more specific and accessible without needing to know the term "operating system" and without any controversy.
My contention is that there is no operating system "Linux" as described in the article. Such a thing does not exist; there is no single combination of software that could be accurately labelled "Linux" yet not other combinations that, while based on the Linux kernel, are substantially different such that its properties are in conflict with the former, opposite combination of software. It is appropriate to have an article on such a general subject in the spirit of a fulsome disambiguation article that also includes topics which fit under this general subject yet do not or do not yet fit in separate article. This is not equivalent to the current bare statement that "Linux is the name of an operating system".
I cannot find an article that refers to "Linux" as it is defined in the article. When, for instance, articles about Oracle's and Red Hat's training people on "Linux" in their "Singapore Linux Center", they are not talking about training them on an operating system "Linux". They are going to be trained with certain distributions, certain combinations of software, and much of the knowledge they gain will be specific to those certain combinations of software, and irrelevant to other combinations that are qualified to be "the Linux operating system" that is so indefinitely described in the article. When Microsoft is talking about "challenging Linux on supercomputers" they are talking about competing against distributions that are for the s390 and other mainframe architectures, and competing against companies that are supporting and evangelizing such mainframe implementation, like HP and IBM. When articles are talking about Samsung's new "Linux OS-based" flash filesystem solution, they are talking about a module of, or integration with, the Linux kernel. Note also that the word "operating system" is sometimes used in place of "kernel", which is to say that a kernel is an operating system. This is not a usage where there is a "Linux operating system" as explained in the article, rather that the "Linux operating system" is just the kernel. So, it would not follow to say that Linux is the name of two different things, a kernel and separately an operating system. - Centrx 23:18, 24 Jun 2004 (UTC)
If we are to delineate names as they are commonly used, then we should change the article to reflect that it is often used as a name for all manner of Unix-like systems and even all manner of alternative operating systems. This is not a matter of removing useful information, because the article will still explain that the general term "Linux" is a common descriptor for various Linux-based systems. I agree that the article should remain in "Linux" as the name that people will most commonly search for. That has nothing to do with whether the article should contain false information. Furthermore, it has nothing to do with its inclusion in an encyclopedia. Whatever the usage, if there is nothing concrete that can be said about it, it is nothing more than a dictionary definition, and there is nothing concrete that can be said about such an amorphous and self-contradictory general "operating system" of Linux. It should be more in the vein of a cohesive and full disambiguation page, also with information that is very general to the idea. As it stands, the article contains no information that is not properly the domain of the other articles (GNU/Linux, Linux kernel, Linux distributions, SCO v. IBM), aside from the pronunciation of the name. It is an amalgamation of wholly separate topics, from the compatibility, history, and features of the kernel, to the naming conflict, to litigation which is only directly pertinent to the kernel, to information and code statistics about distributions. The reason this is so is because there is nothing concrete and whole that can be said about the amorphous "Linux", all of the information in the article is directly pertinent to something more specific than "Linux". In other words, each part of the article is an explanation of the particular specific, actual topic that one refers to in a particular use of the word "Linux". - Centrx 23:51, 23 Jun 2004 (UTC)
In the same vein, we shouldn't have an article on the United States of America, since there is nothing to be said, besides the dictionary definition, about such a broad topic that is not properly the domain of other articles (on the history, politics, geography, etc.). No place for a single article that summarizes these different aspects and brings them together in one place for the user to branch out from. —Steven G. Johnson 00:27, Jun 24, 2004 (UTC)
No, I think an article that relates its aspects is appropriate. I was only making this point because it means that there is nothing even in the article that provides information on "the Linux operating system", it is all information on other things. It is easy to define the USA, it is is a certain country: it is 'a land bounded by such and such and its sovereign government, history, culture, etc. since the late eighteenth century'. The article as it stands says something more like, "Linux is anything that contains the Linux kernel". This "operating system" is at once a) the kernel, b) GNU/Linux, c) whole distributions, and d) an embedded system. Why is it useful to have these substantially different things under a single "operating system" umbrella when they can be simply described without confusion? - Centrx 23:18, 24 Jun 2004 (UTC)
If I wrote several thousand rambling words on Talk:United States of America pointing out that "America" is ambiguous and that if the entire article isn't deleted I will insist on an accuracy dispute header, I could not reasonably expect such a demand to be respected. In practical working with them, the difference between Red Hat Linux and Debian GNU/Linux as operating systems is less than that between Windows 98 and Windows XP, though both the latter are indisputably Windows and run most of the same apps - David Gerard 23:44, 24 Jun 2004 (UTC)
I have not advocated any such thing. I have repeated many times that I think that this article is appropriate to have and that I advocate a modification of it.
Comparing Windows 98 and XP is not useful. If we were to compare current Red Hat with 1998 Debian, there are substantial differences. Or if we were to compare even modern general purpose distributions with modern embedded systems. The reason these are "undisputably Windows" is because they are created by the same company with the same name and to appeal to similar markets. That is wholly different than many different companies with different names appealing to wildly different markets. It is not because they are the same system. They are substantially different technically. Without similarities in backing and trademark, the products are quite different, in design and backend function, in administration tools, and in the skin. In fact, the only other similarity between all of the versions of Windows is the user interface, which is not difficult to duplicate on other operating systems.
The differences in practical work with them do not mean much, depending on what you mean by practical work. Similarly configured Linux and FreeBSD systems can have the exact same user interface and one can do all the same practical work in the same way on both systems. Indeed, FreeBSD can run most native-Linux apps, so if that is a qualification for Windows 98 and XP being both of the Windows operating system, then the same is true of FreeBSD and Linux. - Centrx 01:25, 25 Jun 2004 (UTC)
That opinion is plausible, but it's also novel, thus original research - justify with good references. (And before you think of DistroWatch, they only listed FreeBSD due to pressure from advocates.) Linux with WINE can run lots of Win32 but isn't commonly classed as "Windows".
Reread original research - even if you're right, original research still doesn't belong on Wikipedia. It's intrinsically a secondary source; your opinions are odd and novel and take far too many thousands of words to explain to be anything other than original research - David Gerard 07:13, 25 Jun 2004 (UTC)
If you mean the statement that "FreeBSD is the same operating system as Linux", I am not asserting that is true, nor am I suggesting it should be put in the article. That statement was predicated upon your assertion, that the fact that they "run most of the same apps" qualifies two systems to be the same operating system, which I disagree with. In other words, if the qualifications you are using for "operating system" are true, then it follows that two systems which are clearly different systems are the same operating system under your criteria. So, the criteria are not accurate.
I'm not asserting that any "original research" should be included. Read the above draft. There is nothing original about it. If your only arguments do not apply to the edit I made, then the edit should be reinstated. How about you make objections that are specific to the edits I made. From what you and Mr. Johnson have been saying, I cannot see how the edit conflicts with your arguments. - Centrx 08:00, 25 Jun 2004 (UTC)
"If there is not a definable thing to which someone is referring, you cannot say that it is a novel usage that is an appropriate part of the language." So, you're saying that if a term is not used with enough specificity to your liking, then the usage does not exist? People use "Linux" to refer to operating systems based on the Linux kernel, usually with enough userland to make it Unix-like, and often including the GUI infrastructure. Of course, there are choices for what you put on top of the Linux kernel. (There are also choices for what kernel version you use — does this mean that "the Linux kernel" is also meaningless and therefore not an "appropriate part of the language?") But you can't jump from variable application to say that the term means nothing, or everything. (No one uses "Linux" to refer to vanilla ice-cream, nor do they use it for "all manner of Unix-like systems" not based on the Linux kernel. You can argue that this is not fair, that the kernel is not the most important contributor to what the user sees, and I might even agree with you...but that doesn't change the fact of present usage.) —Steven G. Johnson 23:58, Jun 23, 2004 (UTC)

Further for Centrx: Wikipedia:NPOV_tutorial may be instructive, particularly the section on word ownership - David Gerard 13:26, 24 Jun 2004 (UTC)

Side Comments

Hi, very good article! I esp like:

Litigation
Main article: SCO v. IBM

How you take segments out of other articles to briefly summerize the major points. I hope this style catches on for other articles, it is quite good. --ShaunMacPherson 01:59, 25 Jun 2004 (UTC)

That's a pretty standard format around Wikipedia when there's a long, long side article - David Gerard 07:01, 25 Jun 2004 (UTC)

Modification of introduction

Objections to and suggestions on any of the following?:

In the field of computing, Linux is a multi-purpose computer operating system kernel that is the basis of a host of similarly named software compilations. It is the most famous example of free software and of open-source development.

Strictly, the name Linux refers only to the Linux kernel, but it is often used to describe entire Unix-like operating systems (also known as GNU/Linux) that are based on the Linux kernel and libraries and tools from the GNU Project. Compilations of software that are based on these components, called Linux distributions, typically bundle large quantities of software, such as software development tools, databases, web servers like Apache, desktop environments like GNOME and KDE, and office suites like OpenOffice.org.

The kernel was originally developed for the 386 microprocessors but now supports a variety of computer architectures. It is used in applications ranging from personal computers to supercomputers and embedded systems like mobile phones and personal video recorders.

Initially developed and used mostly by individual enthusiasts in the early 1990s, Linux has since gained the support of industry heavyweights like IBM and Hewlett-Packard, overtaking many proprietary versions of Unix, and challenging the dominance of Microsoft Windows in some areas. Many analysts attribute this success to its vendor independence, low cost of implementation, security, and reliability.

- User:Centrx

"the field of" is redundant text and I will shoot it on sight. Still seems in severe need of tightening. The first sentence still attempts to claim against the common usage. This text appears to have no actual advantages over the present intro, being a needlessly wordier version with problematic implicit claims. What were the advantages again? - David Gerard 07:01, 15 Jul 2004 (UTC)
It is appropriate and informative to put the subject in context in this way (In...computing). It is not redundant, as there is nothing in the introduction that states that Linux is something in this field, that it is more than just "in a computer" as an "operating system kernel" but rather its code is educative and useful in other elements of the field, and its development model is widespread in the field.
Please inform me how it is in need of tightening, especially severely and in light of the fact that the present text requires greater tightening. It is closed-minded to say that the modified text has no actual advantages over the present text, when there are numerous minor changes, unrelated to the major objection, that greatly improve the text without degrading readability. In fact, I will make those changes now. The second paragraph of the present text, for instance, is a single, unnecessarily long sentence. How is breaking that up into something more readable not an advantage?
Nevertheless, somewhat wordy text is preferable to inaccurate text. It is false to say that "Linux" as a total operating system including the GNU tools is the common usage. It may be common, but it is not more prevalent than the use of the word to mean the kernel. The first sentence makes no statements that conflict with what you consider the common usage. It may not explicitly state that usage in the first sentence (but, after all, that usage is most commonly implicit), but what it does state there is in accord with that usage, and the text of the second paragraph explicitly states that usage.
The major advantage of the updated text is that it does not state that Linux is an operating system. You may think it is the name of a group of similar operating systems, but it is not one operating system. The following three phrases are false: "...is the name of an operating system..." and "...used to describe the entire Unix-like operating system..." and "...the name is commonly used...for whole Linux distributions". Do you think, for instance, that the article on Red Hat Linux should describe it as "the Linux operating system" or even "a version of the Linux operating system" (in light of the use of the word version in the computer world, not a general English use). After all, it is a common usage, and it is quite in line with the present text of the article. Also, it is false to say that Linux, in its use as the name of the kernel, is only the name of the kernel when used as part of the larger "operating system". This is blatantly false, as the kernel is often used in software that far from that "operating system". I am going to have to put the factual dispute header back if the article doesn't undergo some change. - Centrx 21:08, 15 Jul 2004 (UTC)
Also, the present text uses several terms that a naive reader will not know as though they are obvious, like "the GNU libraries and tools" and "Linux distributions". This is an introduction, just because the definition of these terms in the modified text makes it longer or add more clauses, does not mean it is not appropriate. - Centrx 21:19, 15 Jul 2004 (UTC)