Talk:Linux/Archive 16

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 10 ← Archive 14 Archive 15 Archive 16 Archive 17 Archive 18 → Archive 20

New Picture

Hi. As you can see I'm in the process of adding stuff from the French article. I think we need a new picture that explains the makeup of a typical installed distribution (I just realized what I consider a difference between Linux and Linux distribution, btw... Linux distribution needs to describe not only the end result, but also the groups that distribute and how they are organized, including technical information (e.g. summary of the Debian release process, which does not really belong on this page)... but that's off-topic for this section, we can discuss it elsewhere.)

I think this one from the French wikipedia is just OK, but it isn't detailed enough and it could be a lot prettier too (no offense to the original author/artist). These are the components I'm thinking about for the picture:

  • Linux kernel
  • Open source drivers
  • Closed source drivers
  • libraries
  • shell (for console access)
  • All five of the following overlap shell and window manager:
    • free and open source applications
    • non-free applications, which are distributed binary-only or non-free source license
    • development tools
    • distribution-specific installers + package management
    • manuals, documentation
  • X server
  • window manager
  • support (phone, IRC, forums, mailing lists)

I can produce such a picture myself and it will look pretty nice, but it takes a while to do (I use Inkscape a lot for technical drawings and I'm familiar with this kind of thing; the .svg's it produces are great for wikipedia too). However, if you have "the skills of an artist" (Trogdor reference), feel free to pitch in! (Not that I'm a real artist, although I do have a Graphire4 tablet......). I'd really like some feedback on these ideas, if anything is missing, and discussions as to how it can all be organized. Chris Pickett 16:23, 13 February 2007 (UTC)

For one, the kernel is blown way out of proportion in the picture.  :)
Personally I think "telephone support" is blown way out of proportion :)
  • Kernel modules are part of the kernel thus no distinction needs to be made.
  • Development files are not normally part of a default installation -- they're of course part of the distribution, but if you want to mention them, you'll first have to mention the whole bunch of development tools, such as the compiler, binutils, autotools, etc. Not to mention that a variety of languages are used for development, some of which don't have a concept of header files at all.
  • Generally "binary" means that software is distributed in a compiled form. Debian does this with free software programs all the time.
80.233.255.7 17:38, 13 February 2007 (UTC)
Thanks! I took some of your comments into account and adjusted my list/figure outline
I renamed kernel modules above to device drivers. I do think it's important to bring up that there are closed source blobs, considering that many users will install them (NVidia, ATI).
I understand that binary means compiled form, but I've heard it used and used it myself frequently to imply binary-only, but it's usually just laziness... mea culpa. Furthermore, you can have so-called "open source" software that is proprietary, e.g. Inform. So, you're right, it's important to avoid that confusion.
Is it better now? Worse? Chris Pickett 18:15, 13 February 2007 (UTC)
Better. The usage of "open-source" and "free" might come across as confusing or inconsistent. Not sure how to fix that, though. 80.233.255.7 19:19, 13 February 2007 (UTC)
Oh, right, because it could be "free as in beer". In the article text I've been writing "free software and open source <foo>". I don't really like FOSS or FLOSS for some reason I haven't identified yet, but maybe those are good terms. Anyway, that kind of thing is easily changed. Chris Pickett 20:05, 13 February 2007 (UTC)

Changes I'd make:

  • Manuals are part of software - maybe the diagram meant to say "printed manuals"?
  • "commercial" is meaningless and confusing - free software can be commercial, and proprietary software can be non-commercial, so "commercial" says nothing
  • If "commercial" software is supposed to be "proprietary software", it should be noted in some way that not all distros have this block of stuff
  • Kernel is out of proportion
  • If components are to be mentioned, GNU should be mentioned, maybe: "Free software - GNU system components and third-party system software"?

Gronky 02:39, 14 February 2007 (UTC)

Gaming

The article does not make it clear that Xbox DOES NOT run Linux. I am aware of this link, but it was a project to run Linux on the Xbox, and only means that the Xbox CAN run Linux, not that the Xbox DOES run Linux. I have an Xbox. It does not run Linux. It CAN run Linux, but it does not. I think that the article implies that the Xbox runs Linux by default. This is not true, the Xbox is made by Microsoft.

So the argument. Should this be referenced in the article? I removed it, but it was reverted, so I'm bringing it here. What do you guys think? At very least, it should be reworded. - Thekittenofterra 01:39, 15 February 2007 (UTC)

If Linux can be installed on a piece of hardware, that hardware can and does run Linux. It is extremely uncommon to buy a computer from any manufacturer that runs Linux by default. Nevertheless, Linux can run on more hardware than any other operating system, save perhaps NetBSD. (Not sure about this, but apparently NetBSD runs on a toaster.) I do think however that you have raised a more important point, namely that Linux is usually not the default OS when hardware is purchased. In fact, in writing this response to you, I realized it is the subject of the next subsection I plan to add to the Distribution section, when I find the time to translate it from the French Linux page. This will probably be today or tomorrow. Sound good? Chris Pickett 01:50, 15 February 2007 (UTC)
You raised a whole different subject. Most users do not even realize that consoles have OSes. Consoles are designed to be locked down systems, and thus, developers do not develop for Linux on the Xbox, which makes it very confusing to even mention the Xbox as something developers develop Linux games for. Maybe its odd because the other consoles run Linux for the most part, but the Xbox uses Direct X, not Open GL to run games, meaning, it runs a modified Windows kernel.
Now, on the part that you brought up, where OEMs don't install Linux, that is a problem for Linux. As a fan of the Open PC architecture, it does me great pain to see that Linux is vastly unsupported, leaving your OS sometimes locked down, which goes against the Open PC architecture. This is something I have been looking to talk about for a long time, and any further talk on the subject of OEM Linux, or Linux distros for the general public, please contact me on my Talk page. This means you Chris. - Thekittenofterra 02:02, 15 February 2007 (UTC)
I'd prefer to have discussions here, so that everybody can participate. I think if you feel the article needs to discuss this, you could propose some new paragraphs here. Or you could wait, because as I said, I am going to translate the OEM section from the French page shortly. The problems with confusion surrounding Linux running on consoles are not just limited to consoles; this is true for several non-OEM embedded Linux systems, for example iPodLinux. (Note that the Embedded Linux article is rather confusing; any Linux running on an embedded device is embedded Linux.) I think the appropriate place to discuss it is earlier in the article, and then just a quick mention again in the Gaming section. Chris Pickett 02:21, 15 February 2007 (UTC)
You are correct, there is a difference between "does" and "can". As an example: he can walk 10 miles but he does not walk 10 miles. Mike92591 22:46, 16 February 2007 (UTC)
I think it's clearer to say "does not run Linux by default" instead of relying on does/can distinctions. —The preceding unsigned comment was added by Chris Pickett (talk • contribs) 23:00, 16 February 2007 (UTC).
I think someone likes Linux more then the straight facts. There is no way in hell that Xbox uses Linux to help developers develop, as the article implies. Microsoft would never hear the end of it. It uses Direct X, under an NT kernel, thus the name Xbox, for Direct X box. Does and can are two diffrent things, and I think this should be changed or removed as soon as possible, to ensure the integrity of Wikipedia. Perhaps we can settle on what is already in the Xbox article, where there is already a reference to the project. I am removing it now to reflect the truth. - 68.228.56.158 05:04, 9 March 2007 (UTC)

We need to reorganize this

I'd imagine that most would agree that the whole linux category could be more organized. I suggest that it be reorganized so that it's similar to Isaac Newton and it's related articles(for example linux as the primary page with short descriptions of it's more detailed secondary pages). Mike92591 22:05, 16 February 2007 (UTC)

Hi Mike. I'm still translating sections from the French Linux page and merging them in here. I'd really like to complete that work, and I'm managing about one subsection per day. However, Linux is 50+ kb right now, could probably be shorter. How about a new page for each section? Linux history, Linux philosophy, Linux distribution, Linux usage. The problem is, right now, it's not clear to me what's appropriate for the main page and what's not. And I still think there is more that should be discussed here. Basically my approach to writing articles is to write past normal limits, and then trim it down to size by rewording for brevity and keeping the essential stuff. I don't know if that's appropriate for Wikipedia but it seems like it could work. The goal I have in mind is to bring Linux to FA status.
So, basically, here are the questions:
  • What's still missing from the article? I would say the remaining info from the French page.
  • Is the current ordering of sections and subsections logical? It seems best to get that sorted out before splitting out major sub-pages.
Cheers, Chris Pickett 22:59, 16 February 2007 (UTC)

To answer your first question: nothing that I can think of

Okay. The remaining sections on the French page are: "Desktop", with "Traditional GUI", "GNOME and KDE", and "Desktop software" subsections, "Linux as a server", "Other uses", and then finally "Linux security". I think that some writing about media related to Linux would be good: books, movies, news, slashdot, etc.

To answer your second question: the usage section is really oddly organized and history could be a little better. Mike92591 04:16, 17 February 2007 (UTC)

Yes, I'm working on the usage section. What do you think is a good organization for it? I'm aiming for the French organization presently. What do you think could be better about history? Chris Pickett 04:37, 17 February 2007 (UTC)
The growth of Linux and Market share and uptake seem too similar to be separate but perhaps thats just my weird opinion. Mike92591 04:28, 18 February 2007 (UTC)
Yeah, possibly. The market share bit is meant to be talking about how many people are using it, whereas the growth part is listing off major events in the history that contributed to Linux as we know it today. Chris Pickett 04:39, 18 February 2007 (UTC)
I thought about it, and it seems to me that a logical organization for the Usage section would be to split it into Interface and Applications sections. The verb "to use" is notoriously generic (c.f. LucasArts adventure games). Chris Pickett 18:53, 17 February 2007 (UTC)
Sounds good Mike92591 04:28, 18 February 2007 (UTC)

CLI

FTFA: "Linux includes a command line interface (CLI) as part of its Unix-like functionality." with a screenshot of Bash. I'll let other people speculate on how to fix such nonsense. --MarSch 11:28, 17 February 2007 (UTC)

I'm sorry, it was a translation and a good-faith effort. I think it's okay, and I do most of my work in a terminal. I'd appreciate it if you could explain what's wrong with the sentence. Thanks, Chris Pickett 18:28, 17 February 2007 (UTC)
It's one thing to be sloppy with the name Linux in general and use it when you really mean GNU/Linux. It's quite another to basically say that GNU Bash is part of Linux. That rates on the same scale as calling Linux part of GNU.
It is GNU that includes Bash, not Linux. --MarSch 10:00, 19 February 2007 (UTC)
I see what you're saying, but the problem is this: we're referring to the OS as Linux, not GNU/Linux. The GNU components are one part of Linux, and so if something is part of GNU, it is generally considered part of Linux too. I know it is confusing that the kernel and the entire OS get the same name, but that's just how events have played out. In the same way that Bash is part of OS X, Bash is also part of Linux. The section on the CLI doesn't actually mention Bash or GNU, it just has a screenshot of a Bash terminal. It probably should have a Bash wikilink somewhere, but I want to make sure things aren't confusing for laypeople; explaining that Bash is the default shell but that there are also other shells such as dash and sh and tcsh might be too much. Furthermore, I think it is the POSIX sh compatability that is of chief importance at this level, not that the particular implementation/extension of that functionality is usually Bash. (Most shell scripts used in a Linux system start with /bin/sh so as to remain as portable as possible; Debian relies on this so that it can use the faster dash shell to run scripts.) Chris Pickett 18:54, 19 February 2007 (UTC)

A lot of systems include bash. BeOS included bash. --FOo 10:40, 19 February 2007 (UTC)

Most importantly: its about it's command line interface not just bash but, bash is by far the most common. Additionally with bold and capitals letters for emphasis: LINUX IS NOT AN INCORRECT NAME FOR THE OPERATING SYSTEM; LINUX IS NOT THE NAME FOR THE KERNEL. Mike92591 16:16, 19 February 2007 (UTC)

"Linux" is a confusing name. GNU+Linux is clearer. But we're talking about the same OS. The shell in most GNU+Linux distros is Bash, but even when it isn't, the commands used on the CLI are usually from GNU Core Utilities. Anyone can invent hypothetical corner cases, research projects, and toys that use Linux but aren't more GNU-based than Linux-based, but having that as a motive just leads to a misleading or inaccurate article. Gronky 16:30, 19 February 2007 (UTC)

LINUX IS NOT THE NAME FOR THE KERNEL

I guess that must be why Linus Torvalds refers to it as "Linux" in RELNOTES-0.01 and ever since. Oh, wait. 80.233.255.7 16:52, 19 February 2007 (UTC)
You're right, Linux can refer to both the kernel and the operating system. Now it is no wander why people advocate the name "GNU/Linux" and "GNU+Linux". I apologize for my ignorance. Mike92591 18:16, 19 February 2007 (UTC)

What MarSch doesn't seem to accept is that the name "Linux" is generally understood to be a broad term to refer to an operating system based on the Linux kernel and other software (including GNU). The opening paragraph makes this quite clear and it's been this way for 15 years that I can remember. Do a Google and you will get 23 million hits for GNU/Linux vs 290 million hits on Linux. An encyclopedia publishes the FACTS, it doesn't promote people's political agendas. As for bash not being Linux, this has been the shell of choice from the very beginnings of Linux and it is a perfectly reasonable illustration of the command prompt. 87.74.2.97 23:40, 24 February 2007 (UTC)

"LINUX IS NOT THE NAME FOR THE KERNEL" is really nonsens. You can call the kernel "the kernel" or you can call it linux. It does not have other names. If you run "Linux From Scratch", and it should be small, you end up with linux + busybox. The name of the system would rightly be "Linux From Scratch System" (the distro name + "System"). It would be wrong to call it "GNU/Linux" as there is no GNU in it.
The worst thing with the article is that only a few percent is about the Linux kernel, and the rest is about GNU. It is not necessary to have a picture of Linus but that is OK. On the other side should there be no picture of Richard Stallman. RMS did not code anything of Linux. It is disappointing that there is written more about GNU on the Linux page than there is written about GNU on the GNU page.
GNU/Linux is also a confusing name. What is it? It is much more clear if you say Debian or Fedora. And if you tell your mother, you just say Linux because there is no reason to bother her with technical details. Hans --85.81.46.206 03:21, 27 February 2007 (UTC)
Hi Hans. We all understand that Linux is the proper name for the kernel. This article is about the notable software that is included in what most people call a "Linux operating system". For a large number of machines, this includes GNU software, and that is why it is discussed here. I don't like the pictures of Linus and RMS either, but I think other people do. I agree that Busybox should be mentioned properly, and the distinction should be carefully made. I looked at the Busybox README and it indeed provides replacements for many GNU utilities. Do you know if it includes any GNU utilities at all? Does it require GCC for compilation? Can you run an X server on such a system? How do you propose we integrate Busybox into the article text? Thanks, Chris Pickett 03:46, 27 February 2007 (UTC)
Hi Chris. Please don't add Busybox - remove a lot of the other stuff. Busybox is for embedded systems and some very small systems does even not use Busybox. The best thing to do is to delete or move all the GNU-stuff to the GNU article and the Linux-stuff to the Linux Kernel article. Keep what is relevant to describe the single word Linux. If the article is short it will force people to go further to either GNU og LinuxKernel to read more. Richard and Linus are NOT best friends so it looks strange with pictures of both of them. Hans --85.81.46.206 04:09, 27 February 2007 (UTC)
Hi again. If you look through the archives at the top right corner, you'll see that there has been a lot of heated discussion on this page. I don't see a big rearrangement like that being popular---when most people talk about "Linux" they are talking about what is described on this page. However, I really do think it is important to describe Embedded Linux better. The article *is* currently too long, but we can deal with that by moving information to daughter pages. Chris Pickett 04:33, 27 February 2007 (UTC)

Please fact check

We've had a message to OTRS stating that this article has attack information in it. I can't see it myself, but if people would not mind giving the article the once over, that would be good. Thanks, Guy (Help!) 22:39, 24 February 2007 (UTC)

I don't know what OTRS is. Can you ask for clarification and give a link to the complaint? I don't know of any specific attack info in the article and I'm monitoring it. Chris Pickett 23:37, 24 February 2007 (UTC)

working on other enthusiasts?

The following sentences (in the intro) make it sound like Linux kernal developers began working on other enthusiasts after 1991:

"Most development from 1984 to 1991 was done by the GNU project. After 1991, the Linux kernel developers began working on it as well as other enthusiasts."

This can't possibly be what it means??? I am unfamiliar with this subject but I assume "the GNU project" refers to a group of people who were working on developing Linux since 1984, and that "Linux kernal developers" refers to another group of people who began working on Linux after 1991. "Other enthusiasts" is then probably a catch-all phrase for some people who are neither associated with "the GNU project" nor with "Linux kernal developers", but who nevertheless began working on Linux after 1991. Is this correct? Are the "other enthusiasts" important enough to keep in this introduction? If the sentence read "After 1991, other enthusiasts, among them the Linux kernel developers, began working on it as well," would that place too much emphasis on the "other enthusiasts"? 216.165.237.51 00:38, 25 February 2007 (UTC)

I'll fix it, thanks. Chris Pickett 01:22, 25 February 2007 (UTC)
There is nothing particular about 1991 with respect to people other than Linux kernel developers. The X Window System is an example of work done by "other enthusiasts" and its development definitely did not start only after 1991. 80.233.255.7 02:26, 25 February 2007 (UTC)
Well, my opinion is that the lead needs to be rewritten entirely at some point... and before that happens, X needs be accounted for both in the lead and the history sections, and possibly TeX and other major pieces of the GNU system (after all, you can't really claim that "the GNU project was working on X", even though it's an official part of GNU). Did anything come from BSD? Chris Pickett 03:34, 25 February 2007 (UTC)

Market share

Market Share of ~3% is laughable. To get this based on "web browser traffic" is almost as laughable... Dare I say it, but the average Linux system is useless without internet access, and would be the way a large majority are ever able to get a copy. Ie... Linux computers, as a percentage, would be far more likely to have internet access than your "average" Windows system.

Edit it to something more realistic - like 1/30th of 1% ? (ok I made that up, but clearly - so did the person who wrote the article!) —The preceding unsigned comment was added by 149.135.103.46 (talk) 12:18, 25 February 2007 (UTC).

Initially worked on by ???

The second paragraph mischaracterises the initial developers as "individual enthusiasts", and says nothing more about them. Most of the initial developers were working for the GNU project to make a free software OS exist. This seems important and there's no reason to gloss over it, but given recent debates about the three letter word, I thought I should ask here first if there are objections to adding a few words on this. Gronky 19:21, 6 February 2007 (UTC)

Yawn. Be Bold. "Discussing this" (i.e. having the same argument again for the fiftieth time in a year) on the talk page isn't going to improve the article. You haven't even suggested an alternative phrasing to discuss. Chris Cunningham 20:57, 6 February 2007 (UTC)
I'm a few weeks late to this, but I'm removing most of this change. Why? Look at the facts: Linux didn't start in the 1980s -- Linus Torvalds himself will tell you that. GNU != Linux -- you can use the GNU tools on non-Linux systems (Mac OS X and Windows Vista, e.g.). Linux != GNU -- you can use the Linux kernel in embedded environments and build your own userland from scratch, if you like. Also, X Window System and TeX have nothing to do with Linux or GNU... they are cross-platform applications. -/- Warren 08:03, 26 February 2007 (UTC)
I reverted your change. If you'd like to write about non-GNU systems that use the Linux kernel, please tell us about them here, maybe it's important enough to have a section in the article. Last time we had this discussion all that turned up was that one could use an alternative libc. Otherwise, the lead should at least reflect what the history section says. (It doesn't presently mention TeX in the history section. In fact, I added the mention of TeX and X because Chris C. was complaining about too much GNU emphasis and I somewhat agree---those large pieces not directly developed by the FSF should be given their own mention. TeX is needed for documentation, btw.) Chris Pickett 16:42, 26 February 2007 (UTC)
As to your argument about cross-platform applications: almost everything on a Linux machine is a cross-platform application, except the kernel and distribution-specific stuff; this article just explains what common Linux operating systems are like. —The preceding unsigned comment was added by Chris Pickett (talk • contribs) 16:45, 26 February 2007 (UTC).

Non-free Minix

http://en.wikipedia.org/w/index.php?title=Linux&diff=111381733&oldid=111374657

Chris, as I said in the summary, "non-free" suggests that Torvalds was motivated by the need to have a free software kernel. By "free software" I meant free software as defined by the Free Software Foundation, which is different from non-commercial software. You can not use "free software" for one and "free software" for the other and then use "non-free" for the opposite of both.

  1. The original Linux kernel's licence was too restrictive to qualify as a free software licence. It did not allow any and all commercial distribution. With this in mind it is obvious that being free software was not a goal.
  2. In the page pointed to by the link you added, Torvalds only speaks of price, as indicated by "for free".

80.233.255.7 19:10, 27 February 2007 (UTC)

AFAICT the license was a free software license except for the fact that it didn't allow commercial redistribution.[1] Specifically, it had a "give all changes back" clause: Linus was not just creating freeware, and this was also mentioned in the original kernel announcement. [2] My understanding of "freely available" and "freely redistributable" is different from "for free", and its clear that source code availability is important here (nobody is disputing that, right?)
I think the most appropriate way to clear this up is to say that Linus changed his license to GPLv2 when he realized that it was better, and that the significant difference was that it allows commercial redistribution. Is that accurate? It's not mentioned in the article at any rate. Chris Pickett 19:29, 27 February 2007 (UTC)
There, it's done now. Is it misleading in any way? Chris Pickett 20:04, 27 February 2007 (UTC)
I got an edit conflict so I'll reply here as well. My concerns still stand even after viewing the change. Both kernels were non-free, so it's misleading to suggest that the reason for developing Linux was Minix's non-freeness. (It might have not be the intention, but that's how it comes across, anyway.) 80.233.255.7 20:17, 27 February 2007 (UTC)
A licence is not a free software licence if it prohibits commercial distribution. A "give all changes back" clause is also restrictive: it's important that the users be free to use their modifications privately. I am not suggesting that Linux was "freeware", but it definitely was not free software. I recognize that Linus changed the licence later, but at that point it is no longer relevant to the original motivation, and that's what we're really talking about here. With regards to "for free", earlier in the same paragraph you quoted in the summary Linus wrote:

[..] look at who makes money off minix, and who gives linux out for free. Then talk about hobbies. Make minix freely available, and one of my biggest gripes with it will disappear.

80.233.255.7 20:17, 27 February 2007 (UTC)
I don't think we should be guessing at Linus' original intentions nor go around trying to source them. I don't want this article to turn into even more of a pissing contest. I personally think "source code available free-of-charge (or for a nominal fee)" and "share and share alike" are the most important parts of free (specifically copyleft) software, and in my opinion it seems like Linus had nearly everything right except for the commercially redistributable bit. However, I haven't read the original licence and I'm also too busy to play lawyer right now. Nevertheless, I took out the "non-free and commercial" bit regarding Minix. Does it read more fairly? Chris Pickett 20:54, 27 February 2007 (UTC)
I also added that Linux was a non-commercial replacement for Minix, and emphasized that the key difference between his licence and the GPLv2 was commercial redistribution. So if there's anything left that is suggestive or misleading or factually inaccurate please bring it up. Chris Pickett 21:09, 27 February 2007 (UTC)
Much better. 80.233.255.7 21:33, 27 February 2007 (UTC)

History

Does the history section need to include the information on GNU? To me, the way it reads is that GNU/Linux supporters are trying to improve their case by including information on it. While I do agree GNU and Linux are closely related I personally don't think that the information on GNU in the history section belongs there, I do think that it needs to be looked at. Socerhed 13:16, 2 March 2007 (UTC)

  • GNU was a OS, complete except for the missing kernel. Linux was nothing, but a kernel. If you are arguing that the OS we now know as Linux would have become what it now is, had it not found a proper ground (the GNU movement) to grow in.. then any further discussion with you is probably futile. I am a bit fed up with anti-GNU zealots: first you bundle GNU and Linux into a single package, and call the resulting OS "Linux", devoting an article to it. Well, I might accept that, provided that almost everyone refers to GNU/Linux by the name "Linux". It is reasonable. Now, you propose to move one step beyond that, and actually erase GNU from the history section, effectively deleting a fundamental part of the early history of Linux. What next? Identify the whole OS with the kernel, and say that all we have today was written by Linus Torvalds? It is the pro-GNU zealots who say that Linux is just a kernel, and you appear to agree with them. If you say that Linux doesn't come from what once was just GNU, then you are actually accepting that Linux is just the kernel, and hence you should oppose to this article. — Isilanes 15:19, 2 March 2007 (UTC)
All this nonsense stems from the fact that this page is still named Linux and not GNU/Linux. There is already a page on Linux at Linux kernel. Can we please rename this page? --MarSch 16:12, 2 March 2007 (UTC)
No. The operating system has no official title as there is no authority in place to give it one. It thus goes by a variety of names, of which "Linux" is by far the most common. A very, very vocal minority has drunk the kool-aid and is willing to endlessly parrot the position that collaborative software is named in some procedural manner (where the procedure is determined by the Free Software Foundation), but this is a minority position indeed and warrants little more than a footnote in the OS article. The counter-argument is that "all this nonsense" stems from the shameful manner by which the FSF has attempted to tack an Obnoxious Advertising Clause onto GNU, Chris Cunningham 16:35, 2 March 2007 (UTC)
The argument isn't about naming or trying to say that GNU isn't important to Linux, its whether or not GNU needs to be included in the history in the context that it currently is. The way that the section currently reads leads the reader to think that Linux came about because of GNU. The Linux kernel wasn't made for GNU. The Linux idea had nothing to do with GNU. GNU had nothing to do with the Linux kernel. They are two separate things. Go look at Apple, UNIX is mentioned 3 times. This article is not about GNU. Want to put a link? Sure, but I mean really this is crazy, Linux is Linux and GNU is GNU. People act like GNU was THE ONLY source that Linux had to pull from. HELLO. It was the 90's, there were programmers coming out of God's ass and they were just as capable of making programs. Besides its hard enough to explain Linux anyways, I don't even want to try to explain GNU. Socerhed 04:57, 3 March 2007 (UTC)
  • "The way that the section currently reads leads the reader to think that Linux came about because of GNU. The Linux kernel wasn't made for GNU.". This is the article about an OS referred to as "Linux". For comments about Linux (kernel), please go to ... yes, you guessed it: Linux (kernel). The creation of the Linux kernel has little to do with the FSF or the GNU movement, yes. They (GNU OS and Linux kernel) are two different things. However, the Linux OS comes from the inclusion of the Linux kernel into the GNU framework, when this synergy generated a complete OS. Either you accept that, or you support dropping "Linux" as a correct name for the OS this article refers to, which is the OS steming from that cooperation. — Isilanes 13:23, 3 March 2007 (UTC)

I think that in addition to GNU, there are other components that need more thorough treatment and integration to provide a balanced picture of the history. At the top of the list we have the X Window System, which currently gets half a sentence: "The upper level could be supplied by the X Window System, ..." I would also suggest finding out and then discussing the role that BSD played. Chris Pickett 16:59, 2 March 2007 (UTC)

I second this. Components such as X and TeX should be more represented. Although this needs to be done a bit carefully. Currently, representation of X has been added by giving it equal rank with GNU, which ignores the size and intent of X's contribution. By "intent", I mean that like TeX and Linux kernel, the (two) X developers were not trying to make an OS exist. X just accidentally later became useful in the creation of an OS. But that's getting into details. The more important point is that there is general disproportion in the article, and fixing that includes adding more detailed discussion of components such as X and TeX. Gronky 00:39, 3 March 2007 (UTC)
What about if we try and separate technical contributions and philosophical/organizational/licence/legal ones? This seems like a pretty clear dividing line. Then again, it might be difficult to do properly. Also, the second category probably pertains more to History of FOSS, and we have a huge Philosophy section already. Chris Pickett 01:11, 3 March 2007 (UTC)
From a technical contribution angle, as well as contributing N lines of source code, GNU also put work into design of the OS. Also, separating the other angles from the whole might indeed be difficult. Imagine trying to describe Windows Vista without mentioning Microsoft - think of the difficulty and of the quality of article/intro it would produce. The situation is not completely analogous, but both Windows Vista and GNU+Linux did have organisations backing them, both did have a legal support structures, both did have licences drafted and enforced, and both did have philosophies, and neither would be here today without them. ...but we're getting to far into article theory again. Gronky 04:36, 3 March 2007 (UTC)
All I meant was to discuss them more or less independently of each other, but in the same section, say using "Technical" and "Social" subsections. Maybe it really is a bad idea. (btw, I think LOC is a terrible metric; good code has had lots ripped out of it.) Chris Pickett 05:22, 3 March 2007 (UTC)
LOC is indeed a bad metric. It ranks OpenOffice.org as a bigger contribution than the Linux kernel or Glibc or probably GCC etc. Gronky 12:51, 5 March 2007 (UTC)

The notion that Linux development stemmed from TeX or X Window System is patent, unsupportable bullshit. It stinks of agenda. You won't be able to find quotes from Knuth or Torvalds that support this fantasyland version of reality, because those quotes don't exist. It also belies a fundamental misunderstanding of what Linux is -- it's not just a kernel, and it's not just a desktop operating system. You don't need GNU, TeX -or- any kind of X implementation to have a fully functioning system that could be called Linux. That's easily proven by all the embedded systems out there that employ Linux. This fact isn't news to anybody reading this, but I guess it's easy to stick fingers in ones ears and ignore it in favour of promoting one particular and popular rendition of Linux, that of a GUI-based desktop system. I will revert further attempts to promote this view, unless such statements are effectively backed up with proper Wikipedia:Attribution to reliable sources. That's how we do things on Wikipedia. -/- Warren 20:52, 4 March 2007 (UTC)

Well, I personally would like to see mention of Busybox. The info I found suggested that it started in 1996 as an installer for Debian and grew from there. Would you like to do some research and write a bit about non-GNU embedded Linux operating systems for us? I would really like to see that added. By the way, I'm not one of the "GNU/Linux" pushers but the fact remains that GNU/TeX/X and other pieces of software are an important parts of Linux desktop and server operating systems. As far as I am aware these came before embedded Linux systems. Chris Pickett 21:08, 4 March 2007 (UTC)
BusyBox? How does BusyBox define this OS? Gronky 12:51, 5 March 2007 (UTC)
The point is that Busybox and uClibc can apparently be used instead of the GNU userland in embedded Linux systems, and the systems (apparently) look about as much like GNU as BSD does. Since (it appears, and seems natural) that there are a LOT of embedded devices sold that use the Linux kernel but really not that much of GNU, where do they get discussed? In this article? Or do we put a disambiguation clause at the top to clarify the split? I've seen this complaint a couple of times now, from Warrens here and from Hans (in archive 15). There is a distro called TinyLinux that uses them, and it looks like a lot of times there isn't really a "distro" per se but in fact the manufacturer of the device rolls the OS from scratch. I would like the following things need to be discussed in the lead and history sections: GNU, X, embedded Linux, and desktop/server Linux (a.k.a. GNU/Linux). Since the lead is so volatile I'll just try to get the rest of the article straightened out first. Chris Pickett 15:19, 5 March 2007 (UTC)
Then again, the lead is less cluttered as it stands. Chris Pickett 15:34, 5 March 2007 (UTC)
I wasn't suggesting that development of the OS stemmed from X or TeX, I'm just saying those packages deserve a mention. Development of the OS stems from the GNU project, but for completeness, it's worth mentioning that GNU didn't write everything, and that the OS now incorporates third-party packages. Incorporating some, such as X, TeX, were part of the original plan. And the incorporation of others, such as Linux kernel, were by unforseen circumstance. Gronky 12:51, 5 March 2007 (UTC)


Citation consistency and style

Some of the citations in the article are cited like Linux Usability Report v1.01. Convert these over using citation templates. -- Ash Lux (talk | contribs) 20:05, 4 March 2007 (UTC)

Do you know of a way to add full cite templates that doesn't clutter up the body text OR require separate Notes and References sections like Finnish Civil War? Chris Pickett 20:52, 4 March 2007 (UTC)
I'm unaware of any way to do that, but if you find a way, let me know. -- Ash Lux (talk | contribs) 06:12, 5 March 2007 (UTC)

Embedded Linux

I agree with Warrens that this article is biased towards Linux desktop and server operating systems. I found a reference to embedded Linux operating systems and distributions that seems to be pretty comprehensive. [3] I think perhaps the reason we are ignorant of these operating systems is that we are used to using Linux desktops and servers. Some of the operating systems are based on desktop/server distributions, others are open source and rolled from scratch, and yet others appear not to be open source in any significant way. Since there is overlap in some cases, it's even more confusing. The main point is that these devices and the OS they run are significant and worth covering. Unfortunately I don't know enough about them to do it properly. Perhaps an early embedded Linux subsection would suffice, with a clarification that the rest of this article talks about the OS as installable on desktops and servers? Chris Pickett 18:53, 5 March 2007 (UTC)

Embedded Linux has its own article. Server Linux doesn't. I appreciate your attempt to increase the scope of this article, but it can't be all-encompassing. Chris Cunningham 09:01, 6 March 2007 (UTC)

Operating systems "appearing"

I accidentally hit return while writing my edit summary. I was going to say that operating systems don't "appear". People spent a decade writing it. Gronky 00:43, 6 March 2007 (UTC)

Not that it matters but, the use of "appear" was correct. Mike92591 01:40, 6 March 2007 (UTC)
Y-awn. In other words, GNU GNU GNU GNU GNU NOTHING WITHOUT GNU. People couldn't run an operating system with "Linux" in the name until the kernel was released, after the kernel was released it wasn't long before people could run an operating system with "Linux" in the name, so it basically did "appear". You'd be far more constructive nitpicking on GNU than continually finding new and obscure ways to complain about being marginalised in Linux. Chris Cunningham 08:59, 6 March 2007 (UTC)
The lead seems to be both biased and factually inaccurate now. I doubt GNU worked on Linux in 1983, because there was no such thing until Torvalds announced it on August 25, 1991. We can not have original research and choose who to credit and how much, but only document how other sources see the history of this OS. The BBC didn't celebrate ten years of the Linux operating system in 1993, but in 2001. Similarly, Encyclopaedia Britannica article starts with "In 1991 Linus Torvalds of Finland began asking for volunteer programmers...". This article gives much more weight to GNU than most reliable sources, which leads to neutrality concerns. Prolog 12:36, 6 March 2007 (UTC)
Yes, this should be corrected. Some serious pro-GNU bias has crept into the article over the last couple of months to the detriment of the previously-fine intro. Chris Cunningham 13:10, 6 March 2007 (UTC)
In 1991, Linus Torvalds released a barely working kernel that he called Linux. Most of the software that was used in combination with this kernel had already existed before that. It doesn't take much thought to realize that "history of 'Linux' the operating system" includes history of such software, (unless you're a Torvalds fan boy that truly believes that the tarballs distributed at kernel.org contain the source code of an operating system, as stated in their own README) and that this is where GNU plays a significant role. It may not be a well-known fact, but it is neither "biased" nor "factually inaccurate". Saying otherwise obviously shows a lack of competence. 80.233.255.7 14:32, 6 March 2007 (UTC)
Utter rot. Does the article on cars go on about the long and glorious history of the internal combustion engine? This is glory-hunting. GNU would probably have a fairer representation in the article if people didn't keep getting reverted for artificially boosting its prominence. Chris Cunningham 15:59, 6 March 2007 (UTC)
  • You are 99% right. Except it's just the other way around: here the car is the OS (GNU) and the engine is the kernel (Linux), and the car was invented before the engine was. Simply the name "Engine" (Linux) caught on, and the whole car (OS) started to be called "Engine". — Isilanes 17:09, 6 March 2007 (UTC)
  • Actually, the key point is simple: we are now calling "Linux" (as legitimate a name as "GNU/Linux" or "Bananas") an OS that was started by the GNU movement in 1983. This is undeniable. What we now have is the continuous development of that OS, created as a free replacement of Unix, still being Unix-like. Now, we are at a crossroad here: either we drop the name "Linux" (bad move, in my opinion), or we admit that the "Linux OS" started its development in 1983 (regardless of the homonym kernel being introduced in 1991). Saying that Linux "appeared" in 1991 equals admitting that Linux is only a kernel. — Isilanes 15:02, 6 March 2007 (UTC)
There was no operating system called Linux (or GNU/Linux, or Jigglypuff Luxury-Yacht Linux) in 1990. That internal combustion engines existed before automobiles does not mean that automobiles should be stated as having existed in some embryonic form since the invention of the internal combustion engine. Your logic is faulty. Chris Cunningham 15:59, 6 March 2007 (UTC)
Good analogy. Analogously, the internal combustion engine is like X Window System. It was useful for the final product, but it was not developed for the final product, and indeed it was in a "ready" state before the final product wasn even planned. There was no "car" when the internal combustion engine was being developed, but there was such thing as a "car" when the car was being developed (during development it was incomplete, but the plans were there and it was partly there) - actually it was probably called an "automobile" or something more historic and it later became known as a "car". Similarly, during the 80s, the operating system we're talking about existed, although in an incomplete form, and with a different name. Just like cars. The plan and the work to write an OS began in 1983. Gronky 16:40, 6 March 2007 (UTC)
  • Indeed. The analogy is the other way around: here the "automobile" is GNU, and the "engine" is Linux, and the automobile was invented before the engine was. Maybe the engine (kernel) was most important for an automobile (OS) to become what it now is, but the automobile was not "invented" with the engine, but much sooner, when an engineer (Richard Stallman) saw the need for a fast and affordable transportation method (a free Unix-like OS). The engineer invented and started to develop the automobile, but his work stalled until a clever fellow engineer (Torvalds) independently invented an engine, which happened to fit in Stallman's automobile. Thereafter the name "Engine" (Linux) caught on, and we now call automobiles "Engines". To be a purist, one could claim that the name is not correct (maybe "engine-powered automobile" is better, i.e., "GNU/Linux"), but all is fine if we simply agree that the name "Engine" is a convention, and that it is immaterial whether we call the car "automobile" or "Engine" or "Banana". The problem comes when poor-memory people start arguing that what the first engineer invented was not an "Engine", because engines were introduced by the second engineer. This is silly! We had already agreed that the name "Engine" (Linux) was not being used just for the actual engine (kernel), but rather the whole vehicle (OS), which started to be developed by the first engineer! — Isilanes 17:04, 6 March 2007 (UTC)
You're talking about naming, but that's not what this thread is about. The issue is that the article said that the GNU project worked on the operating system from 1983 onward, and someone replaced that with "Linux appeared in the early 90s". Why hide who worked on it? [P.S. and not crediting the authors of the OS gets even sillier when Novell, Sun, IBM, and HP get credited for supporting the OS] Gronky 17:41, 6 March 2007 (UTC)
  • Hey, I agree with you. However, the naming is relevant. We must all understand what we are calling "Linux" (a Unix-like OS, that was created as a free replacement of Unix), so we understand what the article must describe (the 1983 GNU project, which evolved to 2007 Linux). The name is irrelevant, as long as we understand what we refer to with it. This thread proves that naming an OS by its kernel is an unfortunate choice, because it provokes the "controversy" of whether the OS was created in 1991 (with the kernel) or in 1983 (when it actually was). Anyway, I find not objection to the name choice, if we all agree what we are referring to with it, which is not clear for some, it seems. — Isilanes 17:54, 6 March 2007 (UTC)
You're right that naming is relevent, but I've noticed that when people discuss what it is and what it's called at the same time here, nonsense ensues. For example:
Assertion: The OS we are talking about is more GNU than it is Linux kernel
Response: But everyone calls it "Linux"!
or
Assertion: The GNU project's contribution to making this OS began in 1983
Response: But there was nothing called "Linux" before 1991, so nothing before that year could be a contributing factor to the OS
And this goes on and on. So when I see someone talking about what it is, and someone "rebutting" the argument with an off-topic point about what it is called (or vice versa), I ask people to separate the issues. When talking about what it is, it's clearest to refer to the OS as as "the OS that we are talking about", or "the OS that this article is about". That avoids the naming issue. Unfortunately, we don't all agree on what we are talking about. There are still many who think that Linus started writing an OS from the ground up, kernel first, and then "Oh, what good fortune" all of the other "tools and libraries" were just lying around, and Linus and friends then made an OS based on Linux (plus the other, less important 99% of the software they "found"). People do believe that, and magazines love to publish it "The OS that a finnish college student wrote that is now taking on Microsoft!" - it's not true, but it's a great story. Gronky 22:53, 6 March 2007 (UTC)
No, Linus didn't start building his kernel to be a GNU kernel. I agree, saying "the OS that we are talking about"-like things would be better. Get it right, the less important part is 92% and was something less than 99%. —The preceding unsigned comment was added by Mike92591 (talk • contribs) 00:45, 7 March 2007 (UTC).

We can argue about intentions all day. It doesn't matter whether Linus intended Linux as part of GNU, or whether Stallman intended to adopt Linux as a kernel for GNU. Neither one of these changes the facts:

  • The Linux kernel was not created by Project GNU. Unlike GNU software, the copyright to Linux was not assigned to the FSF; individual contributors hold it severally.
  • The Linux kernel, like a lot of other non-GNU software, is licensed under the GNU GPL.
  • The Linux kernel is compiled and used with GNU software. Likewise, the highest-volume distribution mechanism for GNU software, is the set of software distributions which use Linux.
  • Software distributions which use Linux for the kernel bundle software from a large number of other sources, including GNU but also including BSD, Perl, Python, X11, KDE, SDL, Videolan, and many others.

The question of whether Debian, Red Hat, SuSE, etc. are "Linux distributions" or "GNU/Linux distributions", is an abuse of language. It is a case where the "is of identity" makes it seem that we are talking about deep essential realities, when really we are talking about what names people choose to use for things. --FOo 06:15, 7 March 2007 (UTC)

I don't think anyone has disputed those four facts you've listed, and the debate here is not about naming, it's about how to describe the OS we're talking about. See my comment just above your's. Gronky 10:48, 7 March 2007 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Linux/Archive_16&oldid=1050741620"