Talk:Operating system/Archive 3

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Archive 6

No user interface?

The statement from the intro paragraph: "Operating Systems themselves have no user interfaces; the user of an OS is an application, not a person." Is that true? What about DOS? What about Windows? Sure, if the OS has a separate kernel that doesn't provide UI, then the kernel has no user interfaces, but the operating system encompasses more than just the kernel. I'm not an expert, it just sounds suspicious. Fuzzypeg☻ 00:25, 29 August 2007 (UTC)

I hadn't noticed that bizarre edit until now as I don't monitor this article too closely. Thanks for flagging it. I just traced it to this edit [1] by User:Buonoj on 30 April 2007. Oddly, it was that user's ONLY edit ever. I believe the edit is so inaccurate as to constitute possible vandalism and should be deleted.

User interfaces are an integral portion of operating systems and are taught as such in computer science courses. Indeed, the user interface is usually discussed as one of the services provided by the operating system to applications. I am fixing the lead paragraph immediately. --Coolcaesar 18:56, 29 August 2007 (UTC)

I have not got teaching by that way, "OS does need to include the UI", because the application it self can proof the UI for other application, example OS > Xorg > Firefox. We could edit firefox and Xorg on those manners that Firefox is the GUI itself while Xorg acts as UI manager and OS just gives services for Xorg to run. User only starts the computer, what starts the OS what runs Xorg and starts firefox what is shown for user. This does not mean that UI cant be included in OS or even in the kernel (monolith) if wanted, this is teaching what I have got, everything just depends the way how the OS or Kernel OS is build. Golftheman (talk) 08:28, 17 July 2008 (UTC)

The OS of a general purpose computer has I/O (Input Output), but not necessarily what most people would call a "user interface". The autodiagnostic electronics in a car don't talk to the driver (although they could and in some cars might), but they have a port to offload data to a handheld device, and **that** device has a display and buttons for the mechanic (the user). A router may not have a UI, but it has input (ethernet) and may have a port open for telnet (an application that could be used by a person to reconfigure the router) etc. Tavern games I worked on used ROMDOS, and had no UI, but at boot time loaded the game framework, which was an application with UI (touchsceen and coin counter) that loaded games (when user selected a game thru touchscreen and paid coins). UI refers to interfaces intended specifially for people, like buttons, windows, keyboards; I/O generally encompasses any signal, which might be telemetry from a satellite or an infrared proximity detector. Pete St.John 19:28, 30 August 2007 (UTC)

An operating system may provide facilities for constructing user interfaces, such as a graphics display driver or a keyboard driver. However, what we now call a "user interface" (such as the Macintosh or Windows user interface) is a congeries of libraries and application programs which use those OS facilities.

A Web browser (like Internet Explorer) or file manager (such as the Macintosh Finder) is only part of an "operating system" when that word is used as a marketer or product manager would use it, rather than as a computer scientist would use it. These are distinct uses of the word, just as "set" means something different to a collector (as in "Collect the whole set!") from what it means to a mathematician. --FOo 07:40, 24 September 2007 (UTC)

But the web browser or filemanager (mediaplayer) is part of OS if parts of those are integrated to same services what are OS services. Then they are not just by talkin marketing "part of OS", but part of OS by technically and then there is not false information by marketing. Isn't this what Microsoft did on first place, first they bundled the browser with OS and then they integrated the browser to part of OS services, so it was not just bundled with OS? Golftheman (talk) 08:28, 17 July 2008 (UTC)

I agree with the above clarification of the difference between how a computer scientist or scholar would define operating system versus how a marketing person would define operating system. Its pretty clear that marketing people conflated "computer system' and "operating system" a long time ago and as a result caused great confusion about what an operating system is in the general public. Using the marketing definition of OS would allow you to make a reasonable claim that solitaire was part of the Microsoft Operating system. :-) And it clearly is not. It is part of the "computer system".

Since this entry is part of the Computer Science category on the Wikipedia lets make sure we are using the proper computer science definitions. If it were a marketing entry we could use the marketing definitions. Jjk 05:20, 12 October 2007 (UTC)

"Computer system" typically refers to the computer hardware (the computer itself, screen, mouse, etc.), not the software. Also, people here appear to be confusing "Operating system" with "Operating system's kernel", which is the basic process and memory manager that people are speaking of. Although an OS can be composed of just a kernel, nowadays it typically includes various inseparable additional services, for example a gui framework server such as a x server. Rami R 12:41, 12 October 2007 (UTC)

But OS #1 reason is to share hardware services for applications so they can work together. most kernels (microkernels) does not include all needed services to be alone a OS, but few kernels (monolith) does include all services and those are then alone a OS. if X needs Y but Y does not need X, is the Y part of X or X part of Y, if is, then we dont have only a X or Y? Golftheman (talk) 08:28, 17 July 2008 (UTC)

• The sentence in the lead paragraph, [a]n operating system is an interface between user and hardware furthers the confusion discussed above, instead of clarifying it. An operating system may include an interface between application programs and hardware; in some embedded devices the application is built into the OS and there is no user interface at all, e.g. electronic fuel ignition in an automobile engine, it doesn't even have an "on" switch. In some OSs the GUI is built into the OS and the OS cannot boot without it. Those are entremes. In a typical general-purpose computer, the OS provides an interface between applications and hardware, and the applications typically include user interfaces. Pete St.John 17:01, 3 December 2007 (UTC)

"the OS provides an interface between applications and hardware, and the applications typically include user interfaces." This says the all what is needed. And only extra information is, the OS can include the UI itself if wanted, but it is not OS service what is wanted from software so it can be a OS. Golftheman (talk) 08:28, 17 July 2008 (UTC)

To answer Fuzzypeg and Rami R's questions, "What about DOS?" -- even in DOS, we can clearly distinguish between the OS kernel and the user interface. The command-line user interface in DOS is usually "COMMAND.COM". (The "cmd.exe" implements more or less the same command-line user interface in a "Windows DOS box"). The pretty graphical user interface that shows the desktop on a Windows box is the "Windows Shell".

I think we all agree that the kernel does *not* include "COMMAND.COM" or the "Windows Shell" or "solitaire".

Now please answer my question: If you think there *is* a difference between an "operating system" and a "operating system's kernel" -- if you think that "Windows Shell" is part of the operating system, even though we all agree that it is not part of the kernel -- then what exactly is the difference? And how do we distinguish between that "operating system" and applications such as "solitaire" ? --68.0.124.33 (talk) 02:29, 28 April 2008 (UTC)

It is question how the system was build, was the solitaire integrated to OS services or even kernel itself? Does removing solitaire affect that you cant run any applications on the computer?

Difference is the services what the OS is. I/O, Memory management, Progress management, Filesystems, System calls etc. If you integrate the solitaire to progress management service, it is part of OS because you cant run any application without progress management. But microkernels does not include all those services in kernel, but separeted and even placed to userland. Monolith kernel runs all services alone in same address space, kernel space. And so called hybrid kernel does run some services in kernel space separeted from kernel and can run some services on the userland, still the kernel is microkernel. Micro kernel alone is not OS, but microkernel + OS services is the OS. And if you integrate applications to OS services, so application can not be removed (does not mean that application does not have uninstall feature!) without compromising the OS services, it is part of OS. We cant draw one rule what would work for all systems, what application is part of OS and what is not. It is question about how the whole system was build and we need blueprints of the OS itself until we can say by sure. Golftheman (talk) 08:28, 17 July 2008 (UTC)

Your hands and eyes can not directly touch or alter the operating system short of touching your finger on a boot floppy's magnetic film. The OS manages the hardware (attached devices: monitor, keyboard, mouse, floppy) and software programs (Windows Explorer, Solitaire) you use via device drivers and APIs. To interact with the OS you need a shell (such as the "DOS PROMPT" C:> command line interface), or other program (software application) such as a utility scandisk, fsck, dd, fdisk, or application such as a click of a mouse pointer on the "Start Menu" to open Excel, or Internet Explorer. All of these rest above the kernel with its support libraries, and interact by means of application programmer interfaces (APIs), pipes ">,|,<", and sockets such as TCP, or UDP, clipboard, and memory snapshots, configuration and storage files. The distinction for modern systems is that the kernel is "protected" from end-users, unless you happen to click on your favorite booby-trapped website link, or email trojan designed to escalate privileges. :bwildasi Fri May 16 22:02:42 UTC 2008 —Preceding unsigned comment added by Bwildasi (talk • contribs) 01:03, 17 May 2008 (UTC)

I agree the operating system may have no uesr interface. Best examples would be embedded microcontrollers, like the above mentioned injection/ignition control units, which work completely non-interactive. It would be easy to make one too, out of some of the ordinary OS-es like DOS - if you removed COMMAND.COM and replaced it by a program that, say, computes Pi number, which does not interact and just sends results over a network - here it is, an OS without a GUI. :arny (talk) 17:59, 7 June 2008 (UTC)

Unix vs POSIX

I reverted a recent addition which attempted to clarify the difference between "Unix" and "POSIX"; but was just too long, rambly, and error-ridden (e.g. "Postix", you have to correctly spell the key word of the section). I'll drop a note on that editor's page to suggest composing a tighter contribution. Pete St.John 16:26, 4 October 2007 (UTC)

Anonymous user needs help with homework?

what are some of the utilities in a operating system . please —Preceding unsigned comment added by 216.255.113.5 (talk) 13:45, 10 October 2007 (UTC)

Customising the desktop?

Me, the anonymous person, saw the image on the article showing a curstomised OS showing the computer stats, like: RAM: 250/3000 Can anybody explain how to make that happen? I really need that. —Preceding unsigned comment added by 83.183.127.178 (talk) 20:33, 5 December 2007 (UTC)

Multics

There is no reference to multics except in external links! Multics IS the major development in operating systems, was just an experiment by a developer who worked on multics, who wanted a stripped down version of multics for himself--85.96.29.197 (talk) 10:02, 29 December 2007 (UTC)

Sure, there is an article Multics. It was a big effort and has a place in history. And I'm guessing you meant, unix was an experiment by a developer (Ken Thompson) who had experience in MULTICS? Pete St.John (talk) 17:29, 3 January 2008 (UTC)

sorry, your guess was right, I somehow deleted unix after the comma. Anyway, there is an article about multics, but this article (operating system) has no reference to it--160.75.91.122 (talk) 08:04, 9 January 2008 (UTC)

OK, two things bugged me; the absence of historical context (too broad an axe since creating the seperate history article, which should be linked) and burying Thompson under Plan 9, which is merely the most recent infra-Bell development along the Unix path. So I put a new historical sentence: I changed the title "unix like" to "Unix and unix-like", added Thompson at the top, connected the languages (B --> C) with the development (MULTICS --> Unix --> ....) and linked to the History article. My intent is a short hub connecting history, people, and technology at a very influential place and time. Pete St.John (talk) 17:04, 9 January 2008 (UTC)

Modularity != User Mode

I think the discussion of GUI's contains a slightly inaccurate comment about modularity. You can say that X-Windows runs outside the kernel, but that is not the same as saying that it is modular and Windows or Mac OS were not. Windows (9x and NT) was considerably more modular than UNIX, in the sense that it was built out of components with object interfaces (COM). The GUI was contained in one of three primary dll's that made up the kernel (kernel.dll, user.dll, gdi.dll). UNIX is more modular today, but for a long time was entirely monolithic, to the point where installing a new device driver required compiling the entire kernel (e.g., SUN OS in the 1980s). DonPMitchell (talk) 03:15, 23 January 2008 (UTC)

Much is made of the awkwardness of recompiling the kernel to install a device driver, and it's a good point, but modularity comprises more. I suggest as a thought-experiment, administering an NT server for say Exchange, without a GUI, or with a third party GUI. Compare that to admin'ing a Unix server with your choice of fvwm, other window managers; Gnome, other desktops; or none of that at all. If you ever had to use something like PCAnywhere (pushing pixels over a telephone line, as opposed to concise geometry as in an X or HTML communication) then surely you have some sympathy, at least, for the idea that modularity of the GUI was an issue for us in MS OS's. 3.11 was great, it was a UI on top of the OS, DOS 3.x, but since Win95 about, that's been problematical for developers. Programs don't interoperate easily in an environment which assumes communication with an application is by clicking on coordinates. And while recompiling may be awkward, at least it's possible. Pete St.John (talk) 18:29, 24 January 2008 (UTC)

Extensible GUI

The section about GUI doesn't discuss the important issue of extensibility, for example OLE Automation in Windows and OpenDoc (not to be confused with ODF) on the Mac. An attempt to implement this on Linux was called Bonobo, but I'm not sure what the current situation is on Linux. Extensible GUI is very important commercially, since it permits many applications to act as development platforms, for example the many specialized extensions to MS Office (e.g., for dentists offices, and such). This also permits third party extensions of functionality, for example adding natural language translation or voice recognition plugging into a word processor and extending its interface with new groups of controls or buttons. DonPMitchell (talk) 03:24, 23 January 2008 (UTC)

Lock This Page ??

I have noted needless addenda made to this article by zealots of specific 'OS Camps'.

While this factual freedom of notation does add to the information contained within this article , it does produce some form of redundancy which is not suited to encyclopedia articles e.g (previous revisions of article, and updates that have occurred)

the seperate lines

"Windows is also used on servers." "Mac OS X has both server and personal versions."

was condensed into

"Linux, Mac OS X and MS Windows all have server and personal variants"

On another more obvious note, the history of certain operating systems and the origination of their code and concepts (apple Gui, the circumstances of MS purchasing DOS, the patent wars) and some other PR related matters are at high risk of being changed by the PR departments of those parties.

Saijao (talk) 02:22, 13 February 2008 (UTC)

yeah, this page gets alot of attention, but it's sort of a continual sprinkle, never downpours. I think a zillion people have this watchlisted, and there are few big fights. Every once in a while an "OS Camp" gets a small change in, and later some other camp will get back a small change. I think in general CIS students who care about operating systems, as opposed to avid consumers of particular packages, have disproportionate say and the article stays more or less in bounds. It's kinda funny to watch, but most of the changes are small and they are generally overwritten with new small changes. Probably there are bigger fights at the articles for particular OS's. Pete St.John (talk) 22:40, 13 February 2008 (UTC)

Fair use rationale for Image:Linux sreenshot.jpg

Image:Linux sreenshot.jpg is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to ensure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images lacking such an explanation can be deleted one week after being tagged, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot (talk) 23:13, 13 February 2008 (UTC)

So, I went and looked, and noted this (boldface mine):

If this is a screenshot of free software (GPL, BSD, etc.) do not use this template! Use the license tag that correspond to the license instead, optionally alongside {{Free screenshot}} to categorize as a free screenshot

In other words, since this is a screenshot of free stuff, it's free to use. It merely has to quote the Gnu Public License, which basically is "use as you see fit, but don't lie about authorship". We're not lying about authorship so we're OK. But I'm not a WikiLawyer myself. Pete St.John (talk) 23:24, 13 February 2008 (UTC)

GNU/Linux as the correct naming convention.

"Linux" is not an operating system in itself. It's a kernel. When referring to the complete operating system that combines the GNU userland with the Linux kernel the correct name for that system is the GNU/Linux operating system. It's not correct nor precise or accurate to use just the word "Linux".

The use of GNU/Linux as the correct naming convention for Wikipedia has been confirmed by its founder Mr. Jimmy Wales.
Lightedbulb (talk) 03:16, 15 February 2008 (UTC)

Might be nice to link us that sentence, it might be interesting. Anyway, I generally prefer to speak of FOSS unices (such as FreeBSD or Red Hat) myself, because generally we are distinguishing development environments, and not kernels, and the term "Operating System" is somewhat broad. Even in this article. My own soapbox is that the term "linux" is over-used; Torvalds wrote the most common kernel for the wintel platform in use today, a great contribution, but there have been many contributions to the widespread adoption of Unix (tm) and it's numerous descendants (unices). I can't even cope with an NT environment without gvim. Pete St.John (talk) 19:17, 15 February 2008 (UTC)

If you want to talk about the naming convention, please visit the appropriate place for that discussion -- Talk:GNU/Linux naming controversy.

Discussion of the naming convention has moved to Talk:GNU/Linux naming controversy.

That would be a great place to point out what Mr. Jimmy Wales has said.

Have a nice day. --68.0.124.33 (talk) 14:23, 18 March 2008 (UTC)

CLI

I see no mention of CLIs (Command Line Interface) when I used "Ctrl + F" on the article. Can someone add this in? Possibly after the mention of the GUI? Glad to see BIOS is in there, but shouldn't it also be called the Menu Interface? —Preceding unsigned comment added by 86.2.24.211 (talk) 19:09, 28 March 2008 (UTC)

Article name change to Operating system (distribution)???

This article apparently uses the "marketing" definition of "Operating System" as opposed to the original, technical, and dictionary definition (according to Merriam-Webster:[2] ) of "Operating System" (as already noted in the topic "No user interface?"). By the technical definition, the operating system is a piece of software that is an interface between the system's hardware and other software programs to request services of said hardware (memory management, process management, network communication, etc), things such as user interfaces (graphical or otherwise) fall under the category of "other software programs" as that particular program doesn't make requests directly to the hardware but instead goes through the kernel process for this request.

Unfortunately, marketing and media outlets seem to have a history of misusing and/or under-qualifying terms (technical and otherwise) which already have very specific and set definitions (for example, the term "hacker" has been used inappropriately since at least the early 80's when compared with it's original sub-cultural definition that originated at least a decade earlier, especially concerning the added criminal connotation where as the previous definition was more positive ... or at the least judgment-neutral). To avoid the whole "Operating system" ambiguity, I propose that the article title be changed to "Operating system distribution" with a disambiguation note titled either "Operating system (Computer Science)" or "Operating system (kernel)" or something similar that redirects to the article which uses the technical definition of the word. "Operating system" could still direct here as the "misused" definition is more popularly used among laypersons, but the distinction still should be noted somewhere.MerlinYoda (talk) 19:48, 2 May 2008 (UTC)

Actually, some people (e.g. the multitude of fanatics who insist on referring to "GNU/Linux" in all contexts) believe the opposite is true: the operating system is the "other software" and the kernel doesn't count. The GNU page refers to it as an operating system. I don't really care one way or another, but can we at least make up our mind? I think it would be better to hedge the definitions on all of our pages by saying "some people define this more broadly as..." than to be internally inconsistent in articles that link to each other. « Aaron Rotenberg « Talk « 07:20, 7 May 2008 (UTC)

I agree that there needs to be some sort of consistency, "hedging the definition" might be a good idea to get some semblance of consistency. The whole "Linux or GNU/Linux" fiasco that you touch on comes from this very problem of having various definitions/interpretations for the same word (not counting homonyms) and having no way to infer which interpretation should be used ... at least not from any apparent contextual clues. I assume that when someone says "The operating system on my computer is Linux" that they are using the "technical" definition that is synonymous with "operating system kernel", and if someone says "The operating system on my computer is GNU/Linux" that they are using the "marketing/modernized" definition that is synonymous with "operating system distribution". I've wondered (aside from "official" reasons given) sometimes why "GNU/Linux" instead of "Linux/GNU" ... I think I know why, but that isn't valid to the discussion here really. Then there are those that will say the Operating system on their computer is "Windows" and "Mac OS" which really isn't correct in either sense of the word as "Windows" could be anything from "Windows 95" to "Windows Vista" distribution-wise and I don't know how many different "kernels" all those distributions tie to.MerlinYoda (talk) 20:33, 8 May 2008 (UTC)

The best thing to do would be to fix the article. We do need an article simply titled "Operating system"; we just need to establish what it should talk about.

Regarding your specific complaint: although the original operating systems did not have graphical user interfaces, modern PC operating systems like the Mac and Windows do, and on Windows they are in fact in the kernel (for performance reasons). I do not think it is a misuse of the term; definitions change all the time and we should serve the expectations of our readers and not be pedantic. Fundamentally when writing this article we should be thinking, "What will most readers be looking for when they search on 'Operating system'?" The fact that Unix-derived OS's run the GUI as a user-level task is a technical detail that average users are not aware of (although we probably should mention it in this article). --slashem (talk) 08:00, 7 May 2008 (UTC)

While it is true that more and more is getting put directly into the "operating system kernel" as system capabilities progress for various reasons (performance usually the one given, but I have doubts as to whether it's always the primary one), I could have sworn that, at least for Windows, the actual GUI is a process entirely separate of the kernel (i.e. nothing in the interface is making direct requests of the hardware). This may have changed with Windows Vista or maybe even somewhere in Windows XP, but short of trying to find technical sources it's hard to say one way or the other as I haven't kept track.

Putting this aside, I don't think it's that definitions for words change over time so much as definitions get added over time (and some become archaic and are dropped). This may seem like a pedantic point, but it's important as offhand I cannot think of any term where a possible definition for it includes a previous definition and then adds on top of it ... especially terms originating from fields where precise definitions are important (silly example: a "thigamabob" can, by definition, be a "whatsit" or can be a "whatsit and a thingamajig"). For example, knowing whether a "megabyte" meant 1,000,000 bytes or 1,048,576 bytes was an issue previous to the coining of the term "megbibyte"(sp?) but they were completely separate definitions given for completely separate reasons. "Operating System" on the other hand seems to have had a set definition that was "misused" (for lack of a better term) at a later point to include components that were not part of that definition (i.e. there was a need to coin a new term). It's that misuse that is problematic and is no different than when other words are misused, this is usually due to a perceived understanding of a term while lacking knowledge of the actual definition(s) (many people, myself included, have likely been guilty of this). I don't think an article on a topic stemming from such a a technical background could be well written by thinking "What will most readers be looking for when they search on 'Operating system'?" as most "readers" are like looking to be informed on what exactly an "Operating System" is as opposed to having expectations as to generally what should be in the article.MerlinYoda (talk) 20:33, 8 May 2008 (UTC)

I'm not going to argue with you here over how language changes over time. I do feel very strongly that Wikipedia in general is not here to assert that a common usage of a term is a "misuse"; it is something that dictionaries avoid, for instance. Wikipedia believes no one "owns" an article, well it is even more true that no one "owns" a word. When the vast majority of the world (outside your technical clique) uses a different definition, you are tilting at windmills.

Fundamentally you are not trying to understand your audience, which is not technical and will indeed expect discussion of GUI's here, but trying to change common usage, which is futile. Are you going to try to remove references to cracking from Hacker? --slashem (talk) 22:15, 8 May 2008 (UTC)

Perhaps "misuse" isn't the best word to use ... still the "common usage" definition is pretty much a bastardization of original "tecnical" definition (and I'm sure "Operating System" isn't the only word to be "redefined" in common usage this way despite having an otherwise accepted, long standing, and standardized definition). Also, I would not try to remove references to cracking from Hacker as long as they are made clear that the context in which the word is being used is from the more recent "media-created" definition of Hacker as opposed to the original definition. If anything, I would try to frame them into that context rather than remove them, but that is a different situation all together. Still, just because there's a "common usage" of a term, that doesn't necessarily mean that it should be the presented over all others. It depends greatly on the circumstances of that usage.

In any case, the recent rewrite as of now is actually pretty good. Although, if we really are to "understand our audience" as you say, it might be of some worth to add a note on how the "common" perception of an "Operating System" is "The software that provides an interface to the hardware as well as the the software that are packaged with it.". A layman would probably agree with that statement "Notepad is part of the Windows Operating System" even though the original definition would disagree with this claim. MerlinYoda (talk) 21:28, 9 May 2008 (UTC)

Agreed. --slashem (talk) 21:39, 9 May 2008 (UTC)