Talk:Operating system/Archive 5

Archive 1

←

Archive 3

Archive 4

Archive 5

Archive 6

Market share changes; the alphabet doesn't

Although I can see why it is tempting to reorder the examples of OS's to suit our view (or the view of the market), the market and our cultural view changes over time. In this case another editor will want to swap the list around to suit and then another and another. To see examples of this behaviour we need only look at the article history. I set the list in alphabetical order some time ago (and at the time added a note hinting as to why). I strongly suggest that the article features lists that will be reordered (time and time again for various and often pointless reasons) to be ordered alphabetically and left that way. fg^t_c 21:26, 5 November 2011 (UTC)

• I was the editor who recently changed the order to reflect market share. I did it because I saw another re-ordering get reverted and thought "oh, the ordering seems to be chosen at random, so I'll order it in a less arbitrary way." If you prefer to order it alphabetically, I have no specific objection, so please proceed. SocratesJedi | Talk 03:09, 9 November 2011 (UTC)

No particular order matters which is actually the problem. Everyone has their own view of which way around they should be so the order gets shifted around for all kinds of reasons (I think typically it comes down to favouritism). The only order that has no connotations seems to me to be alphanumerical. Unless we are enslaved and indoctrinated by 12 fingered telepathic aliens, I think there is little chance the order will get out of date. I added a note a while back reading that the list was not exhaustive and was alphabetical in order to try to avoid this constant flux but the note was thought to be un-encyclopaedic. Alphabetical or alphanumerical lists are however very encyclopaedic so I really think that would be best. Thanks for your input. fg^t_c 03:28, 9 November 2011 (UTC)

I also agree it should go alphabeticaly. As you say it will never change. --JetBlast (talk) 04:13, 9 November 2011 (UTC)

me too; go the alphabet Steev (talk) 20:46, 14 January 2012 (UTC)

Mini computers

I'm not much of an author so I will leave editing the article to a better writer with more details. This article misses not just an operating system but a class of computers. 'Mini computers' include IBM's AS400 which can run several operatng systems including OS400, and HPs HP3000 using the MPE operating system. (I was an HP3000 system operator.) Introduced sometime in the 1970's, HP stopped making the HP3000 in 2002 and stopped supporting it in 2004. I think that the MPE operating system is now owned by OpenMPE. — Preceding unsigned comment added by 4.254.228.86 (talk) 14:33, 15 May 2012 (UTC)

"User interface" section

I guess that the statement

“ The user interface is not actually a part of the operating system ”

was a result of incorrectly narrow perception of the "operating system" concept by some user, not some point of view well-established in the industry. Incnis Mrsi (talk) 16:45, 15 September 2012 (UTC)

Computer crashes

I do not think that the IP’s addition is something really “out of place”, so not only this Wtshymanski’s edit constitutes a biting of a newcomer, but his edit summary deceives Wikipedia users about the true nature of that reverted addition. Opinions? Incnis Mrsi (talk) 06:43, 8 April 2013 (UTC)

Normally I would assume good faith, but given Wtshymanski's long history of deliberately making "mistakes" in order to remove content (look at how many times has he proposed a speedy deletion for a reason not on the list of criteria for speedy deletion), I must conclude that he knew perfectly well that the top part of his edit "moved" a reference without changing what is displayed on the page and that the edit comment was a deliberate attempt to deceive the reader into missing the content removal below.

There is a partially-constructed WP:RFC/U at User talk:DieSwartzPunkt/WTS, and this search turns up many previous issues with Wtshymanski's behavior. --Guy Macon (talk) 08:36, 8 April 2013 (UTC)

What *are* you talking about? Is that still around? --Wtshymanski (talk) 13:54, 8 April 2013 (UTC)

If you think it's an attack page, report it at WP:ANI.

If you keep getting into conflicts with other users, perhaps you should use a bit of troubleshooting logic and figure out what the common factor is. --Guy Macon (talk) 15:53, 8 April 2013 (UTC)

Linux is not an operating system

Linux is just the kernel of several operating systems. Operating systems provide the interface between applications and the hardware device handling program. The kernel handles the hardware devices. The operating systems that use the kernel Linux are GNU, BSD, Unix and others. So the article should replace the term "Linux" with "GNU, Unix and BSD". Quiliro (talk) 22:17, 20 December 2011 (UTC)

Too much water under that bridge to turn back now. Even if strictly correct, "Linux" is known to be an operating system. fredgandt 00:00, 21 December 2011 (UTC)

We don't entertain the GNU/Linux naming controversy here. The Wikipedia manual of style specifies that we use the most commonly used names for things. Thus the article about Linux is titled Linux, not GNU/Linux and there is a separate article on the Linux kernel. (As a caution, editors who edit war over this generally earn blocks for violating this policy.) Yworo (talk) 00:06, 21 December 2011 (UTC)

Technically Linux kernel is the operating system, and it is not wrong to call Linux kernel only as Linux because it is its name after all. It is just plain wrong to call anything else than Linux kernel (bundle of different software) as Linux. And Linux is not used by any other OS because it is technically impossible. Linux is a monolithic operating system, like original Unix and BSD's are. While HURD is server-client operating system. Golftheman (talk) 08:43, 26 April 2012 (UTC)

Fire fox — Preceding unsigned comment added by 65.91.116.34 (talk) 22:27, 26 April 2012 (UTC)

Although you're correct about the general usage, I think that in the article that is about operating systems it's perhaps worth being a little more precise. —mako๛ 23:15, 18 May 2013 (UTC)

Content needs updating

Some of the content has been recently made outdated; an example is in the OS X section which states that releases are named after big cats. As of OS X Mavericks they are now named after places in California. Wikipedia does not allow me to edit this. DwellingDreamr (talk) 21:40, 26 October 2013 (UTC)

Semi-protected edit request on 14 December 2013

In the beginning of the second sentence in the first paragraph, there is a grammar mistake. Instead of "An operating system", you wrote "The operating system". That has to be corrected please!

Hadestempo1 (talk) 05:52, 14 December 2013 (UTC)

• Declined The sentence is both gramatically and contextually accurate as written. Of the system software in a computer system (which may include an OS, BIOS, etc.), the operating system is an essential component. Dwpaul ^Talk 05:58, 14 December 2013 (UTC)

Further Reading

There are many more external references about Operating Systems that could be added to this page.

One good example: a link to the free online book "Operating Systems: Three Easy Pieces" (http://www.ostep.org). Unlike the other books mentioned here (which is not a particularly comprehensive list), this book is accessible online without cost.

Other books include classics about particular operating systems, such as the BSD "devil" book - there are many others like this.

Unfortunately, the page is not editable. — Preceding unsigned comment added by 24.196.83.146 (talk) 12:28, 29 January 2014 (UTC)

Simple Defination of Operating System

In Simple way we can discribe as o/s is a collection of softwares that work as interface between computer hardware and operating system.

Some mention of the following might be in order: Atlas Supervisor, seeing as it has been descried as "the most significant breakthrough in the history of operating systems." — Preceding unsigned comment added by 80.7.147.13 (talk) 12:16, 12 April 2014 (UTC)

Semi-protected edit request on 25 April 2014

2.68.231.247 (talk) 21:45, 25 April 2014 (UTC)

Not done: Empty request. —KuyaBriBri^Talk 22:28, 25 April 2014 (UTC)

Semi-protected edit request on 1 May 2014

2.51.250.32 (talk) 17:25, 1 May 2014 (UTC)

Not done: it's not clear what changes you want made. Please mention the specific changes in a "change X to Y" format. NiciVampireHeart 17:31, 1 May 2014 (UTC)

Linux Kernel is mentioned without reference to Linus Torvalds. Several other names are cited in the article. The strategy that Linus used to develop, improve, and register changes to the official kernel, namely the internet, is noteworthy as it used the world wide web "system of operating systems" to foster a new UNIX like OS. Lmanderville (talk) 13:32, 14 May 2014 (UTC)

Real-time

someone should point out that real-time operating systems are defined to be operating systems with exactly known wall-clock response time. that is, they respond to specific requests in specific amounts of time, like 2 seconds. — Preceding unsigned comment added by 173.228.28.61 (talk) 18:24, 17 May 2014 (UTC)

Under the 'Types of operating systems' section in the 'Real-time' paragraph the writer starts to talk about event driven vs time shared systems. I fail to see how these design notions are specific for real time operating systems. I would agree that in a real time OS these things require a closer look, but the concepts are so universal that they are wider (and older) than the concept of the OS itself. — Preceding unsigned comment added by 83.87.238.229 (talk) 17:45, 29 January 2013 (UTC)

OS vs. system software

"An operating system (OS) is a collection of software". Oric-1 (my old computer) links here, and similar ZX Spectrum lists Sinclair BASIC as an operating system.. These and similar old computers had what I would not consider operating systems (in the modern sense) or what..? [And DOS..] Why am I bringing this up. They certainly are system software. They act as a shell and provide a user interface and I consider that a part of an OS. Their ROMS are a collection of software routines, but are they one software, not a collection? A kernel on it's own (such as the Linux kernel) is not a full OS or would you disagree? "Linux" with a shell (and/or only GUI) is. Mac OS (older versions) are also not an operating system without their UI.

System software: "The operating system (prominent examples being z/OS, Microsoft Windows, Mac OS X and Linux), allows the parts of a computer to work together by performing tasks like transferring data between memory and disks or rendering output onto a display device. It also provides a platform to run high-level system software and application software." [.. and then a separate example..] "A user interface". UI is system software, is it being treated as separate from the OS? comp.arch (talk) 15:28, 22 May 2014 (UTC)

controls the system hardware including memory, processor time and storage. — Preceding unsigned comment added by 78.101.223.69 (talk) 20:03, 26 May 2014 (UTC)

Semi-protected edit request on 7 November 2014

Linux is not an operating System its a kernel, if you're talking about the operating system that uses it you're talking about GNU/Linux. thanks. Sand49 (talk) 12:47, 7 November 2014 (UTC)

See the second last box at the top of the page on Talk:Linux. As that box explains, we refer to "Linux" as on OS because that's what the sources say. Stickee (talk) 13:22, 7 November 2014 (UTC)

Semi-protected edit request on 12 November 2014

could you please change Linux to GNU/Linux, as linux is just the kernel where GNU/Linux is the operating system. Thanks Sand49 (talk) 06:10, 12 November 2014 (UTC)

See above response. Stickee (talk) 10:07, 12 November 2014 (UTC)

Operating Systems

It should be called GNU/Linux, not Linux.

http://www.gnu.org/gnu/why-gnu-linux.en.html — Preceding unsigned comment added by 178.9.181.142 (talk) 17:45, 26 March 2015 (UTC)

Operating System and Kernel Distinction

The definition of an operating system is included, the contradicted with examples that are do not perform those functions. Microsoft Windows does not manage hardware resources and give software the ability to use a diverse number of hardware components through common interfaces. The NT kernel (I would tend to argue operating system) does.

This aspect of the article should be evaluated. — Preceding unsigned comment added by 75.170.136.2 (talk) 18:44, 30 May 2015 (UTC)

Essential

"The operating system is an essential component of the system software in a computer system."

No it isn`t. It is perfectly possible to run a computer system without an operating system and in times past this was frequent, if not normal, practice in small systems. All that is essential is either a pre-loaded application or OS, or some form of bootstrap loading mechanism. In the modern world even embedded processors tend to have operating systems - after all it makes it all much easier. But that doesn`t make it essential. Bagunceiro (talk) 22:58, 10 July 2015 (UTC)

Curious

What does this sentence from the article actually mean? 70.247.167.104 (talk) 06:03, 19 February 2014 (UTC)

All methods require some level of hardware support, which doesn't exist in all computers.

What we're trying to get at is that -- no matter which method of memory protection an operating-system programmer chooses when writing an operating system -- there will be some computers out there that can't run that operating system because they don't have the necessary hardware to support that particular method of memory protection.

There are lots of different techniques that an operating system can implement memory protection.

Each specific technique requires a specific kind of hardware support -- for example, paging requires a paging memory management unit. For example, memory segmentation requires segment registers. For example, efficient hardware virtualization can be implemented on hardware that supports the Popek and Goldberg virtualization requirements.

The entire point of the μClinux operating system is to support the many computers do not include a MMU, and so they cannot run standard Linux which uses a memory protection system that requires a paging MMU.

If you choose any one of those 3 techniques, there are lots of computers that don't have the hardware support for it -- even though they may have hardware support for one or more other techniques.

Is there some way of clarifying that section of the article without going into so much detail?

--DavidCary (talk) 15:44, 13 August 2015 (UTC)

Semi-protected edit request on 24 November 2015

The most recent version of OSX is El Capitan, not yosemite. 75.157.255.134 (talk) 00:52, 24 November 2015 (UTC)

 Done - I've removed that sentence, as it contains too much detail and would be something else to maintain as new versions of OS X come out, if it were merely modified. Dhtwiki (talk) 11:36, 24 November 2015 (UTC)

SteamOS

New to wikipedia talk section so excuse any protocol breaks I make, but SteamOS is definitely based on Linux, so the header section really shouldn't list it as a separate UNIX based OS. — Preceding unsigned comment added by 18.189.70.27 (talk) 16:15, 4 December 2015 (UTC)

I've put SteamOS inside the parenthesis after Linux, and near Debian, with which it's associated. The OS isn't mentioned later on, so we've got more detail in the lead than in the body, which shouldn't be the case. Dhtwiki (talk) 23:26, 4 December 2015 (UTC)

There`s far too much in the list in the lead. We don`t need to list every existing OS as an "example of [a] popular modern operating system". I think we should restrict it at this level to Windows, Linux distros (without listing them), Android and OS/X. Further explanation and expansion to follow in the body. Bagunceiro (talk) 00:24, 5 December 2015 (UTC)

Change seven to six

In the fourth alinea it states that the first seven of above mentioned operating systems are unix based, with the recent edit of putting steamOS under the linux category this makes it now the first six. JJK96 (talk) 21:49, 10 December 2015 (UTC)

I've made the change. Thanks for pointing that out. Dhtwiki (talk) 08:59, 11 December 2015 (UTC)

First PC to run Windows

The article claims that the first PC to run Windows was the PS/2. However, this was launched in 1987, two years after the first release of Windows. Suggest this needs clarifying or scrapping. Bagunceiro (talk) 16:12, 19 December 2015 (UTC)

Maybe someone was thinking of OS/2. That sentence can be scrapped. It would be hard to determine the first computer to "run" any operating system. Also the article says "Microsoft Windows originated in 1985...", which is probably better put as "first released in 1985". Dhtwiki (talk) 11:19, 20 December 2015 (UTC)

List of "popular" operating systems

I want to simplify the last paragraph in the lead but I`m not entirely sure what to do. As it stands it is inaccurate and selective, but not selective enough - worse than useless I think. My preference is simply to delete it; I don`t think it adds anything to the article and left there it is an open invitation for everyone to add their own preferred OS and it just grows out of all proportion as it is now. The alternative might be to place a strict limit on it - something like "Examples of popular modern operating systems include Microsoft Windows, Apple OS/X and Linux variants including Android" would do as what I suppose it is trying to do is make clear what is meant by an OS. However, without some agreement here neither of these would hold. Does anyone have any views on this matter? Any other proposal? Bagunceiro (talk) 23:19, 10 December 2015 (UTC)

I'd think about moving some of the detail further down, rather than just deleting. It might be of some use to place some of the more popular offshoots with their parent OSs. I don't like peremptory deletions, unless the material is of absolutely no value. Someone wanted to see that information placed, although I agree that mere laundry lists of OSs can grow and be hard to maintain. Dhtwiki (talk) 09:04, 11 December 2015 (UTC)

I´d agree with the principle of that, except that I think it is already there. Bagunceiro (talk) 11:32, 11 December 2015 (UTC)

I thought that wasn't the case, but see that it is. So, simplifying the lead, as you suggest, especially reducing the too-long parenthetical after Linux, is OK, without worrying about losing information (I'd not heard of SteamOS before). Dhtwiki (talk) 17:51, 11 December 2015 (UTC)

OK, well with the lack of further input I'm going to make the change. Bagunceiro (talk) 11:49, 16 December 2015 (UTC)

Yes, I put real-time operating system (RTOS) in the list. They are also operating systems, and the article should reflect that, and I think recognize the most popular ones, even if the general public doesn't know them. Taking them out, that are more popular than say OS X (while leaving that in), seems biased to me. comp.arch (talk) 11:44, 4 February 2016 (UTC)

It doesn't claim this list is exhaustive or even "popular" any more. It's just a short list of examples which a typical reader may have heard of.

Actually I'd like to remove the list altogether - I don't think it achieves anything except to generate controversy since people get upset that their own favourite is missing. Suggest we delete this and leave it to the examples section (and add RTOSs, for example, to that). Bagunceiro (talk) 12:13, 4 February 2016 (UTC)

Maybe I can live with you just dropping the list. At the time I just added the most popular RTOSes. The list was already there. I'm not sure it helps to list OSes in an example section, then you might just get controversy there.. [and the lead should summarize it/articles]. The list I left in included only popular (I expect you had issue with z/OS, that is most popular in its category if you define them narrowly, maybe it's not "modern" however..). It seems strange to list no examples [in the lead].. but then again Car doesn't.. Programming languages, name C and Perl. I'm ok with that. At least C. comp.arch (talk) 13:22, 4 February 2016 (UTC)

I don't have any quarrel with any of them per se. What I object to is a long and complicated sentence which just boils down to a list of names for people to fight over - it adds nothing. Bagunceiro (talk) 17:06, 4 February 2016 (UTC)

Before I saw this discussion I re-added the short list of OSs as it existed on February 3, along with its accompanying hidden comment. I don't think that it hurts to exemplify in the lead the concept of an OS, using a short list of software people are likely to be acquainted with. The comment explains that the list isn't meant to be exhaustive. Dhtwiki (talk) 16:49, 5 February 2016 (UTC)

Agreed, provided it stays that way. We shall see. Bagunceiro (talk) 17:35, 5 February 2016 (UTC)

Question in Arabic

هل يمكن للرواى الخاص باى جهاز ينطق باللغه العربيه ؟بدلا من الانجليزيه . اذا افترضنا جدلا ان دور الراوى لذوى الاحتياجات الخاصه من ضعف النظر والسمع اذا لماذا لايكون الراوى عربى ؟ — Preceding unsigned comment added by 156.202.219.99 (talk) 23:11, 18 March 2016 (UTC)

Is your question why equal attention isn't paid to Arabic, as to vision and hearing impairment, with regard to accessibility? In my experience, operating systems come with considerable "locale" facilities, to change the prompts, error messages, etc., to the language of one's choice. This might be better addressed at Arabic Wikipedia, in any case, although their OS talk page hasn't shown activity since 2008. Is that why you're asking here? Dhtwiki (talk) 21:53, 19 March 2016 (UTC)

Where is Unix?

I don't mean to be an "operating system chauvinist" but the History section seems to jump directly from mainframes to microcomputers, omitting minicomputer OSs, and particularly Unix, which most people would agree is a historically significant OS. The word "Unix" only occurs 4 times, in the context of the development of Linux. Chetvorno^TALK 17:22, 25 March 2016 (UTC)

I see that there is little to no discussion of minicomputers per se in the History section. However, there is plenty of discussion of Unix, and Unix-derived systems, in the Examples section, more than on any other operating system, "Unix" being mentioned more than four times, I would say. Dhtwiki (talk) 23:58, 25 March 2016 (UTC)

Incorrect statement

The statement on realtime OS that "A real-time operating system is an operating system that guarantees to process events or data within a certain short amount of time" is misleading. The realtime characteristic isn't so much about being fast as being predictable in its time characteristics and considering deadlines in its scheduling algorithm. Granted, realtime operating systems often are, in fact, very fast but speed alone isn't the defining characteristic. — Preceding unsigned comment added by Ejs1985 (talk • contribs) 16:56, 24 June 2016 (UTC)

Possibly change "within a certain short amount of time" to "by a specific moment in time". Dhtwiki (talk) 02:41, 25 June 2016 (UTC)

Semi-protected edit request on 7 August 2016

Please link Virtual machine image to the stub article I just created. Thank you. --2601:285:101:A67A:88DB:4009:605E:A8E9 (talk) 22:24, 7 August 2016 (UTC)

 Not doneSee Below The article is in draft space, and it looks pretty rough to me. Let's wait until the article is accepted before linking (although I see that you've been linking at other articles). Dhtwiki (talk) 23:50, 7 August 2016 (UTC)

Really? Why not let a more experience editor who understands red links on Wikipedia better than you do, respond to this request. Thanks. --2601:285:101:A67A:88DB:4009:605E:A8E9 (talk) 00:21, 8 August 2016 (UTC)

Even without an article the link is appropriate. "Looks rough" as an editorial comment has no value for rejection of a non-controversial link. --2600:387:1:811:0:0:0:99 (talk) 12:13, 8 August 2016 (UTC)

 Done You shall have your red link. I was in part reacting to the literal wording of the request, which implied an existing article (a "stub" is an article type in main space). The policy on red links is more about not removing those already placed. I think this is a questionable case, where you're requesting a protected edit to establish one and where one editor has systematically placed them on many other articles (where the lack of such links suggests that no need is felt by other editors). Dhtwiki (talk) 21:28, 8 August 2016 (UTC)

"Prominent" operating systems?

Hi

Looks like Dhtwiki and I are entangled in a bit of dispute. I contend that "Linux" is neither popular nor prominent.

But the solution is simple: Wikipedia:Verifiability. I added source. Anyone contending that Linux is popular, please add one.

Thanks.

Best regards,
Codename Lisa (talk) 06:44, 24 August 2016 (UTC)

This type of laundry list is always unhelpful. They do not have any useful content and simply provide a platform for squabbling around the editors' respective favourites. I still contend that we would be better off without it altogether. But as an attempt to bring an(other) end to the constant bickering here I've made making the statement more concrete. Perhaps this will at last put an end to it? Bagunceiro (talk) 11:56, 24 August 2016 (UTC)

Change ReactOS classification

ReactOS in its' current state is by no means a "hobbyist operating system": It's being developed as a full operating system by a dedicated team, whose developing it based on funds and donations. To top it all, the ReactOS team releases new versions from time to time whilst improving the operation system. Please change this. --ArmyMan007 (talk) 12:47, 20 September 2016 (UTC)

Add Link to Artificial Intelligence Operating System

If you could add a link in the text to the topic of Artificial Intelligence Operating System, for this new class of OS, it would be appreciated. (i.e. - "Other specialized classes of operating systems, such as embedded, real-time systems, and artificial intelligence, exist for many applications...") — Preceding unsigned comment added by Aios3837 (talk • contribs) 00:20, 23 September 2016 (UTC)

Semi-protected edit request on 13 January 2017: Grammar and clarity edit for third paragraph

I am proposing grammar and clarity edits to the unwieldy third paragraph in the introduction. This paragraph is a point of contention in the talk page, but I am not proposing any substantial edits to rectify its relevancy issues, only clarity and grammar.

The sentence which reads: "In the mobile (smartphone and tablet combined) sector, based on Strategy Analytics Q3 2016 data, Android by Google is dominant with 87.5 percent or growth by 10.3 percent in one year and iOS by Apple is placed second with 12.1 percent or decrease by 5.2 percent in one year, while other operating systems amount to just 0.3 percent.[4]" Should read: "In the mobile (smartphone and tablet combined) sector, according to third quarter 2016 data, Android by Google is dominant with 87.5 percent of the market and a growth rate of 10.3 percent per year. iOS by Apple follows with 12.1 percent of the market and a per year decrease in market share of 5.2 percent.[4]" Base0x10 (talk) 18:50, 13 January 2017 (UTC)

I've incorporated (some of) your suggested changes in that lead sentence (which I've maintained as one sentence). Dhtwiki (talk) 11:29, 14 January 2017 (UTC)

Semi-protected edit request on 6 June 2017

IBM made an approach to Burroughs to licence MCP to run on the AS/400 hardware.

In the above sentence the accepted usage is license, as in the act of obtaining permission, rather than licence, a certifying document. 96.84.88.99 (talk) 20:59, 6 June 2017 (UTC)

In my experience in the software industry, both terms are used in American English somewhat interchangeably. Please provide a WP:RS for the claim that the difference in spelling has a functional effect on licensing agreements. Eggishorn (talk) (contrib) 22:29, 6 June 2017 (UTC)

Done Despite Eggishorn's objection to the validity of the nominator's claim, I went ahead and made the change because Wikipedia Manual of Style requires spelling consistency and there were two other instances of "license" on the article already. That said, I did look at www.dictionary.com/browse/license and www.dictionary.com/browse/licence just to be sure there is no unforeseen circumstances. —Best regards, Codename Lisa (talk) 18:44, 7 June 2017 (UTC)

Is Chrome OS section WP:UNDUE

I'd prefer not to modify that section myself, but I'm not sure how ChromeOS is notable enough to have a separate section on this page. Power~enwiki (talk) 22:52, 13 July 2017 (UTC)

Have to agree with you here, @Power~enwiki:. However, I do understand your reticence because I can envision a potential reversion on the grounds that it's a "sub-subsection" of a subsection (Linux). However, I am in agreement with you since ChromeOS and android are mere flavours of the linux kernel, and thus do not warrant any special attention because market share is not relevant to this topic. — Preceding unsigned comment added by 66.186.166.65 (talk) 23:03, 13 July 2017 (UTC)

My reticence is only because I was previously a Google employee and wish to avoid any appearance of WP:COI. Power~enwiki (talk) 23:07, 13 July 2017 (UTC)

Yes, I think it is. We could add links to Chromium OS and Chrome OS to the ever-growing list at the end of the Linux section. ~Kvng (talk) 04:36, 21 July 2017 (UTC)

Firmware functioning

I want to know that why firmware doesn't need os to function. — Preceding unsigned comment added by 39.42.55.174 (talk) 18:00, 4 December 2017 (UTC)

Because it's the other way around - The computer requires the firmware (by which we're talking about BIOS, UEFI etc.) to run the OS. In effect the OS sits "on top" of the firmware.

Having said that, the sentence you are referring to (All computer programs, excluding firmware, require an operating system to function.) is at best misleading. Only programs that are written to run in an OS environment require an OS. That is the vast majority, of course, but excludes much embedded and specialised code. Bagunceiro (talk) 19:37, 4 December 2017 (UTC)

I've removed the sentence. There are a great many examples through history of computers that had no operating system, yet they seemed to be able to run programs just fine. Jeh (talk) 19:47, 5 December 2017 (UTC)

I was trying to think of the best way to improve it, and I think you've found that! Thanks. Bagunceiro (talk) 20:06, 5 December 2017 (UTC)

Sure. Decommission Wikipedia and you don't need to improve it anymore.

Also, search the history and find all the examples you want. There are trillions of apps that need an operating system.

On the other hand, computer programs that don't need an OS are firmware, bare-metal hypervisors, bootloaders (well, a bootloader is always part of an OS) and bootkits.

FleetCommand (Speak your mind!) 21:22, 5 December 2017 (UTC)

Hyperbole much? Nobody is saying this is the last improvement that needs to be made. I'm used to much better from you than this, FC. Jeh (talk) 21:26, 5 December 2017 (UTC)

@Jeh: Really? Thanks. I take that as a complement. Nobody has ever said that he or she is used to "much better" from me. It is, however, safe to say that all application software and all mobile apps (which could either be application or utility, despite the word "app") need an OS. Supreme regards, FleetCommand (Speak your mind!) 21:32, 5 December 2017 (UTC)

I don't think it's safe to say that at all. But I guess it depends on what you mean by an OS. Is a simple loader, one that provides no services to the program after it's been loaded, and which must be started from scratch in order to load the next program, an "OS"? I don't think so. I've used and written a lot of "application software" that depended only on a loader that was more or less hardwired into the machine. Jeh (talk) 22:13, 5 December 2017 (UTC)

A simple loader is not an OS. But also, a computer program that does not require an OS is, by definition, neither application nor utility, bearing in mind that these two categories do not encompass all computer programs. But they do encompass all apps; because apps are either applications (mostly) or utility (rarely).

If what you make supervises hardware resources, then it is an OS. "Supervisor" is another name for operating system. FleetCommand (Speak your mind!) 09:49, 7 December 2017 (UTC)

"a computer program that does not require an OS is, by definition, neither application nor utility". I don't know how you can believe that, other than that your experience in the industry must not include older machines. There was no "operating system" on the very successful IBM 1401, for example, but a great many shops ran business applications (payroll, billing/receivables, banking,general accounting, inventory, student management, etc., etc.) on them, as well as development for all of the above, plus utilities (in particular tape-to-printer and card-to-tape, in service to e.g. a 7090) on it. No job monitor, no OS "supervising hardware resources", etc. (What "supervised hardware resources" was the operations staff. ;) ) Jeh (talk) 12:01, 7 December 2017 (UTC)

Semi-protected edit request on 2 April 2018

Please change "This is achieved by time-sharing, dividing the available processor time between multiple processes that are each interrupted repeatedly in time slices by a task-scheduling subsystem of the operating system." to "This is achieved by time-sharing, where the available processor time is divided between multiple processes. These processes are each interrupted repeatedly in time slices by a task-scheduling subsystem of the operating system." The original sentence is hard to read because of the length so I propose breaking it down into two.

Please change "Unix-like operating systems, e.g., Solaris, Linux, as well as AmigaOS support preemptive multitasking." to "Unix-like operating systems, such as Solaris, Linux, and AmigaOS, support preemptive multitasking." GrammarAndShape (talk) 00:48, 2 April 2018 (UTC) I thought using the plain English "such as" will make for easier reading.

 Done. Slightly modified the second change because AmigaOS isn't a Unix-like. Bagunceiro (talk) 21:18, 2 April 2018 (UTC)

Semi-protected edit request on 22 May 2018

In Examples section, macOS is not a BSD descendant. Can you please fix that? 192.107.120.90 (talk) 17:24, 22 May 2018 (UTC)

 Not done: please provide reliable sources that support the change you want to be made. Eggishorn (talk) (contrib) 18:15, 22 May 2018 (UTC)

Semi-protected edit request on 10 June 2018

72.2.76.227 (talk) 19:19, 10 June 2018 (UTC)

 Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format and provide a reliable source if appropriate. LittlePuppers (talk) 19:26, 10 June 2018 (UTC)

Say your thoughts about this, and whether or not this should be added to the article:

An operating system/kernel manages the computer's resources and distributes them among the apps that need those resources.

Windows NT is an operating system, but it's not a kernel because windows nt's kernel can't work alone, as it's not monolithic.

Linux is a kernel and an operating system because Linux can work alone, because linux is a monolithic kernel.

Windows and Ubuntu are distributions (distros), as they are a ready to use environment, as they include an operating system and pre-installed apps, as well as a graphical user interface.

Problem is, there's only one Windows NT distro, windows, but there are hundreds of linux distros. so, should windows be called a distro or not?

What's the difference between kernel and operating system?

"An operating system (OS) is system software that manages computer hardware and software resources and provides common services for computer programs." "The kernel is a computer program that is the core of a computer's operating system.... The kernel's primary function is to mediate access to the computer's resources.... The kernel has full access to the system's memory and must allow processes to safely access this memory as they require it.... " btw take a look at the spanish language wikipedia articles for linux, sistema operativo, and microsoft windows. Pancho507 (talk) 05:06, 16 May 2019 (UTC)

What means “Linux can work alone” – without any user-space process? It could act as a network router, for example, but generally it is not used in this way. What means “windows nt's kernel can't work alone”? Surely the sole ntoskrnl.exe image is insufficient for anything, but the file contains only a part of Windows kernel. One has to load several files to obtain a viable NT kernel in the primary memory, but otherwise it functions no different from Linux in principle. There is a lot of confusion in the Pancho507’s posting above. Incnis Mrsi (talk) 12:31, 18 May 2019 (UTC)

Semi-protected edit request on 18 June 2019

The Linux subsection (section 3.2.1) should be on the same level as macOS or Microsoft Windows (i.e. should be section 3.3 in the current document structure). Linux does not evolve from MacOS and in fact predates it, so there is no justification for it being under its subsection. 153.98.68.196 (talk) 11:21, 18 June 2019 (UTC)

 Done Linux is Unix-Like, so I moved it to that section. RudolfRed (talk) 18:54, 18 June 2019 (UTC)

I've made some further adjustments, putting the macOS subsection ahead of the Linux one because the former began as NextStep OS in 1988, although not becoming the underpinnings of macOS until ten years later, while Linux started in 1991. Dhtwiki (talk) 10:44, 19 June 2019 (UTC)

First OS written in a high-level language

Both Burroughs MCP (1961) and Unix (1973) are claimed to be the first operating system written in a high-level language. One or both of those claims should probably be qualified. It's certainly true that Unix portability owed much to being written in C, but MCP definitely predates Unix. Nukeqler (talk) 15:55, 8 November 2019 (UTC)

Semi-protected edit request on 26 August 2020

Change "Interrupts are central to operating systems, as they provide an efficient way for the operating system to interact with and react to its environment. The alternative – having the operating system "watch" the various sources of input for events (polling) that require action – can be found in older systems with very small stacks (50 or 60 bytes) but is unusual in modern systems with large stacks. Interrupt-based programming is directly supported by most modern CPUs. Interrupts provide a computer with a way of automatically saving local register contexts, and running specific code in response to events. Even very basic computers support hardware interrupts, and allow the programmer to specify code which may be run when that event takes place" to " Interrupts are system calls, generated by user or software using operating system. They provide the efficient way for the operating system to interact with user and the system components. Most modern CPU's support interrupt-based programming. This makes the system responsive and efficient." VidyaVinaya (talk) 18:28, 26 August 2020 (UTC)

The suggested rewording makes no sense at all. Interrupts are not necessarily system calls, they are not generally generated by the user or software using the operating system. DOS did use software interrupts as system calls, as a means to do an indirect jump so that BIOS and DOS routines were not tied to specific addresses. This was essentially an implementation detail. Interrupts are no more efficient than normal system calls – indeed they almost inevitably cause a context switch which a straightfoward subroutine call does not. The proposed rewording seems to show a basic misunderstanding of what interrupts are used for. In most modern systems (even embedded systems), very little code is run under interrupt: instead, the interrupt will generally just add something to the scheduler's event queue and return pronto, if for no other reason than interrupts generally operate in a privileged mode of execution. (There's all kinds of other potential problems with priority inversion and so forth., which a good scheduler should sort out.) Actually, embedded systems – especially safety-critical systems – often do poll hardware rather than use hardware interrupts, because it makes for a more deterministic system. The reason the original text is somewhat long is that it is a far subtler balance than the proposed rewording makes out. I suggest rejecting the req8est. 84.236.27.182 (talk) 06:29, 2 October 2020 (UTC)

 Not done @VidyaVinaya: I agree with this IP editor that your suggested edit is a fundamental misunderstanding of how modern operating systems work. If anything, since interrupt handlers can in general require a complete context switch in effect, interrupts can be quite inefficient.--Jasper Deng (talk) 06:39, 2 October 2020 (UTC)

Broader Scope of the Concept

The phrase "operating system" can refer to important systems other than software systems. For example, capitalism, democracy, mass communications, journalism, internet, and other institutions create the social, political, and economic operating systems we live in. I suggest creating a disambiguation page that lists these broader operating systems. See, for example: https://en.wikiversity.org/wiki/Living_Wisely/Improving_our_Social_Operating_Systems --Lbeaumont (talk) 12:02, 28 September 2020 (UTC)Thanks!

 Disagree Wikipedia is not a dictionary and that use of the term is extremely uncommon (I was able to find just one book using it, nothing more than that). The page you mentioned above was written by yourself - Daveout(talk) 12:32, 28 September 2020 (UTC)

 Disagree To disambiguate you would first need another article (or section) for it to be disambiguated from. Even then, this is clearly the primary topic. If you can link to relevant text elsewhere in WP, just add a hatnote. I think what you are describing are just examples of systems. 84.236.27.182 (talk) 06:53, 2 October 2020 (UTC)

NPOV: undue emphasis on microprocessors

Large swaths of the article are written as though every computer was a microprocessor. Many of the OS concepts predate microprocessors, and most of the referencesa to microprocessors violate WP:NPOV. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 23:49, 30 June 2021 (UTC)

User Interface

The article begins with "An operating system (OS) is system software that manages computer hardware, software resources, and provides common services for computer programs." What kind of operating system does not provide a user interface? — Preceding unsigned comment added by B0ef (talk • contribs) 22:59, 13 July 2021 (UTC)

Not all operating systems have human users. Or, running a display panel is just another task for the operating system. I speculate most computers don't run Windows because most computers are embedded in non-computery things like engine control systems or flashlights. --Wtshymanski (talk) 00:31, 20 July 2021 (UTC)

Interrupt pins

I question this sentence in the article: "When an interrupt is received, the computer's hardware automatically suspends whatever program is currently running, saves its status, and runs computer code previously associated with the interrupt." This seems to be co-mingling the Interrupt Service Routine (ISR) and the Process Control Block (PCB) together. The ISR is in hardware but the PCB is in the OS (software). The word "status" seems to mean "registers". Upon a context switch, the OS copies the registers to the PCB -- a software process. (Or maybe it done in hardware now. Please educate me.) Also, "computer code previously associated with the interrupt" sounds like the ISR. The article should be specific and differentiate between the hardware functions and the OS functions. Timhowardriley (talk) 14:45, 20 January 2022 (UTC)

That sentence is in Operating system#Components - #Kernel - #Interrupts, which is not specific to a PC architecture, and is relevant to systems without interrupt pins. In general, there is a hardware context and a software context; an interrupt causes the hardware to save the hardware context, switch to a more privileged context and transfer control to a first level interrupt handler. Code in or called by the interrupt handler may eventually switch back to the original context or may select a different process and switch to its software context.

Status may be as simple as a PC or PSW, or may include a large set of registers.

Operating system code called from the interrupt routines may switch to, e.g., a new Address Space Control Block (ASCB) and Task Control Block (TCB), a new PCB, a new page table, a new segment table, but that is well after the context switch in the interrupt. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 17:45, 20 January 2022 (UTC)

Wonderful, but I meant to be educated in the article. ;-) Timhowardriley (talk) 21:36, 20 January 2022 (UTC)

NPOV: assumption about hardware and OS

@Timhowardriley: The description of interrupt handling is incorrect for some hardware and for some software. I changed the text

Interrupts are central to operating systems as they provide an efficient way to react to the environment. Interrupts cause the central processing unit (CPU) to have a control flow change away from the currently running process.

to

Interrupts are central to most operating systems as they provide an efficient way to react to the environment. Interrupts cause the central processing unit (CPU) to have a control flow change away from the currently running process.

and

If a computer program executes a system call to perform a DMA I/O write operation, then the system call may execute the following instructions:

to

Typically, if a computer program in a Unix-like operating system on Intel x86 hardware executes a system call to perform a DMA I/O write operation, then the system call may execute the following instructions:

The list of instructions that follows is incorrect for, e.g., OS/360, Multics; in fact, it is incorrect for any OS with nonblocking writes, and for any hardware that presents data for an interrupt in a different manner, which is why I added the qualification. I also added the paragraph

For other hardware or operating systems the processing is conceptually similar but some details may differ, e.g., how the interrupt is processed by the CPU, how the OS preserves the associated data, whether the OS allows the process to continue running in parallel with the write.

Timhowardriley changed the second paragraph to

If a computer program in a computer with a direct memory access chip executes a system call to perform a DMA I/O write operation, then the system call will execute the following instructions:

and removed the paragraph

For other hardware or operating systems the processing is conceptually similar but some details may differ, e.g., how the interrupt is processed by the CPU, how the OS preserves the associated data, whether the OS allows the process to continue running in parallel with the write.

With those changes, the text is again incorrect except for specific systems. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:18, 30 January 2022 (UTC)

I think OS/360 and Multics are not much used anymore. However, if you have sourced information about how they handled interrupts, it would add to the section. As an aside, please don't accuse me of violating Wikipedia:Neutral_point_of_view. There's no bias in my edits; everything came from Tanenbaum's book. Timhowardriley (talk) 23:03, 30 January 2022 (UTC)

The issue is whether the text as written is true in general, not how old or how common the exceptions are. However, CP-67, DOS/360 and OS/360 all have current successors. I believe that since the older systems are simpler and have publicly available logic manuals, it would be better to cite them rather than their contemporary incarnations. Would anybody object to restoring the dropped text and adding these citation? Is three enough, or should I cite a few more systems?

citing an article as violating WP:NPOV is not an accusation against the editors. It is simply a claim that the text as written fails to present a balanced picture. That is an issue that can arise for several reason, and I would be surprised if editorial bias is the most common. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 01:03, 31 January 2022 (UTC)

You win. Timhowardriley (talk) 10:19, 31 January 2022 (UTC)

This user believes peripherals should not cloud the essence of the message.

This user believes Be Afraid is an unfortunate reality. "You should also learn that Wikipedia users often display ownership of articles they've edited, and jealously guard the fiefdoms and cliques they've developed..."

Sigh! I wasn't suggesting removing the scenario, just appropriately qualifying it. Also, I find it ironic that after your "please don't accuse", you chose to unjustly accuse me. "Things should be as simple as possible, but no simpler."

I asked in my previous post "Would anybody object to restoring the dropped text and adding these citation?" I ask again, would you object if I restored the scenario with a qualification and added the citations?

As a side note, would you object to adding MVS UNIX System Services (née OpenEdition) and z/VM OpenExtensions Shell and Utilities (née OpenEdition) to #Unix and Unix-like operating systems? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:15, 31 January 2022 (UTC)

Notes

References

Over-representation of 'Desktop Systems'

Ctrl-F

• Windows: 56 (minus a few ambiguous uses for ui)
• Linux: 45
• macOS: 15
• Android: 4
• iOS: 1

Xbox? Nintendo? PlayStation? All systems used by millions everyday. Notable omission imo.

Article acknowledges Android has the most market share, but is mentioned only four times, one of which is in a reference.

Bsdrevise (talk) 19:12, 12 February 2022 (UTC)

Desktop Operating System marketshare and Statistics

Would it be possible to include this?: Desktop OS marketshare I think is relevant information, as people all around the globe is subjected to follow market tendencies more or less. — Preceding unsigned comment added by Wiki Axman (talk • contribs) 18:51, 5 March 2022 (UTC)

Interrupt disclaimer

These two sentences read like a disclaimer. The details of how a computer processes an interrupt vary from architecture to architecture, and the details of how interrupt service routines behave from operating system to operating system. The scenario below is typical, although the details for some other [43][44][45][c] hardware and operating systems vary significantly. The citations, [43][44][45][c], point to IBM's System/360 specifications and are decades old. Instead of disclaiming the "scenario below", the article should articulate how IBM's System/360 handled interrupts before the direct memory access chip was available. Timhowardriley (talk) 00:21, 7 April 2022 (UTC)

Here's a good find: http://bitsavers.org/pdf/ibm/360/os/plm_1966-67/Y28-6616-1_IO_Supervisor_PLM_Apr67.pdf . Page 8 begins a documentation of the pre-DMA interrupt process. IBM called the technology "The Input/Output Supervisor". Unfortunately, the documentation is cryptic because it's filled with outdated acronyms. For example, FLIH means "First Level Interruption Handler". Nonetheless, documented are some key instructions. SVC means "Supervisor Call" and is followed by an index to an interrupt table. EXCP is a machine instruction to "Execute Copy". It talks about storing the "Status Word" and saving the the contents of the general registers. It would be interesting if an editor could articulate the I/O interrupt algorithm contained in this document for modern-day students. Timhowardriley (talk) 20:24, 7 April 2022 (UTC)

There's nothing wrong with an appropriate disclaimer. The article is about operating systms in general; OS/360 is only an example.

The cited reference describes how the IOS component of OS/360 uses interrupts but does not describe the S/360 interrupt process and assumes knowledge of other manuals.

I would encourage editors to add subsections for the interrupt architecture of a few other systems, both older and more recent. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 12:46, 11 April 2022 (UTC)

There is no execute copy. Execute channel program (EXCP) is not a machine instruction; it is an OS service. There is also an EXCP SVC and an EXCP macro instruction for invoking that service. OS/VS added EXCPVR, which uses the same code as EXCP but bypasses translating the channel program. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 17:49, 13 April 2022 (UTC)

DMA existed long before DMA chips and before S/360.

This article is about software; IBM System/360 architecture#Interruption system articulates how IBM's System/360 handled interrupts. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 12:46, 11 April 2022 (UTC)

• Regarding "There's nothing wrong with an appropriate disclaimer.": I disagree. It's a judgement saying, "The following information is being questioned." Instead of a disclaimer, add subsections for the interrupt architecture of a few other systems, both older and more recent. Timhowardriley (talk) 17:10, 11 April 2022 (UTC)
• Regarding "I would encourage editors to add subsections for the interrupt architecture of a few other systems, both older and more recent.": I agree. Timhowardriley (talk) 17:10, 11 April 2022 (UTC)
• Regarding "DMA existed long before DMA chips and before S/360.": I agree DMA existed long before DMA chips. Many technologies were implemented in software and later made faster in hardware. If S/360 had DMA, and if there's still a document describing its technology, then a subsection for the interrupt architecture for the S/360 would make an interesting read. Timhowardriley (talk) 17:10, 11 April 2022 (UTC)
• Regarding "IBM System/360 architecture#Interruption system articulates how IBM's System/360 handled interrupts": It's probably technically correct, but contains jargon I wish were better explained. For example, what's the difference between "old PSW" and "new PSW"? And why is "replaces the instruction address" qualified as "generally"? Is the "instruction address" what we now call the "program counter"? I'm curious about these things. If they were explained in a subsection for the interrupt architecture of S/360, then I would be an eager reader. Timhowardriley (talk)
• The main purpose of Interrupts is to avoid spin waiting. The technology used to achieve avoidance should be the thesis. Timhowardriley (talk) 17:58, 11 April 2022 (UTC)
  • No, a disclaimer does not imply "The following information is being questioned."; it simply limits the context of the statement.
  • Issues regarding IBM System/360 architecture#Interruption system should be addressed in talk:IBM System/360 architecture; if you start a discussion there, please ping me. The description of old and new PSW locations seems clear to me, but the next sentence is problematical.
  • Nomenclature is highly balkanized; instruction address, instruction counter, location counter and program counter have all been used.
  • The first use of interrupts was for handling errors. The use of interrupts for handling completion of I/O came later. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:19, 12 April 2022 (UTC)

References

Interrupt citations "[2][3][4][a]"

These citations point to IBM documents from the 1960s and are very long. Instead, the reader should be referred to a specific page to go to. After all, the purpose of a citation is to document the source material of secondary research. Moreover, the sentence containing the citation seems like a rewording of the prior two sentences. The prior two sentences have the complete message: The details of how a computer processes an interrupt vary from architecture to architecture. And the details of how interrupt service routines behave vary from operating system to operating system. This sentence is redundant: The scenario below is typical, although the details for some other[42][43][44][c] hardware and operating systems vary significantly. Timhowardriley (talk) 18:03, 13 April 2022 (UTC)

I've consolidated those sentences and added location data to some of the citations. Two of the citations are to directories containing multiple relevant manuals; I'm not sure what to do with them, as listing the relevant manuals might be TMI.

I added an example of something that varies from system to system; I don't believe that an exhaustive list would be appropriate.

IBM looking for a current description of IBM's object code only (OCO) policy and will add a citation once I have the URL(s). --Shmuel (Seymour J.) Metz Username:Chatul (talk) 12:54, 14 April 2022 (UTC)

I restored the tag b/c the 486M of material dated from 1967 to 1973 is not a citation. I restored the prior two sentences because they are intuitively correct. The sentence in question duplicates the message of the prior two sentences and contains original research. Timhowardriley (talk) 17:33, 14 April 2022 (UTC)

Organization of Interrupts section

I propose that #Interrupts be restructured as

1. A general prelude, mostly existing text
2. A subsection on processor handling of interrupts, with the {{main|Interrupt}} template

   Mostly existing text

   Refer reader to Interrupt for examples

3. A subsection on OS responses to interrupts
   • A general prelude, mostly existing text
   • Susubsections on a few examples
   • Explictly identify the systems to which the examples pertain --Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:40, 15 April 2022 (UTC)

     My vision for the section's expansion is to have a subsection for each technology. Interrupts currently describes only one technology -- DMA. I researching IBM's S/360 technology for another subsection. My vision has each subsection being self-contained. Timhowardriley (talk) 18:25, 15 April 2022 (UTC)

     Now that you've got me thinking, each entry in the interrupt vector table could be a subsection. Timhowardriley (talk) 18:44, 15 April 2022 (UTC)

     Another subsection I'm considering is the interrupt controller chip: https://archive.org/details/structuredcomput00tane/page/125 . Timhowardriley (talk) 21:17, 15 April 2022 (UTC)

     From the perspective of interrupts, a DMA chip and a timer chip are the same technology; they both signal an interrupt in the same fashion and the processor respnds to the signal in the same fashion. Similarly, on an IBM S/360 an I/O interrupt, an external interrupt and an SVC in interrupt are the same technology. A possible breakdown might be

     1. Interrupt architecture
        1. Burroughs B6500
        2. CDC CDC 3600
        3. DEC PDP-10/DECsystem-10/DECSYSTEM-20
        4. IBM S/360
        5. Intel 8088 including interrupt controller
     2. Interrupt handling by OS
        1. IBM OS/360
        2. PCDOS
        3. Windows NT
        4. Specific *bsd, Linux or Unix on specific processor

     I suggested these particular examples because they are markedly different from each other; many manuals are available on bitsavers and there are aftermarket books for the exceptions. I'm not sure whether these lists are too large or too small, but each should be at least 3 and no mre than 10. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:10, 17 April 2022 (UTC).

Operating systems beyond computers

I would like to see treatment of the term "operating system" to go beyond computer science. Our actions in real life are governed by a number of non-computer operating systems. These include our highway system, our money system, our governance systems, our healthcare systems, etc. See, for example: Improving our Social Operating Systems Thanks! --Lbeaumont (talk) 20:49, 19 April 2022 (UTC)

Your idea is profound. Whereas this article is about the computer context of the term, I can see how the term can be expanded. Timhowardriley (talk) 22:01, 19 April 2022 (UTC)

NPOV - Interrupt-driven I/O

@Timhowardriley: The details of interrupt-driven I/O vary from system to system. Such systems may have an interrupt per character or an interrupt per word, and may support far more than keyboards and pointing devices, e.g., tape drives.

The paragraph doesn't not contain bias and does not violate the neutral point of view requirement. It seems your concern is the paragraph is not complete. Well, the source of this paragraph is from Andrew Tanenbaum's book here: https://archive.org/details/structuredcomput00tane/page/292/mode/2up . You are free to edit the paragraph to your interpretation of this reliable source. Timhowardriley (talk) 16:09, 19 April 2022 (UTC)

The paragraph is not neutral because it makes a general claim rather than establishg a context to which it applies or using qualifiers such as some. There is more than one source, and your citation of Tannenbaum does not does not provide context. What platform does Tannenbaum describe? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:53, 20 April 2022 (UTC)

NPOV - I/O channel

@Timhowardriley: I/O channels come in a variety of forms, varying in complexity from simple character assembly/disassembly to Turing complete. While some channels are separate physical devices, others work by cycle stealing on the CPU microcode. Also, some connect directly to memory rather than using a shared bus.

The most common type of channel these days is that of the IBM System/360 and successors, which uses a simple channel program, but that is not the only type still in use. #Channel program should be split into generic and specific text, and should distinguish between "program" and "channel program". --Shmuel (Seymour J.) Metz Username:Chatul (talk) 12:37, 19 April 2022 (UTC).

The subsection doesn't not contain bias and does not violate the neutral point of view requirement. It seems your concern is the subsection is not complete. Well, the source of this subsection is from Andrew Tanenbaum's book here: https://archive.org/details/structuredcomput00tane/page/295/mode/2up . You are free to edit the subsection to your interpretation of this reliable source. Timhowardriley (talk) 16:13, 19 April 2022 (UTC)

The paragraph is not neutral because it makes a general claim rather than establishg a context to which it applies or using qualifiers such as some. There is more than one source, and your citation of Tannenbaum does not does not provide context. What platform does Tannenbaum describe? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 12:02, 20 April 2022 (UTC)

@Timhowardriley: The last edit made it worse. I don't have a copy of the Tannenbaum book, but I'm quite certain that he didn't claim To engage an I/O channel, the operating system generates a computer program and places it into the I/O channel. The steps for channel I/O are:

1. Do any relevant housekeeping.
2. Queue request if resources busy
3. Prepare the I/O buffer
4. Prepare the channel program if applicable.
5. Set any required registers or main storage locations.
6. Perform the relevant I/O instruction(s) to initiate the I/O.
7. For blocking I/O, do a context switch, otherwise return to the current process.
8. When the channel causes an interrupt, update the status of the request and, as appropriate, process other requests waiting for the channel or device. If appropriate, do a context switch.

I don't understand the rationale for Finally, the memory address containing the START I/O instruction is set into the CPU's program counter. SIO, SSCH, etc., are no different from other instructions, and normally an OS will arrive at those instruction via normally next-instruction fetches. Why not simply say something like "Perform the relevant I/O instruction" without mention of the PC? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14reference:08, 20 April 2022 (UTC)

• Regarding "The paragraph is not neutral because it makes a general claim rather than establishg a context to which it applies or using qualifiers such as some.": Then add the word some. Wikipedia:Tag bombing says, "If you can tag an article, you can also edit it in other potentially more helpful ways." Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  There are two reasons why I didn't simply edit that section:

  1. There are multiple issues.
  2. I've had too many edits reverted without discussion; given that, I prefer to discuss first.

• Regarding "There is more than one source,": What source are you referring to? If you have a source that says what Tanenbaum published is incorrect, then that would warrant a discussion. More likely, the source you are referring to describes a different implementation of the technology. Well, that would be useful information in the article. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  Among others, I'm referring to citations that you previously deleted. These should be enough to show that Tannenbaum's description does not apply to a large mainframe that uses data channels in general, and, in fact, it does not apply to any mainframe from, e.g., CDC, IBM, RCA, UNIVAC. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "citation of Tannenbaum does not does not provide context. What platform does Tannenbaum describe?": I disagree. Tanenbaum says on page 55, "Large mainframes use the design shown in Fig. 2-19. In this design, a computer system consists of a CPU (or possible multiple CPUs), a memory, and one or more specilized I/O processors called data channels." The context is a large mainframe that uses data channels. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  I'm aware of many mainframes that use data channels and do not operate in that way; I'm aware of none that do. On none of, e.g., CDC 3600, IBM 7090, IBM System/360, UNIVAC 1107,, is it possible to place a computer program into an I/O channel. In fact, many systems with I/O channels don't have channel programs at all. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "The last edit made it worse.": Wikipedia:Civility says, "[Editors] should focus on improving the encyclopedia while maintaining a pleasant editing environment by behaving politely, calmly and reasonably, even during heated debates." It also says, "Passion can be misread as aggression, so take great care to avoid the appearance of being heavy-handed or bossy." Please keep your comments respectful. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)
• Regarding "I don't have a copy of the Tannenbaum book,": Tanenbaum's book is inline. A link to the actual text is embedded in the citation. For example, you can view page 55 here: https://archive.org/details/structuredcomput00tane/page/55/mode/2up . Moreover, I placed a link to the PDF version of Tanenbaum's 6th edition on your talk page: User_talk:Chatul#Operating_system.Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  Thanks; that was in January and I'd forgotten about it. I've downloaded it for future reference. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "but I'm quite certain that he didn't claim 'To engage an I/O channel, the operating system generates a computer program and places it into the I/O channel.'": Now that you know you can witness the book, you can verify Tanenbaum wrote this on page 55: "When the CPU wants I/O performed, it loads a special program into one of the channels, and tells the channel to execute it." Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  Okay, I see it, which raises the question of which mainframe with I/O channels he was writing about. Certainly none that I'm familiar with. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "I don't understand the rationale for 'Finally, the memory address containing the START I/O instruction is set into the CPU's program counter.'": Now that you know you can witness the book, you can verify Tanenbaum wrote this on page 296: "To perform I/O on a computer with data channels, the CPU first creates a program for the channel and stores it in main memory." (Yes, this contradicts page 55 where he says it's stored in the channel. I surmise both are true). Anyway, he continues, "Then it [the CPU] executes a START I/O instruction specifying the channel and the I/O device." Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  Given the contradiction, it is clear that page 55 is incorrect. A source that contradicts itself isn't reliable.

  There is a major difference between "Finally, the memory address containing the START I/O instruction is set into the CPU's program counter." and "Then it [the CPU] executes a START I/O instruction specifying the channel and the I/O device." --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "The steps for channel I/O are: ...": These steps are original research. Whereas they may be true, Wikipedia:No original research says, "The phrase 'original research' (OR) is used on Wikipedia to refer to material—such as facts, allegations, and ideas—for which no reliable, published sources exist." You are providing no reliable source for your algorithm. Of course, we contribute our knowledge and experience to the articles we edit. For example, I'll add sentences that are intuitively correct. Nonetheless, performing secondary research is how to avoid edit wars — like this one. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)Please suggest language that is generally valid but that you don't consider to be OR.

  I've tagged and discussed here precisely to avoid edit wars and formal dispute resolution. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

• Regarding "SIO, SSCH, etc., are no different from other instructions, and normally an OS will arrive at those instruction via normally next-instruction fetches.": I don't have the knowledge necessary to digest this. Tanenbaum was being generic when he wrote "a START I/O instruction". Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  The entire reference to the program counter is extraneous. Why not simply write a short generic sentence like "The operating system then uses the relevant I/O instruction to initiate the I/O on the channel."?

• Regarding "Why not simply say something like "Perform the relevant I/O instruction" without mention of the PC?": Because that's not documented in any reliable source. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  Your description of setting the program counter is alsonot documented in any reliable source.

• Regarding this edit to the article " and ": Wikipedia:Tag bombing says, "Tag bombing is the unjustified addition of numerous tags to pages ..." It also says, "Editors who engage in tag bombing after being asked to stop may be blocked from editing Wikipedia." Please stop tag bombing this article. Your offensive tag is Wikipedia:Neutral point of view. NPOV says, "All encyclopedic content on Wikipedia must be written from a neutral point of view (NPOV), which means representing fairly, proportionately, and, as far as possible, without editorial bias, all the significant views that have been published by reliable sources on a topic." There is no editorial bias to my edits. Every sentence I submit is based upon a published and reliable source. I take this tenet seriously. Please only judiciously apply the NPOV tag. Timhowardriley (talk) 19:31, 20 April 2022 (UTC)

  I don't believe that I am tag bombing. Is it time for formal dispute resolution? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:34, 21 April 2022 (UTC)

Notes

References

Organization of Interrupts section

Given that the article is about software rather than hardware, I believe that it would be more productive to organize the #Interrupts section by OS rather than by type of interrupts. Also, the material on various types of I/O belong elsewhere, e.g., Device drivers, Disk access and file systems. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:22, 25 April 2022 (UTC)

* Regarding "more productive to organize the Interrupts section by OS rather than by type of interrupts": the organization parallels the source textbooks. Timhowardriley (talk) 03:52, 1 May 2022 (UTC)

* Regarding "the material on various types of I/O belong elsewhere": the various types of I/O are described in the context of the operating system implementing them. Timhowardriley (talk) 03:52, 1 May 2022 (UTC)

What modern CPU doesn't have an interrupt pin?

Regarding this edit https://en.wikipedia.org/w/index.php?title=Operating_system&curid=22194&diff=1085535669&oldid=1085342379 : What modern CPU doesn't have an interrupt pin? Is its market share greater than Intel's? Timhowardriley (talk) 03:42, 1 May 2022 (UTC)

Followup: The Advanced RISC Machine (ARM) (version 7) CPU has 21 interrupt pins. See the features section in this document: https://www.elprocus.com/arm7-based-lpc2148-microcontroller-pin-configuration/ It says bluntly, "Outside interrupt pins-21". Timhowardriley (talk) 05:54, 1 May 2022 (UTC)

Followup: Intel processors still have interrupt pins. This is the Xeon 3100 series specification document: https://www.intel.com/content/dam/www/public/us/en/documents/datasheets/xeon-3100-series-datasheet.pdf . It describes its APIC interrupt pin saying, "LINT[1:0] (Local APIC Interrupt) must connect the appropriate pins/lands of all APIC Bus agents. When the APIC is disabled, the LINT0 signal becomes INTR, a maskable interrupt request signal, and LINT1 becomes NMI, a nonmaskable interrupt. INTR and NMI are backward compatible with the signals of those names on the Pentium processor. Both signals are asynchronous." Timhowardriley (talk) 17:39, 1 May 2022 (UTC)

Followup: AMD processors still have interrupt pins. This is the AMD's 2004 datasheet: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.183.8059&rep=rep1&type=pdf . On page 42, it talks about its PCI interrupt pins saying, "Several internal interrupts are shared with the PCI interrupts pins. These internal interrupt signals drive the PIRQ[A,B,C,D]_L pins Low as outputs;" Timhowardriley (talk) 18:30, 1 May 2022 (UTC)

Followup: IBM's System/360 (1960s) did not have interrupt pins. However, all of these computers have been retired. (See IBM_System/360#Remaining_machines.) This is the 1968 IBM System/360 Principles of Operation: http://bitsavers.org/pdf/ibm/360/princOps/A22-6821-7_360PrincOpsDec67.pdf . On page 17 is says, "The external interruption provides the means by which the CPU responds to signals from the interruption key on the system control panel, the timer, and the external signals of the direct control feature." It obviously had things called the "interruption key", the "system control panel", and the "direct control feature". A search in this document for "interrupt pin" showed no matches. Timhowardriley (talk) 19:30, 1 May 2022 (UTC)

All of the S/360 machines have been retired, but IBM is still shipping IBM Z machines and has recently announced the z16.

The S/360 certainly had an interrupt key on the system control panel, but neither of those involved interrupt pins, nor did the external control feature. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 22:25, 1 May 2022 (UTC)