Talk:Operating system/Archive 4
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 | Archive 3 | Archive 4 | Archive 5 | Archive 6 |
Rollbacker intervention requested
With respect to the addition of "Google Chrome OS" under "Examples of operating systems", I object on these grounds: 1) The OS is already listed under "See also", List of operating systems. 2) The OS is already listed in the section as the sub-section "Unix and Unix-like operating systems" (redundant editor was being redundant). 3) It is fancrufty: I add my favorite example OS, you add your favorite example OS, all 47,906,025 registered editors add their favorite OS to the article. That's a really good idea. We should just go ahead and somehow transclude "List of operating systems" in that case. 4) The article is already too long, removals under WP:SS (not losing any content, simply moving it), not additions, should be considered good editorship of the article. 5) The subject, Chrome OS already has a whole article for itself. 6) The section "Examples of operating systems" already has consistency problems, and weight problems, this makes those problems worse, not better.
I therefor request the 12 edits it took to add the sub-section be rolled back. —Aladdin Sane (talk) 17:46, 26 November 2009 (UTC)
- I feel extremely tempted, given the recent re-addition of Chrome OS to the article, which is on the general concept of Operating systems last I heard, to pull a WP:POINT and add LEAF, my favorite OS, to the article. Is this acceptable to everyone? —Aladdin Sane (talk) 06:07, 29 November 2009 (UTC)
- OK, that's it. The article is off my watchlist. This non-notable fancruft piece of crap article ought to speedy-deleted for its complete POVish (and highly immature) boosterism. —Aladdin Sane (talk) 06:24, 29 November 2009 (UTC)
- Too easy. Lose "money". My suggestion is to simply make the article theoretical. The focus should a drum beat of OS concepts, repeated in as many ways as possible in the space allowed. This is a B-Class article because it is missing factual information. I imagine if the irrelevant POV material were removed, then there would be less shit hitting the fancruft. For example, who cares how long it took Microsoft to deliver some version of OS, when the article is missing basic factual information about OS's? — CpiralCpiral 18:21, 29 November 2009 (UTC)
- OK, that's it. The article is off my watchlist. This non-notable fancruft piece of crap article ought to speedy-deleted for its complete POVish (and highly immature) boosterism. —Aladdin Sane (talk) 06:24, 29 November 2009 (UTC)
- I think you just started the ball rolling, Aladdin. Each of those example OS sections has a main article. I request the examples section be removed entirely, and it's salient information integrated into their main articles. This will resize the OS article as per WP:SS: don't lose any information of WP. — CpiralCpiral 06:56, 29 November 2009 (UTC)
- Google Chrome OS is just one of about 500 linux distros and it's not even released yet (plan for release is winter 2010). It's very weird that it's listed here.. --ramvi (talk) 13:34, 3 December 2009 (UTC)
- I removed it. — CpiralCpiral 06:43, 11 December 2009 (UTC)
- I see that. Thanks. Doing WP:SS work is difficult and time consuming I note.
- Now that Commodore has its own section, is it OK if I add sections for Sinclair QDOS, OS/2, Yggdrasil, Amiga (several subsections under Commodore now need to be added I suspect), TRS-DOS, PascalDOS, and Be OS? (Please excuse the sarcasm, it's late.) —Aladdin Sane (talk) 07:43, 11 December 2009 (UTC)
- Sounds okay by me, but don't jump to the assumption I know anything at all about PCDOS, or what the difference is/was between Sinclair QDOS and TRS-DOS. There may be a fine and subtle difference, and whatever it may have been, if the article is to benefit readers in general, it should not leave them in a state of confusion. Dexter Nextnumber (talk) 22:46, 20 December 2009 (UTC)
- Come to think of it, there isn't a single mention of game computers like Nintendo and Sega having operating systems. Is this because a contract of nondisclosure prevents developers from mentioning their operating systems? Dexter Nextnumber (talk) 22:49, 22 December 2009 (UTC)
Devoting an entire section to each notable OS
Devoting entire an section to any one, notable, production OS is problematic.
- Each branch of the outline tree better be a conceptual category, not an implementation fruit.
- The size of the article is already twice the limit; these sections can only increase in the future, and notability discussions will then be required.
- An OS is usually personal to an editor, therefore a "representative" often tempts bias. Neutrality and objectivity slip. Troll call. Vandals become even more careless.
The problem sections are 3.x and 2.10.x.
Removing the OS product sections will improve the English Wikipedia.
- It would create space in the OS article for more essential information related to OS in concept and theory. This is a B-Class article, which means it's missing essential information.
- Each of the product sections already has a main article. There is some unnecessary duplication.
- Each of the main articles for these product sections is missing some of this information.
- Many more editors will be able to edit the entire article, as is required for restructuring. Many computers in use today are prevented from editing the article all at once because the size of the browser edit buffer is limited to about 30Kb.
I am going to make sure, one product section at a time,
that the main articles have the information.
The idea is to first duplicate, then delete the original information.
The check mark means that the information in that section has been carefully integrated elsewhere,
either into a more relevant article or possibly within Operating system itself.
- 2.9.1 Example: Microsoft Windows (cloned at Microsoft_Windows#Security)
- 2.9.2 Example: Linux/Unix (went to improve the intro at category:Linux security software.)
- 2.10.1 Mac OS X (Mac_OS_X_Server#Technical_Specifications)
- 2.10.2 Solaris (was already a verbatim copy (of an older version) of File_system#File_systems_under_Solaris.)
- 2.10.3 Linux (improved intro paragraph to Comparison_of_Linux_distributions#Technical.)
- 2.10.4 Microsoft Windows (improved intro to Comparison_of_Microsoft_Windows_versions#Supported_file_systems.)
- 3.1 Microsoft Windows (OS) (improved Comparison_of_Microsoft_Windows_versions#Programs and Windows history.)
- 3.2 Mac OS X (improved History of Mac OS X and Mac OS X Server).
- 3.3 GNU/Linux and Unix-like operating systems
- 3.3.1 Google Chrome See Google Chrome OS
- 3.5 Plan 9 (donated carefully to Plan 9 from Bell Labs and Inferno (operating system))
- What links here has been redirected when check-marked: Scanning and selective sampling found no links to subsections of this article.
This may cause temporary bloat, but the article has been way oversized for well over a year. When these are all checked-off as completed, these sections can safely be deleted. Per WP:SS this is the way to improve Wikipedia.
This plan will reduce this 60Kb article by about 30% to 40Kb. The goal is 30Kb of concept and theory, using product implementations only as necessary for clarification of OS concepts and theory. The next class level above B is GA. If you need any more motivation, look at the other countries versions. We are not #1.
As this is my stated plan to fix the un-contestable, undebatable problem of size and quality issues, it is not so much a discussion, as a long awaited announcement. As such, the above should be edited for improvement of my writing, but there is (in my opinion) no reason to vote it down as if some plan, any plan, was an option.
— CpiralCpiral 21:03, 30 November 2009 (UTC)
- OK, you drew me back in.
- I read your arguments here and in the section above. Almost to a point, I totally agree. I note that WP:SS, is "always do-able, but always tricky". Because we care about not losing information, but we also care about not adding redundant info to the target article while moving, WP:SS requires expertise with both articles.
- There is one point you made (in the above section) that I wish to clarify or contend with. I don't think the article should should be theoretical, I think it should rapidly descend to the theoretical. The reason I have have a problem with "strictly theoretical" is because these articles are written for the reader with "little or knowledge of the subject". That reader needs some practical basis up front to attach the theory to. This is the article I refer people to when I ask them, "What OS are you running?" and the response I get back is "I run Word". Please keep that person in mind when editing this article.
- Good point. — CpiralCpiral 04:33, 1 December 2009 (UTC)
- Back in the murky past, some editor thought it would be a good idea to give a whole section over to "Examples of operating systems" (currently at § 3). That was a bad idea, in hindsight. Nonetheless I think we can leave that editor's section in, for historical referencing, and once the work is done to move info to the right articles, either the 'See also' or 'Main' template should be used to make the whole section look like this:
- Examples of operating systems
- I can see some editors disagreeing, however, I personally find the length arguments totally compelling for losing the info from this article, in favor of other articles. "It's a good reason to move the info." —Aladdin Sane (talk) 21:59, 30 November 2009 (UTC)
- Good one, Cpiral. I hadn't thought about the TOC implications of my markup. (Other editors can know that demo'ing the <h1> tag ain't always the greatest idea, since it makes the TOC "highly misleading" even if you didn't mean it to be (see page history for an example).) —Aladdin Sane (talk) 00:18, 3 December 2009 (UTC)
- Thanks! — CpiralCpiral 02:18, 4 December 2009 (UTC)
- Good one, Cpiral. I hadn't thought about the TOC implications of my markup. (Other editors can know that demo'ing the <h1> tag ain't always the greatest idea, since it makes the TOC "highly misleading" even if you didn't mean it to be (see page history for an example).) —Aladdin Sane (talk) 00:18, 3 December 2009 (UTC)
The implications of this edit scare me
This edit by User:Dexter Nextnumber fundamentally changes the nature of the article, and is far from minor. Now, we get to pull out everything that is there, and talk about Symbian and Ford engines instead, or add another 10 pages to the article about those things. Plus, it is unsourced, and asserts a quantitative measurement. It may rise to the level of 'mere opinion' without a source. I suggest it be removed, to make our lives easier defining the topic and scope of the article. —Aladdin Sane (talk) 01:03, 3 December 2009 (UTC)
- Firstly, the edit doesn't change the whole article, just improves on the accuracy of the one paragraph cited. And it didn't come as a surprise because the gist of the edit had already (a couple days earlier) appeared in the Talk page associated with this page. It appeared that nobody objected, so the edit is merely an act of boldness, something that is generally approved of in Wikipedia. I wasn't playing favorites by emphasizing one automotive engine brand over any other, so this is not a case for nonneutral pov. I think it's pretty easy observing how many modern cars have advanced microcontrollers in them, and their control is distributed among their parts. Hate to trample on your holy grail (perhaps because you appear to be an advocate of some kind of high level language like Unix or Linux, I think, and there's nothing wrong with that, as it gives you something to live for), but there are so many cars on today's streets that a quantitative observation (not necessarily a measurement) really isn't necessary. What kind of measurement would you go for, to give the current edit satisfactory weight for inclusion in the article?
- Secondly, most of that article was really boring for its lack of notability. This small edit at least puts everything into proper perspective. There are a heck of lot cars out there, and all or most of them nowadays have complex microcomputers or electronic platforms in them. Sure, rail away about how unfair it is for the source code to be hidden, but they are examples of operating systems nonetheless.
- Thirdly, if you want to talk about computer systems used in the home or office, or the systems used for controlling washing machines or home heating and air conditioning systems, maybe something along those lines could be slapped together, too, and talked about (to your heart's content) in a separate article somewhere, and then we could link to it. Dexter Nextnumber (talk) 02:20, 3 December 2009 (UTC)
- Now let us review the fundamentals around the 4th sentence of the lead paragraph, and above all, have fun. WP:WNI. — CpiralCpiral 04:32, 3 December 2009 (UTC)
Removing
subsections: Security
Per the proposal above I'm going to remove all the subsections sections under Security section:
- Operating_system#Example:_Microsoft_Windows is now at Microsoft_Windows#Security.
- Operating_system#Example: Linux/Unix is now at Category:Linux security software.
This will happen in a couple of days from now, depending on (and requested) feedback. Thanks. — CpiralCpiral 00:50, 5 December 2009 (UTC)
Done. The size of the article went from 69 to 65kb. — CpiralCpiral 06:04, 11 December 2009 (UTC)
subsections: File system support in modern operating systems
Per the proposal above I'm going to remove all these subsections :
- 2.10.1 Mac OS X (Mac_OS_X_Server#Technical_Specifications)
- 2.10.2 Solaris (was already a verbatim copy (of an older version) of File_system#File_systems_under_Solaris.)
- 2.10.3 Linux (improved intro paragraph to Comparison_of_Linux_distributions#Technical.)
- 2.10.4 Microsoft Windows (improved intro to Comparison_of_Microsoft_Windows_versions#Supported_file_systems.)
This will happen in a couple of days from now, depending on feedback, and pending acceptance in their new homes. — CpiralCpiral 20:26, 7 December 2009 (UTC)
- Wait! Since I don't have high speed internet, it takes me a lot longer to wade through some of that stuff (maybe all of that stuff) than it does you. Could you hold back on deleting the section about Mac OS X? Maybe for a few weeks, or until Jan 1, 2010? I think the Mac OS X stuff might be interesting to new users, especially new owners. Dexter Nextnumber (talk) 21:18, 7 December 2009 (UTC)
- Surely. (2010!!) But would a link from the usual position of Mac OS X, to the article to where it was moved be OK for you? I'm assuming you mean the Examples of Operating Systems subsection Mac OS X, and that by "stuff" you mean doing a search and all that browsing to get back settled in, and that you're getting something special for Christmas. (The section slated for removal is only two sentences.) — CpiralCpiral 00:20, 8 December 2009 (UTC)
- Wait! Since I don't have high speed internet, it takes me a lot longer to wade through some of that stuff (maybe all of that stuff) than it does you. Could you hold back on deleting the section about Mac OS X? Maybe for a few weeks, or until Jan 1, 2010? I think the Mac OS X stuff might be interesting to new users, especially new owners. Dexter Nextnumber (talk) 21:18, 7 December 2009 (UTC)
- I'm fine with moving the information (rarely OK with removing it). The condition is always that the proper {{See also}}, {{Main}}, or related template is added, so that the reader can get to the proper article that now has the info they seek (I note this speeds up, not slows down, the readers ability to read the encyclopedia). Per WP:SS I think your edits will do a great job of maintaining the necessary balance between articles in an an encyclopedia of 6,877,196 articles (including 23 Main articles (not including spin-offs not so titled) on the subject of Mac OS X), Cpiral. The article is titled "Operating system" and I feel must stay focused on that subject, for the article's benefit to the reader of it. —Aladdin Sane (talk) 23:18, 7 December 2009 (UTC)
- Thanks. We could use a {{Main}} template under File system support in modern operating systems (after renaming it Files systimn), linking it to the main article File system. Oops, it has the same problem as we are working on! The conversation there suggests we make a table out of File system support in modern operating systems there. — CpiralCpiral 00:20, 8 December 2009 (UTC)
section: Hobby development
I propose we use remove Hobby development as a subject. The existing Operating system development articulation suffices, and we could simply integrate all the information here into it, and make this B-class article (missing factual, operating system information) more focused, and that hungry article fuller. — CpiralCpiral 06:38, 11 December 2009 (UTC)
section: Commodore
The Commodore section might better be read from computer networks#Functional relationship (network architecture). It is kinda murky, to me, in its linguistic logic, so it's partially unmovable unless citations are added, as I've requested there now. An operating system section such as Commodore might live happily at network operating system. I'd need the cite to have an opinion. But at first glance it sounds like a regular OS with a fancy network protocol, so it's slated as stated. — CpiralCpiral 02:49, 12 December 2009 (UTC)
- If it is removed, maybe you should adjust all of the other sections (above and below) to explain to Commodore users & programmers why all those other computers are unable to TALK/LISTEN, and why files are simply not written directly to Tracks and Sectors like Commodore files are (or can be). Also, if it is moved, somebody might come along and not understand why, and then (out of simple naivete or youthful innocence) add a new section about the operating system (essentially Commodore DOS relying on the TALK/LISTEN protocol common to Commodore devices) - basically repeating anew what is currently already there. True, there need to be lots of citations, and they are currently missing. But considering how many years into the past the protocol is buried, looking for citations could involve scrutinizing a lot of ancient newsletters, pamphlets, and booklets. Which some people might find fun. Dexter Nextnumber (talk) 08:00, 17 December 2009 (UTC)
- WP explains to all people, not just OS folks. All readers need clarity. If you clarify the writing, listen, it will talk to all. If you get me those citations I will be happy to edit our beloved Commodore. This B-class article is missing factual data about operating systems. By eliminating product-named sections (like Commodore) the space allocated to the OS page will then be made available for the immigration of missing data. Now where would you put the parts of Commodore? (The next posting I plan to make is the improved outline/structure/section-naming.) — CpiralCpiral 18:41, 17 December 2009 (UTC)
- How about an entire article on the "Commodore" operating system (as though there were only one), and then link to it? But as you have said before, and which I agree with, there should be more citations so curious individuals can scour internet and doublecheck them? As it is, I am still trying to clean up the article on PETASCII, which is currently doing more harm than good. (For one thing, the uppercase characters A-Z all have bits 6 and 7 set, unlike all other computers in the 1980s.) Dexter Nextnumber (talk) 19:47, 17 December 2009 (UTC)
- That would be ideal. All operating systems you never heard of have there own article already. Try intitle:operating system or {search link|commodore intitle:operating system}} — CpiralCpiral 07:29, 22 December 2009 (UTC)
- How about an entire article on the "Commodore" operating system (as though there were only one), and then link to it? But as you have said before, and which I agree with, there should be more citations so curious individuals can scour internet and doublecheck them? As it is, I am still trying to clean up the article on PETASCII, which is currently doing more harm than good. (For one thing, the uppercase characters A-Z all have bits 6 and 7 set, unlike all other computers in the 1980s.) Dexter Nextnumber (talk) 19:47, 17 December 2009 (UTC)
- Thanks, that is a great way of searching for an argument. Dexter Nextnumber (talk) 22:57, 22 December 2009 (UTC)
section: Example operating systems
In a few days, I will remove the entire section and it's subsections. Please see the above discussion talk:operating system#Devoting an entire section to each notable OS, and the discussions after it.
This is in an effort to focus on what an operating system is, rather than what operating systems there are, their history, etc. The tone set by the first sentence of this article is the tone this article will eventually consist of, entirely consistently. It says "An operating system is...", but by the time we get down to the topic of file systems, the tone has changed to "Support for file systems is...", and then drops to a bass-note of a section devoted largely to subsection after subsection of "<product> is..."
English editors, sharpen your pencils! What is an operating system, pray tell? — CpiralCpiral 03:09, 13 December 2009 (UTC)
remaining File system support in modern operating systems subsections
I've copied the information from the sections Special-purpose file systems and Journalized file systems to improve the articles Journaling file system , file system#Disk file systems, and File Allocation Table. — CpiralCpiral 22:40, 14 December 2009 (UTC)
File system section proposal
The first phase of remaking the section File system support in modern operating systems was to make sure all the information is safely integrated in other articles on the wiki. That's done. The above sections list the information-adopting articles.
It is time for proposals for the section. This is a very rough draft I threw together from the existing text for the wiki to edit.
This does not have to be single section. We don't have to WP:summary style. What about sections titled to organize some of these topics which are not covered here, but that deserve mention:
- abstractions at the application and user levels Done
- levels (logical file system, file system organizational modules, basic file system, file system I/O control, device drivers) Done
- operations on files (mapping, opening, closing) Done
- access methods (sequential, direct, random, mutex)
- directory subsystems (their structures, and how they implement organizational modules search, create, delete, list, backup)
- file protections (rwxrwxrwx)
- allocation methods (fragmentation, block transfers, boot blocks and superblocks) Done
- implementation issues and how various products handled them
- /proc
- volumes (logical, mounting) Doing...
- database-like, record-based file systems used with VM/ESA, OS/400
These are some file system related concepts. Got any more? (No NAS please. Like a memory hierarchy, we have to draw the line at network area storage.)
File system (edit me pleese)
An operating system implements, in its file system, the abstraction of a computer file. The code comprising a file system manages the operations of retrieval (opening the file), data manipulation (reading, writing, encrypting), and storing. To perform these tasks, the file must be mapped, or abstracted, between levels.
An operating system installs it's boot sequences into the booting area defined by a file system. When the machine is activated, the operating system then bootstraps and could then begin to read regular files to complete its initialization. In order to install it must "know", or "support" the very media format or disk format it will later from. A running operating system, coded to particular specifications, will support file operations on other types of file systems. The NT kernel only installs onto the NTFS file system. This means it only boots on NTFS.
Let's see how the physical location of the ones and zeros on the disk gets to the display through various levels of the file system. Each level will "support", or "know" a data scheme. In other words each level's design expects a format, a sort of protocol, a standard structure of information. Assume a disk storage device at the hardware level.
- Disk format: device driver software expects (or can create) partition and sector marks from storage device hardware. The computing device hardware is set with the location where a boot loader is expected.
- File system format: Operating system boot block expects a root directory structure.
- Content format: application expects the data structures defined in its program code.
The application gets the application files starting address from the directory. In a fragment-oriented file system, the first fragment will contain the address to the second, and so on, in a linked list. Once all the data is read, it is sent to the graphics system for display.
It is not necessary for an operating system to support file formats. The interface between the content format and the file system format: If the operating system kernel supports file formats the kernel is the interface. The NT kernel supports file formats.
If the kernel supports a virtual file system switch (VFS), the application, together with the VFS abstraction is the interface. Linux supplies the VFS environment, where support for a file format is relegated be managed by the computer application program.
The operating system uses the file system to maintain the storage media. A later operating system or application might not support an earlier file format or disk format. There are issues concerning the future ability to maintain the usability of file formats[citation needed] and to maintain the integrity of the media, as media sizes grow in size[citation needed]. File system checking with Fsck might take too long, for example. Journaling file systems address some of these issues by adding a metadata journal that constantly checks and silently corrects the file system. Soft updates is an alternative to journaling that avoids the redundant writes by carefully ordering the update operations. Log-structured file systems and ZFS also differ from traditional journaled file systems in that they avoid inconsistencies by always writing new copies of the data, eschewing in-place updates.
— CpiralCpiral 00:34, 15 December 2009 (UTC)
- No, because CBM disk drives do not have a "booting area". Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- That's the best we can do here. We need to articulate the concept of booting, and the definition or form of a separating entity called "the OS". Perhaps you know something important we need to articulate as well, or you can explain how Commodore boots and reads files. Better yet, cite something, please. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- Neither the Commodore operating system nor the Commodore file system has a "boot block". Dexter Nextnumber (talk) 08:05, 17 December 2009 (UTC)
- I see the point. Not only Commodore, but as I understand it, many mini or mainframe systems don't have the concept of "boot block". I would expand the argument against by putting it in these terms: file system is a noun, an actual thing. Booting is a verb, an action. The two probably shouldn't meet and have lunch in a section that purports to be about file systems. It may be better to talk about what it is, rather than what it does (or might do in a given case), in a summary section. Hopefully, these things can be expanded on (the "does" things as opposed to the "is" things) accurately in the main article on the subject. —Aladdin Sane (talk) 19:23, 17 December 2009 (UTC)
- Right. Many such issues as "is" vs. "does" and "PC" vs. "Embedded" to iron out. — CpiralCpiral 09:39, 20 December 2009 (UTC)
- No, because CBM disk drives do not have a "booting area". Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- OK Changed many mentions of "boot block" — CpiralCpiral 09:39, 20 December 2009 (UTC)
- No, the file does not have to be mapped before being transfered. With CBM disk drives, it is sufficient to locate the track & sector a file is located, and then move the read/write head to the proper place, and start reading & transmitting. The 'opening process' does not map the file, it merely makes sure that the read/write head is in the right place. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- Does Commodore have an API? The OS we need to cover is the general OS, the one with the most identifiable parts. The application programmer will write a kernel of an OS in C, and it will offer services to the remaining part of the OS. Each service is an abstraction, or interpretive layer with an API. The function called open takes a few parameters specified by the API, and then it implements the expression "open" in machine language code, as per the compiler. But the programmer does not have to place the heads of the device driver, so the programmer experiences a blessed abstraction. All the programmer does is express a request for the OS to do something. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- I see a problem here with word usage, language, and computer development. I don't have a common usage reference for "API" going back before about the late 80's-early 90's. This does not mean the concept did not exist, simply that it wasn't called that. We may have used it without having a name for it. The concept predates the term, in computer development history. A parallel example is CLI: The term "Command Line Interface" did not come in to my usage until GUI's became common enough to need distinguishing terms. I learned the CLI in the 70's; but I did not call it that then. —Aladdin Sane (talk) 04:39, 21 December 2009 (UTC)
- Object orientation, modularization, functional and data abstraction, these are the concepts we need to write with, and the concept of API has become not only indispensable to the sources we must report on, even if we don't use the acronym itself, but gloriously enlightening as to what the computing phenomena might be. An operating system's entire purpose is to function as API to the machine resources; OS as API makes it fun and easy to handle software modules via management packages with communication interfaces used for protections and synchronizations. The new acronyms are acrimonious at first, but they are all about bringing OS to the common terminology and easy understanding of what now needs for us editors to reflect as the shiny new, formal surface of a former grey fog. Interface is the I in CLI and API and is now the epitome of software computing in general. Interface is the way things get done efficiently and safely by the programmer people, the foundation of the future of computing. API is critical to standards, as now "software as an object in a repository with a standard interface" is the programmer's ordinary experience. Another case, closer to our point: the French plaintiff (below) says The Article makes assumptions without logic and does not reflect operating system science and has an English defendant arguing "...interface, whatever it is, in many cases happens to be tightly interwoven with the OS in rom. So tightly interwoven that it's hard to discern where one starts, and the other begins." while the French operating system is already under consideration to progress from B class to "good article"). (Argh!) — CpiralCpiral 19:55, 21 December 2009 (UTC)
- I see a problem here with word usage, language, and computer development. I don't have a common usage reference for "API" going back before about the late 80's-early 90's. This does not mean the concept did not exist, simply that it wasn't called that. We may have used it without having a name for it. The concept predates the term, in computer development history. A parallel example is CLI: The term "Command Line Interface" did not come in to my usage until GUI's became common enough to need distinguishing terms. I learned the CLI in the 70's; but I did not call it that then. —Aladdin Sane (talk) 04:39, 21 December 2009 (UTC)
- Does Commodore have an API? The OS we need to cover is the general OS, the one with the most identifiable parts. The application programmer will write a kernel of an OS in C, and it will offer services to the remaining part of the OS. Each service is an abstraction, or interpretive layer with an API. The function called open takes a few parameters specified by the API, and then it implements the expression "open" in machine language code, as per the compiler. But the programmer does not have to place the heads of the device driver, so the programmer experiences a blessed abstraction. All the programmer does is express a request for the OS to do something. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- No, the file does not have to be mapped before being transfered. With CBM disk drives, it is sufficient to locate the track & sector a file is located, and then move the read/write head to the proper place, and start reading & transmitting. The 'opening process' does not map the file, it merely makes sure that the read/write head is in the right place. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- Until I came here, I'd never even heard of the term 'API', and I consider myself a fairly competent assembly language programmer. But from the way you describe it, it appears to be nothing more than a fairly ordinary program with CBM Kernal access privileges. But then again, didn't all programs have direct "Kernal" access back then? I don't think anybody even used the word API in the early 1980s. All microcomputers back then were by default single-user systems, and there was no reason to insulate the ROM from direct access by a fairly ordinary, conventional, user-written program. Dexter Nextnumber (talk) 02:26, 22 December 2009 (UTC)
- But there are some file formats it does not support. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- Let me try to understand what you have stated: It currently says It is not necessary for an OS to support file formats, but there are some file formats is does not support? The set is the set of file system formats. The operator is the kernel or it is the app. What precisely is are your associations between the members of the set and the operators on that set? In order to help be constructive, please try to be really specific about the issues you will bring up, about your indications of how things can be improved for the broad generality of our file system section. This section must post with no Commodore wizardry. Thank you Nextnumber. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- After glancing over file:Cbm_kernal_jmp_table3.jpeg, Commodore seems like an excellent example of embedded OS. — CpiralCpiral 19:55, 21 December 2009 (UTC)
- Let me try to understand what you have stated: It currently says It is not necessary for an OS to support file formats, but there are some file formats is does not support? The set is the set of file system formats. The operator is the kernel or it is the app. What precisely is are your associations between the members of the set and the operators on that set? In order to help be constructive, please try to be really specific about the issues you will bring up, about your indications of how things can be improved for the broad generality of our file system section. This section must post with no Commodore wizardry. Thank you Nextnumber. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- But there are some file formats it does not support. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- Most CBM operating systems have a combination of vector tables (located in low and high ranges of memory) and jump tables (mostly in high ranges of memory) that the programmer is recommended to observe, but never forced to observe. Yes, you could store a lot of them on a single Flashrom.
The NT kernel only installs onto the NTFS file system. This means it only boots on NTFS.
- There is a COI and POV problem here. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- I don't think you are correct here. COI violation would assume I'm an NT advocate. POV would assume I needed to say that NT booted to other FS's. Wrong and wrong. That is all. — CpiralCpiral 04:17, 21 December 2009 (UTC)
- I detect precisely zero COI or POV. That style POV/COI means that people from New Mexico can't edit articles about New Mexico.
- I do detect what I thought was misstatement of fact. I thought the NT kernel based OSes were perfectly happy living on a FAT file system. I could be confused; it's been over ten years since I payed that much attention to NT kernel based OSes. —Aladdin Sane (talk) 04:29, 21 December 2009 (UTC)
- There is a COI and POV problem here. Dexter Nextnumber (talk) 23:07, 20 December 2009 (UTC)
- It seems that both of you advocate a FAT system (including by implication FAT12, FAT16, and FAT32 file systems), which has been the subject of heated litigation, as in Microsoft v. TomTom over access to FAT32, or bundling rights. Just because they settled out of court doesn't mean the issue is disposed of. On the other hand, if we are talking out of our hat about pure theory, and what is theoretically possible, there are many other ways of organizing file systems. Myself, I fancy a BAM with pointers to specific tracks & sectors (or even, say, pointers to cylinders & volumes) where the directories can be physically located, and so on. On the other hand, if we are talking about the consequences of monopolistic competition where the only thing that exists nowadays (and therefore very easily cited) are FAT based file systems, then it makes perfect sense to conclude a priori that all file systems have and must have FATs (even though there are many other ways of doing things the same way differently). Dexter Nextnumber (talk) 02:26, 22 December 2009 (UTC)
- I am still trying to wade my way through the articles on the Mac. It's just that I don't have high speed internet (and probably never will), and it takes a while to get through all of it. I am assuming that OS X on the Mac uses some kind of a FAT32, and if it doesn't, it is probably because I haven't made my way over to the right article yet, to clear up my confusion. It would be great if somebody could help me by pointing me to the right article on the Mac filesystem. It is a multi-user system, right? Why somebody would want somebody else to have access to their computer is beyond me. Dexter Nextnumber (talk) 02:42, 22 December 2009 (UTC)
Why don't we just relax and use the Disk access and file systems section we already have? — CpiralCpiral 03:58, 22 December 2009 (UTC)
The Article makes assumptions without logic and does not reflect operating system science
The Article now says that the user interacts with the operating system. That the operating system is between hardware and the user. That is against the operating system science what explains that the operating system is between the hardware and all other software in the software system. French say it nicely, with rough translation, OS is "between hard and soft". The user interface what is made with software, does not belong to the operating system at all.
The article itself now says that user interface is part of the operating system. That is not typical and only few operating systems does so. One example is NT what includes the windowmanager itself in the operating system. Example the Linux, it does not offer user interface and user needs other system software to interact with, like the GNU bash.
The article needs fixing about that. It is not usual that the OS has user interface, just support for such coded to them. Typical OS (like Unix or Unix clones) has interfaces but it is not for the user but for software. The only interface what OS offers are called as system calls and they are API's or ABI's. It is crucial to understand that such user interfaces does not exist in the operating systems, but they are offered by other software what works top of the operating system. It is important to fix this because it makes lots of assumptions for most OS's by only few OS's what does so and they are only few, even when one of them is most used OS by market share. The user always interact through other software to the OS what then executes all on the hardware and pass answers back to other software. User does not use the OS but the programs. 80.248.105.131 (talk) 05:10, 18 December 2009 (UTC)
- Despite expressions in English that are difficult to understand, I'm getting the message expressed here loud and clear.
- The OS and the UI are in all cases separable concepts. They are not in all cases separable or separated in practice. Despite the distinction being obvious to one "skilled in the art", explaining this to this reader with "little or no knowledge of the subject" (our target audience), is difficult, tricky, and way beyond my skill level for the article.
- It strikes me that part of the problem is this: Many editors here are at the level of "expert" in two or more OSes, however no editor here, on this planet anywhere, is at the level of "expert" on all OSes. This limitation makes it difficult to even set benchmarks or standards for the article, much less write the prose that all editors can agree on.
- The problem really is that most people use Windows (any version) or Windows + one Linux distribution (usually Ubuntu now) or Mac OS X. They do not actually understand anything about OS about using those. To understand what the OS is and how it works, is to demand that they understand the differences of software library and program. Binary and source code. Hardware and software. Hardware and machine language. They should as well know the history of the operating system development and how they have been invented and to solve what problems. There is lots of areas what should be know until someone could say "I know this OS". But even then, they could not really understand OS. But it is much easier when user takes a book of modern operating systems and starts reading from there. 99% of it would be technical non-sense for most of the normal people, about how processes gets created, how they are executed and so on. How the memory is used and reserved and cleared. But there are few things what they can learn. Operating Systems has architectures. There is no one architecture for the OS. There are multiple. Every architecture has strength and weakness and question to choose a architecture limits many features away what other architectures has. There is no screenshots of the OS, there is no pictures of the OS. There are only illustrations of very simple way how the OS could work. But not even one illustration can tell it from all architectures as every architecture is different. And every architecture can be coded with different ideas (You can even code the videogame to be part of the OS but it would just be very very stupid. And that does not mean the videogame came preinstalled). People should first knowledge what are OS architectures as monolithic, server-client, layered. Those are the three main OS architectures what OS can use. No screenshots, no images. Golftheman (talk) 07:36, 31 August 2010 (UTC)
- Nevertheless, other areas of the article are being worked on (see above), and the article is being actively improved. —Aladdin Sane (talk) 06:14, 18 December 2009 (UTC)
- I agree that an OS comes between the hardware and the software, but lots of software makes direct system calls as much a matter of necessity, as convenience. Few programmers decide it is high time to rewrite entire subroutines in OS (although, of course, that sometimes happens). The user interface, whatever it is, in many cases happens to be tightly interwoven with the OS in rom. So tightly interwoven that it's hard to discern where one starts, and the other begins.Dexter Nextnumber (talk) 22:42, 20 December 2009 (UTC)
- You can not use one OS as example to proof that all other OS's are like that. And that software is stored to ROM chip is not anyway a explanation or even a proof what OS has. Example, you can have operating system with Server-Client architecture. The operating system A and operating system B both use Server-Client architecture. The other is in embedded systems and other is in normal personal computer of today. They both can be totally same way done, there is no reason why not. But we need different functions and features, so we code and design the OS to be such to offer them. So even that A and B have Server-Client architecture. We can code that other has the graphical user interface coded to it and the other does not have anykind user interface (as it is most times on personal computer OS's!). Even that A and B has same OS architecture, we can code both to in them different functions or leave them away from the OS as normal processes. There is reason why OS should not have anything than most important functions to control the hardware and operate the other software. The one thing what people mistakes is that OS's does not offer user interface, but interface for the user (to the coder. API's and ABI's. Called system calls). Example Bash and Xorg. Both offer a UI. First one a CLI and other one a GUI. But neither of them are part of the Linux OS. Both of them are normal software what just have specific task to offer a UI. You can swap the OS under them but you do not notice it directly any way as the UI (Bash and Xorg) have staid same. What you notice, is that the hardware compatibility has changed, possible the speed, stableness, security, available filesystems and so on. Example, most people do not know that Linux kernel is the operating system itself. Linux is monolithic kernel, not microkernel. Monolithic is the first and the original architecture for the OS. And at that time when Multics and Unix were made, kernel was the synonym for the operating system. The software had many names like: Core, Nucleus, Kernel, Supervisor, Master Program, Operating System and so on. The kernel was the synonym for the software what we today call operating system. The OS was monolithic and they were very simple to code. But later there came new architectures like Server-Client what demanded to take other name in use. The operating system came most popular. Monolithic kernel is still the OS, but in Server-Client architectured operating system, the microkernel (what is many times called as well just kernel) is just part of the operating system while servers (modules) are the other OS functions (device drivers, filesystems, networking, memory management etc). Two totally different architectures, with same features but other is just a kernel while other is (micro)kernel + servers. Top of the operating system there is running a user interface. The Unix idea is not to make the UI part of the OS. That is like in Linux, HURD, FreeBSD, NetBSD, OpenBSD, Minix and so on. The UI is not part of the OS but runs as normal program on it. But some people have made changes (Microsoft) that they replaced the OS functions with the functions what offered as well the graphical user interface and so on. They used same old Server-Client architecture but they developed it with different manner. They included functions to same software components what included the OS functions. Telling what is the OS is not as easy as saying just "between hardware and software" as that is very very very wide definition. It all depends every OS itself what belongs to it and what does not. We can not say by architecture (this does not belong to this architecture) or comparing two OS with same architecture. If we want to explain the Linux OS, we need to explain just that. We can not use it to explain SunOS or FreeBSD (what btw are monolithic (kernel) operating systems as well) as they have different functions and features and other is even proprietary so we can not see the source code what is there. Difficult to read but explains well, pretty boring but accurate about Minix and Server-Client idea vs Monolithic, the basic logic in normal user format and last but not least the very simple comparision of the different OS architectures what actually sums all (but demands still to understand the wider picture). This article is currently totally wrong as it say that the UI (user interface) is part of the OS. While it is not true at all. User interface is not part of the operating system at all. It all depends about the architecture and the implentation of that architecture. Example Linux does not have any user interface, but it can print text and it has the support for graphical user interface (about 11, 000 lines of code) in it. As other example, the user interface is not just what is on screen. It is the devices as well, the keyboard, mouse and so on. Even the speakers are user interface. The hardware is user interface as well. And we can use Linux in embedded systems as well, where the UI can be just the hardware, without any screen. In elevators we can use physical buttons or LED's to give a interface for the user or we could make a touch screen what would as well give more information about the the services what are in every floor. There are so many different ways to offer a UI, but none of them are part of the Linux OS. Golftheman (talk) 07:36, 31 August 2010 (UTC)
Proposing an outline
I think we are having a major makeover. Removing a third of the sections might be an opportunity to rename remaining sections in such a way as narrow the topic, while covering it more fully. At other times renaming section titles could more controversial than now.
Here is the current structure:
1 History 1.1 Mainframe 1.2 Microcomputers 2 Features 2.1 Program execution 2.2 Interrupts 2.3 Protected mode and supervisor mode 2.4 Memory management 2.4.1 Virtual memory 2.5 Multitasking 2.5.1 Kernel preemption 2.6 Disk access and file systems 2.7 Device drivers 2.8 Networking 2.9 Security 2.10 File system support in modern operating systems 2.10.1 Mac OS X 2.10.2 Solaris 2.10.3 Linux 2.10.4 Microsoft Windows 2.10.5 Special-purpose file systems 2.10.6 Journalized file systems 2.11 Graphical user interfaces 3 Examples of operating systems 3.1 Microsoft Windows (OS) 3.2 Mac OS X 3.3 GNU/Linux and Unix-like operating systems 3.4 Plan 9 3.5 Real-time operating systems 3.6 Hobby development 3.7 Commodore 3.8 Other 4 Diversity of operating systems and portability 5 See also 6 References 7 External links
Here is a proposal based on a general modularized OS that supports file system, APIs, compilation. (A section on embedded OS will describe a minimal case without all these generalities.) Here is no product names in the section titles.
- Added Components, Input/output, Protection, Concurrency
- Removed Network and Security, Hobby development and Commodore and Other, Examples of operating systems
- Changed History to Evolution, Features to Services, Diversity of operating systems and portability to Type
X Evolution (human operators to resident monitors, batch to spool, link to History of OS) X Architecture (general HW and SW arch intro) X Management services X.x Process (Process control block · Interrupt · multitasking · Scheduling · Context switch · CPU modes · Thread) 2.3 Modes and spaces (priveledge/user modes, kernel/user space, protected/supervisor mode) 2.4 Memory (link to memory hierarchy, virtual memory, etc) 2.x.x Input/output 2.6 File (virtuality, etc) 2.7 Device (virtuality of spooling, Multiprogramming 2.x Graphics (link to Graphics_processing_unit... X Protection (functions, data, access environments, privileges, type checking, synchronization, addressing) X Concurrency (mutex, transactions) X Design (link to Abstract machines) X Type X.x Distributed X.x Network X.x Hobby X.x Embedded X.x Real time X See also X References X External links
The new structure attempts a complete outline the science of operating systems. The outline is the most important part of the article's future development. Once the sections titles are agreed upon it should be obvious to all what the content will become. It also considers a future where links from other articles will look something like
for precision and efficiency.
Please help with this task. This is article is foundational. There are over 400 articles with the term operating system in their title.
Concerning the titling of names for section headings:
- Names should be unique on a page.
- Names do not explicitly refer to the subject of the article.
- Do not name a sections after a product if the subject is broad and there are many competing products. Broad subjects will contain many subtopic articles, and those will cover notable products. Notability guidelines do not apply to content, but section titles do. Structuring by product name encourages article growth, tempts COI, and gives fertile ground for bias and zealous editing. If important subject matter must be withheld from the article titled after a broad topic, let the product information be withheld before the subject information.
- Consider the contextual needs of the wikilinks to Operating system. They are an unbiased poll of what the article is expected to contain, and should serve. Over 900 articles link to here.
- Section naming is Wikipedian, and is not subject to the field's terminology as usual, but is the subject of Wikipedian, information, management.
Some concerns are facts, and some are opinions based on issues.
Because the information space of this article is shifting, please help consider what place and what form it will take:
- A table of contents of a book about operating systems (It may not be plagarism to reuse a title, or fraction of a table of contents.)
- Category:Knowledge representation
- Category:Operating systems is a wealth of structure. There the concept operating system has parent categories (at the bottom in the categories box), and child categories (listed in the section titled Subcategories.
I appeal to the 320 watchers of this page according to the Page History's Number of watchers report. — CpiralCpiral 08:08, 20 December 2009 (UTC)
What is an operating system?
The section title is a rhetoric question. The next line is a a subtle statement.
I removed from our lead paragraph, the notion of an operating system, that would lead the readers to believe they were reading anything other than the general-purpose operating system, whose parts are many and complex, and the coverage of which can only briefly mention the likes of (paraphrasing and strike-through mine):
Almost all
computersappliances embedded with microprocessors (including handheld computers, video game consoles...), robots, domestic appliances (dishwashers, washing machines), and portable media players usean operating systemmicrocontrollers of some type.[dead link] Some of the oldest models may, however, use an embedded operating system that may be contained on a data storage device.
The link that cited them has been dead for a month. The "subtle statement" (above) should take care of the guidance of wayward users, as usual. It is obviously a valuable sentence, worthy of special handling. But please, let us not present it in our first paragraph in such a slightly awkward way way. The paraphrased sentences is in minor contrast with the other sentences in this importantly staged paragraph.
An general-purpose OS from a design point of view will generate a comprehensive set of OS topics for us. An embedded OS has no compiler, and its development platform no users. We want the general purpose system with both end users and developers. Device drivers, ISAs, and compiler design, are also out of bounds here. A general-purpose presentation, like the article Kernel will discuss all the aspects, including design. OS history is all about software development. It automated SW dev from the beginning, and now more than ever. The now-evolved OS, including the vision I paint of this article is going to focus on SW dev after the reader gets past the lead and history sections.
This OS will need to expand it's topical areas (per the class of article it is) as much as possible, and, as stated, limited to the user's POV. But what is user? OS is a big API for the hardware, an interface. Programmers are users of the system calls via this OS-as-API. Programmers develop applications, some of which also have API's, in other words, are intended for use as modules for use by other developers. In the context of developers, "users" are the API users. API users are developers. These are also OS's users. Developers are users of APIs. OS is big API. End users never touch API or OS.
OS serves both, but historically especially serves the developer user. A hobby OS is defined as having a low ratio of end-user-to-developer ratio. We want to create the environment in OS of high "user" to developer ratio. But we want the developer here who use APIs, not the developers who deal with lower level stuff in lower level languages. End users are users of applications, and applications are developed by developers, and the purpose of OS is to develop applications and run applications. An OS is a self-booting thing. If it does not have a boot area, if it does not have development support tools, it is not an OS here. If it involves ass'y lang and ROM, it is not an OS here. Developers user high level OOP lang these days. We are here now with OS.
Those who use OS's services most fully and completely are developers. A high-level language-programmer deluxe is an OS guru, and leaves no latency unturned. (But alas poor guru who view you to 2009 into here OS.) OS was put there to automatically handle the layer below it. This is called abstraction or interfacing, and it translates or bridges or interprets native signals to other native signals. The OS is a big (P)API made of little (y)APIs. End users will want to know what it is by the end of the lead and history sections, but only the CSs will ever consume the meat of the article, which I have not touched on here. I'm talking about general content, in other words, the all-important choice of section titles.
My vision for this article. I was barber Occaam for the irrelevancy phase of OS. (See the above discussions.) Now it's time to fill in the missing facts as implied by how Wikiproject computing has decided to class this article.
- A design section should be included only for the side effect of listing the tradeoffs involved in offsetting the overhead costs of an OS runtime, as OSs handle there own SW pkgs.
- I have no intention to suggest we write about theory or design or expound in a highly abstract way, especially in the lead and history sections, but
- design has the best topic-list,
- theory the most general purpose information, and
- abstraction the clearest, most precise concept available.
- Devices concerns and "OS design from the hardware side" can both be reduced for the same reasons that microprocessor-based systems can be reduced—space, no space.
- This article should be all about the kind users will really want to know and use. The persons who really uses operating systems or will really closely read anything past the lead and first sections are in for a trip to the realm of a science-for-laymen discussion of how the software of software handling works: the audience becomes application programmers. THis is because the OS is an API for handling software packages.
- The scope of the lower layers, starting from the list of topics associated with "HW arch design considerations", and continuing on to instruction sets and device drivers, is a vast and forbidden realm. For example, File systems (as drafted above) goes immediately upward (abstracts) to the user experience—the abstraction, and the virtual machine that the OS produces from the hardware.
- An OS is like that wonderful, graphic picture, next to the paragraph in question, ({{OS}}). Its only use, why it exists, what it was put there for, is to handle software packages, both (1) the development and (2) the resulting application of SW pkgs. We should discuss from the downward blue arrow on up, giving only mention to the fringe.
When the reader is done digesting operating system in a decent sitting, they should be able to smile and burp to themselves that there are these three layers creating a virtual machine out of a physical one, and the concepts involved in making such an amazingly useful and productive machine are so interesting that I want to memorize the article and calmly respond to my friends at the appropriate time with a golden sentence. — CpiralCpiral 03:30, 22 December 2009 (UTC)
- I've a similar question to "What is an operating system?" I know who to blame the first compiler on. And I know who to blame the first high level programming language on. But I can't seem to figure out who to blame for the invention of the operating system. —Aladdin Sane (talk) 04:42, 23 December 2009 (UTC)
- The first OS was a human being. Like a waiter or waitress for developers, they serviced the needs of developers. That was to keep the developers having to make appts. When appts were made for a length of time, that time was usually too little or too much for the developers needs. Also they developer could be late to the appt, etc. The loss was tremendous, as the value of the computer was very high, and to keep from have it doing nothing, OS was invented. Now we have SETI project that keeps idling PC's on a task. Other network OSs projects do the same. You can volunteer your OS for a project of your choice. — CpiralCpiral 04:57, 23 December 2009 (UTC)
Footnotes?
There doesn't seem to be any footnotes at all in the main article.
I think that there are lots of places where a parenthetical digression or concession interrupts the way a paragraph reads, leading to an interruption that is hard to recover from.
The main article could be improved if a lot of parenthetical exceptions, concessions,or digressions were put at the bottom of the page.
Is there a style manual for Wikipedia that relates to footnotes, and the way they are used? Is a footnote only for a bibliographic reference, or can other matters be put at the bottom of the page in the form of a footnote? Dexter Nextnumber (talk) 08:14, 24 December 2009 (UTC)
- See WP:FOOTERS, and perhaps WP:CITE. The way footer section titles relate to content is sparse. In my opinion (only), when you just can't say enough about the topic, put the fringe in a note, and tuck it further away than a parenthetical phrase could like so
- Notes that are side-notes, love-notes, digressions, trivia.
- Footnotes instead of Notes if big words was the style.
- See also when visual terminology was the style.
- Further reading for elementary or intermediate topics.
- Footnotes is notes at the foot of the document. It has an earthy, somatic style.
- Endnotes is notes at the end of the document. It has a eschatological, absolute style.
- A selected bibliography is a quick off-the-top of most popular favorites used to form the authorship
- Biblography is a list of books of interest. They may or may not be citations.
Remove features sections
Because every technical section in OS is covered by other articles, we can write an article that is non-technical. Because every product section has its own article, we can write anything. What are the world's topmost reasons for looking up this article?
- What is an OS?
- How did it evolve?
- How can knowing improve my understanding of computing?
We could concentrate on the surface issues, and just try to be very user friendly. We could talk about non-technical things like cultures, histories, and products; the way operating systems play their part in the big picture, and not so much what the kernel does. We could be somewhat literary without WP:NOR. Hmmm. We would have to explain enough technical information to make the layering and abstraction digestible.
What is commonly referred to as an "operating system" in our culture, including those references in other articles (as shown in What links here), and including the use of term in the current article here is actually perfectly in alignment with what they are getting. But wait! This is an encyclopedia, not a place for common misunderstands to be standardized. There can be no denying that just exactly what an OS is is missing from most people's knowledge, and from this article, judging by the complaints on the talk page —not scientific, not accurate, not inclusive, too long, to much vandalism. A consensus reality of "operating system" has been at play here, and it needs cleaning up. Why? Google's number one search result for "operating system" for one reason. How? Complete rewrite?? Why? Terminology explanations end general OS illiteracy. Rephrasing in ways that bring new wind to cliches.
Firstly, the phrase "operating system" is misunderstood as "system operations". Windows and Linux are not operating systems, technically speaking. They have an operating system, and they do operate the system, but Windows and Linux are more than an "operating system", because they bundled their OS with user software outside the operating system. Their turnkey computer software operates the computer system, and no one is the wiser... except computer architects, software packagers, software developers, and other industry workers. Wikipedia must explain to the masses what the industry knows as where the boundaries of the OS are, what the definition of OS really is, when to answer that caller ID, and when it is not their responsibility.
The common understanding of "system" is the entire computer. The "system" of "operating system" refers to the kernel and shell. This truth pops up in terms like "system call" and "system software" vs the kind of function call to an application's API.
The term "user" has two meanings. There are end-users and then there are users of the OS API. The users of the OS API who make the system calls also encapsulate there application with an API.
Operating systems |
---|
Common features |
In fact the OS is layered. The purpose of these layers
is the key to defining which layers are the OS, and why. But the concepts behind the layers are (1) evolution, and (2) virtualization. Software developments drove the evolution of the OS, and we will have to touch upon software development concepts the way they are today. We'll need the concept of software objects, messaging, and standard interfaces. No getting around that with computers and software. BTW, OS is all software itself, and all about software development. It just happens to run user applications as a side job. The certified parents of any OS are a compiler and an instruction set.
OS should focus just on defining OS, since it is such a mystery. We should take the reader to the (fuzzy) boundary of the OS from all sides, hardware, application, inside (kernel and shell) and outside (user types). This means focusing on (1)evolution (2)layering and (3) abstraction, while circling around and expounding somewhat on what an OS is not. To do that we have to explain modularization, object oriented software development, and discuss the many shell types and kernel types, and the outsides, the hypervisors, the instruction set and microprocessor.
OS should refer to conceptual boundaries, such as memory hierarchy has. An OS is not a network OS or an embedded OS. that focuses on the definition, the form, the surface, the mental construct. If we talk about file systems, we can focus on the section above that asks "How do the 1's nd 0's get from storage to the screen?"
If all sections here are covered elsewhere our job seems clear. Make the most complex thing in the world enjoyable, and even discernible the way the industry discerns it. We devote a good deal of the article on what an operating system is not.
An WP:ACLASS article will have the WP:LAYOUT formats with lots of citations and graphics, etc. We will have to include complete information about what an OS is not, and what an OS is. An OS is so complex, that could mean only one sentence on each topic. The key to a re-write is an outline a structure.
Quick outline. If we build it, we're done; the wiki will descend in droves. Needs above section-outline for resources mentioned, and fit them into this outline instead. The diagram has four layers.
- Computer system as a whole
- Hardware+instruction set+kernel+shell+user+application.
- OS evolve from mainframes and microcomputers, and finally to PC.
- User culture: Science and art, young and old, business and pleasure, large and small OS
- Hardware Layer
- Classes of electronic device controls determine where OS begins
- Device drivers
- Instruction sets
- Application layer
- How OS serves software application: developers as users
- Abstraction layers enhances production of software
- Operating system=Kernel+shell. .
- OS evolved from a program server to a development platform to a hypervisor. cross-compiler
- portability enhances use of hardware
- Inside the OS (See Kernel, Shell
- Protection of memory and processes
- Abstraction layers of hardware, kernel, shell
- Performance tuning hardware and kernel
- Compatibility and portability
- Boot modes, reboots, software upgrades, boot loaders
- Multiple OS, Multiple Processors, plug and play
- Tradeoffs: OS architecture depend on instruction and compiler
- Specialization (Security, Supercomputing)
- Kernel
- Shell
- Graphical
- Command Line
- What an OS is not:
- Types of bundled software
- Modern automobile does not have an OS, it has Microcomputers instead.
- embedded; An OS is imprinted onto an embedded OS.
- microcontroller or state machine;
- a turnkey software system;
- Hypervisor
This will
- focus on what it is, and avoids how it works;
- answer every talk page question and complaint;
- focus on why it is that way, and explain the multiple meaning of words like system and user;
- avoid OS user-issues like comparisons, purchasing, installation, licensing and the like.
- focus on what it is not, and even clarify the layers above (user space) and below (the instruction set and microprocessor) in order to make it clear.
Yes, this is yet another sloppy outline of sorts. Constructive criticism is preferred. Please scan the entire talk page first, and checkout out some of the questions and answers and complaints. It is helpful to understand my current position: a complete rebuild of what is currently Operating system. For an equally fun alternative to a complete rebuild, which is just as fun, please see my posting at WikiProject_Computing#Merge_Operating_system_and_Kernel.
— CpiralCpiral 20:47, 30 December 2009 (UTC)
- I hope the assumption of the user interface would belong the OS is removed. It is needed to explain that the UI can be part of the OS if coded to them. Example in Linux operating system UI is not part of it but is offered by other software (top of the OS). Difficult to read but explains well, pretty boring but accurate about Minix and Server-Client idea vs Monolithic, the basic logic in normal user format and last but not least the very simple comparision of the different OS architectures what actually sums all (but demands still to understand the wider picture). Golftheman (talk) 07:43, 31 August 2010 (UTC)
Proof of coverage
Proving that coverage of all facts currently articulated by operating systems, exist elsewhere will free the wiki to re-write the OS article in any new structure and style. This table could also serve to list the facts OS must cover to become the general purpose representative it is, but in full, A-class form.
Well wiki me. Hint: the parameters at WP:Search. — CpiralCpiral 09:42, 1 January 2010 (UTC)
rewrite
I think this article may needs a rewrite. (here are some thoughts to consider) e.g. Operating System
1. introduction
- abstraction (brief)
- communication
- (assembly language) - fundamental
- BIOS, EFI, interface (cf I/O Peripherals:)
- communication
2. design philosophy
- inter-process communication, behavioral science, ethology
- (e.g. control panel) - driver
- suite: networking protocol, implementation
3. details e.g.
- mechanism: virtualization
- feature multi-(tasking, processor), cross platform
- component engineering (cbse)
- mode: protected, supervised,
- management: file system
basic os (NOS, GUI) --75.154.186.6 (talk) 07:18, 8 February 2010 (UTC)
Microsoft Operating System Penetration in the U.S
I am trying to locate current data on market penetration of the several active Microsoft Operating Systems (XP Home, XP Pro, Vista, Windows 7), preferably by Service Pack.
If anyone can direct me to a reliable source for this information it would be much appreciated.
Thank you
SteveB 65.10.165.133 (talk) 15:03, 15 January 2010 (UTC) sibjazz@comcast.net
Windows and preemptive scheduling
NT (developed in 1989) did preemptive scheduling, but so did Windows 95. The article states that home uses didn't have this capability until Windows XP, which is incorrect. DonPMitchell (talk) 01:17, 8 February 2010 (UTC)
A rift draft
Happy editing!
I made a subpage, Operating system/Draft, where we can implement the section removals, and address the questions and concerns covered in many of the Operating system discussions on this page. My interest with improving this article starts with the discussion section Talk:Operating_system#Why_is_this_article_considered_too_long.3F and goes on from there.
The sections removed in the Operating system/Draft were slated earlier:
Removing the section Features was a not-so-subtle way of saying the article needs to be re-written, a common recommendation in the discussions.
Once we get a good draft, I'll submit it at OSnews, where the ideal result would be a larger discussion and participation from operating system affectionados. Our Wikipedia Operating system article is the number one search result on Google.
— CpiralCpiral 05:30, 19 March 2010 (UTC)
- Hey, I redid the intro. I tried to make it easier to understand for every day people, maybe we can put more of the technical stuff later on in the article. I know a lot of people here want every detail to be included, but there are a lot of people that are looking up this article because they don't know what the hell an operating system is, and talking about system calls and confusing stuff like that in the introduction felt counterintuitive. You can't explain a definition to someone by using a bunch of other definitions they don't understand, which are in turn explained by a bunch of definitions they still don't understand. I like what you are saying, that we need to redo the focus. This article's sections should pretty much be a bunch of simplified summaries of a lot of other articles—if people want more information on some given topic they can just click a link. --Ferrenrock (talk) 08:56, 23 April 2010 (UTC)
two parallel file system sections
There's "Disk access and file systems" and then there's "File system support in modern operating systems". Simply going by the logical meanings, the latter should have a shorter title and be integrated under the former. --Joy [shallot] (talk) 08:23, 30 April 2010 (UTC)
- Go ahead and clean it up if you can. I'm trying to repair this article, it feels like ten thousand creeping fingers added a bunch of details without any structure.--Ferrenrock (talk) 09:31, 30 April 2010 (UTC)
what about what is OS and not what brands
I'm editing this "Operating System" article so it reflects an applied science and is not too long.
I do not wish to change the whole article. Only to add a bit of science to it.
I'm still editing my changes. I totally agree my changes "don't appear to fit in" and need help. Infact I've followed other's editing so far and have taken much of their advice. Whoever edited the article before appeared not to consider anyone else's advice - including wikipedia's.
I think dropping names of past computer releases is fine for maybe a few paragraphs - but such a thing is not even on the topic of OS; it explains nothing, it invariably slants toward certain manufacturers (mostly US), it rewarding them for things they were sued for. Mostly: it's too hot a topic. It's "name dropping".
Talk of disk formats there is no room for: hd chips are only one often changed standard for one chip of many thousands of chips in devices. Furthermore - there are already wiki's on fs types. An OS deals with many chip interface drivers and many formats for each. competing "non-supported" fs types degrade the science.
Moreso. MS is the subject of many lawsuits. Many resent any slanting history in their favor. Their Window's OS cannot be traced into and thus cannot be reported on. I read the much DOS ASM code maybe 20 years ago: repeating what I know publicly would be illegal - as many sued by MS have found the hard way.
No one "owns the OS article". It's a topic from applied sciences and should reflect that somewhere is all I ask.
- You need to make sure people agree before you delete large portions of the article.
Josh
Thank you Josh. I had believed the article unattended and that the author was acting on personal interests. I'm new to editing wikipedia - this was my first attempt. Apologies.
missing text
A third and more recent type is of user interface is the... what? See the User Interfaces section. Can't find it in the edit history. —Preceding unsigned comment added by 59.93.246.118 (talk) 17:30, 11 May 2010 (UTC)
netcom connect
Hvordan kan jeg få slettet netcom connect? Har prøvd mange ganger men får ikke slettet dette.Får feilmelding om at med det er feil oppsettet.Og at jegikke får innstallert dette. Kan noen hjelpe? Send meg epost. —Preceding unsigned comment added by 88.89.122.93 (talk) 22:49, 2 June 2010 (UTC)
Operating systems
I have an operating system that I made some changes to earlier and now my system is a lot slower. Do you discuss what should be in an operating system? —Preceding unsigned comment added by 206.41.91.42 (talk) 20:05, 13 July 2010 (UTC)
how i can explain window explorer —Preceding unsigned comment added by 117.205.65.111 (talk) 16:49, 16 September 2010 (UTC)
Windows section
The section on Windows reads like an advertisement for this operating system. Please consider re-writing the section or altering the tone. Thank you. 124.148.180.71 (talk) 06:29, 18 November 2010 (UTC)
Copyedit
May I request this article be looked over and emended by a copyeditor? The information is presented in an often confusing manner, and not all the sentences are idiomatic. By contrast, the French version is much more pleasing to read, and easier to follow, though less comprehensive.
COS/360
COS/360 (Compatabililty Operating System) was recently added to the list of "other operating systems used on IBM S/360 series mainframes included systems developed by IBM" in the mainframe section of this article. Is COS/360 a true operating system? I'm not familiar with it, but found this description in the paper "System/370 integrated emulation under OS and DOS", by Gary R. Allred, International Business Machines Corporation, afips, pp.163, Proceedings of the Spring Joint Computer Conference, 1971:
- Two programs, Compatibility Operating System (COS)/30 and COS/40 were developed by IBM which integrated the 1401 emulator on System/ 360 Models 30 and 40 under DOS. At first considered to be interim programs, these programs, because of their wide acceptance and usage, were subsequently upgraded through hardware and software refinements and renamed Compatibility System (CS)/30 and CS/40. For the first time, 1401 jobs and System/360 native-mode jobs could be run concurrently in a limited multiprogramming environment. (Limited multiprogramming in the sense that there were certain restrictions on the Foreground/Background allocation of jobs under DOS.) Single job stream input was also possible. Overall system throughput was significantly improved by eliminating the need to reload the system between emulator and System/360 jobs.
At a minimum we need to fix the spelling of compatibility in the article. Jeff Ogden (talk) 16:37, 10 January 2011 (UTC)
Books or resources
Can somebody add resources, reading materials list, or list of books and where to buy about operating systems? This will be very helpful. Perhaps free downloadable pdf files so we can just download and print. It will save the readers' money and it is easier. — Preceding unsigned comment added by Talaga87 (talk • contribs) 03:27, 26 March 2011 (UTC)
Grammar (a little pedantic)
A recent edit to Types/Real-time (early in the article) seems at face value to be a good one but I think it could do with re-evaluation. Replacing "Objective" with "Value" would seem to make greater sense if the statement is otherwise left as it is. But perhaps I have the wrong idea about what the statement should say. If "Objective" is the most appropriate word then perhaps the statement needs to say something along the lines of "The main objective of real-time operating systems is to react quickly...". I would have simply edited but saw no harm in discussing the course of action first. Hardly the end of the world if the statement stays as it is though. -- FG/T|C 05:56, 26 July 2011 (UTC)
"its successors making Microsoft."
Not sure what the intended meaning of this is, but it doesn't make too much sense in its current form... AnonMoos (talk) 16:59, 13 September 2011 (UTC)
Two Non-free files on this article
I've removed the following non-free files from this article:
My reasons for doing so are this:
- Both images are used in a decorative fashion. It might be appropriate to include both images on an article with sourced discussion regarding differences in these two operating systems, but here that is not happening.
- In no case are the images described in the prose of the article. This is a failure of WP:NFCC #8. With or without the images, the article reads the same and no understanding is lost.
- Both images are used elsewhere. If a reader must learn more of the look and feel of a given operating system, they can go to that operating system's main article page.
- The purposes in rationales are weak and/or copy/paste jobs. This fails WP:NFCC #10c.
Non-free content must be strongly justified. We can't just slap a rationale on it, and say it's ok to use it. There has to be a strong reason why we must use it in order to be encyclopedic. That is clearly not the case here. --Hammersoft (talk) 17:09, 30 September 2011 (UTC)
- How much has the Foundation spent on defending itself from copyright suits? Can you imagine the media stink that would arise if Microsoft sued Wikimedia? And Amiga is a dead issue, so there's no party that would be interested in suing in the first place. Is there a (real) lawyer in the house? --Wtshymanski (talk) 19:29, 30 September 2011 (UTC)
- The issue has nothing to do with whether the Wikimedia Foundation will get sued. Sorry. This isn't a matter of law. It's a matter of our m:Mission, and our policy. --Hammersoft (talk) 20:51, 30 September 2011 (UTC)
Partially proprietary?
This is a meaningless phrase. Either the whole item is generically available from multiple sources, or else you have to get it from one vendor. The "Darwin" article is confusing (as usual for Wikipedia computer articles) as it isn't clear (to me, anyway) that I can take a blank hard drive in a Macintosh, a Darwin disribution and expect to run Mac OS X applications on it. Certainly if you upload a complete Mac OX X distribution to your friendly local FTP server, Apple will quite firmly instruct you to take it down as soon as they find out about it; this indicates a proprietary interest, to me. --Wtshymanski (talk) 15:09, 17 October 2011 (UTC)
- As I understand it, something is either proprietary or it isn't. Mac OS is owned by Apple and is a licensed product, thus it is proprietary. Mine may be a simplistic view but it is ridiculous to use (as Wtshymanski has already said) the phrase "partially proprietary" as this is (as I see it) simply not possible. There may be licencing and/or sourcing concerns that make Mac OS available without paying for it but that certainly doesn't make it free or any less proprietary. -- fgTC 17:46, 17 October 2011 (UTC)
- The underlying BSD Unix portion of OS X is free and open source software. Apple does not own and cannot restrict the copying or use of that portion of the OS. OS X will not operate without this underlying open source component. All the other portions of OS X are Apple-proprietory, but the whole OS is not, being a combination of open source components with proprietary components. "Partially proprietary" is a perfectly good phrase to describe this situation. You can take and sell parts of OS X without transgressing Apple's copyrights on the other parts. Yworo (talk) 23:03, 17 October 2011 (UTC)
- Ducking out on the ground I can see how both sides are correct. "Partially proprietary" does sound twisted however true it may be. -- fgTC 23:42, 17 October 2011 (UTC)
- Can you take and sell enough of OS X that the parts are a complete operating system? Doubt it. Can you sell something called "OS X" and not run afoul of Apple's proprietary interests? Also doubtful. Why is the murky distinction "partially proprietary" useful? --Wtshymanski (talk) 00:39, 18 October 2011 (UTC)
- Yes, you can. It's called Darwin (operating system)! --Cybercobra (talk) 02:24, 18 October 2011 (UTC)
- That's a different product, though? Anyway, "open core" is a better description and not so absurd on the face. --Wtshymanski (talk) 13:52, 18 October 2011 (UTC)
- Yes, you can. It's called Darwin (operating system)! --Cybercobra (talk) 02:24, 18 October 2011 (UTC)
- Can you take and sell enough of OS X that the parts are a complete operating system? Doubt it. Can you sell something called "OS X" and not run afoul of Apple's proprietary interests? Also doubtful. Why is the murky distinction "partially proprietary" useful? --Wtshymanski (talk) 00:39, 18 October 2011 (UTC)
- Ducking out on the ground I can see how both sides are correct. "Partially proprietary" does sound twisted however true it may be. -- fgTC 23:42, 17 October 2011 (UTC)
- The underlying BSD Unix portion of OS X is free and open source software. Apple does not own and cannot restrict the copying or use of that portion of the OS. OS X will not operate without this underlying open source component. All the other portions of OS X are Apple-proprietory, but the whole OS is not, being a combination of open source components with proprietary components. "Partially proprietary" is a perfectly good phrase to describe this situation. You can take and sell parts of OS X without transgressing Apple's copyrights on the other parts. Yworo (talk) 23:03, 17 October 2011 (UTC)
MVS in history
MVS introduced some important concepts, but transparrent data caching wasn't one of them. Both Multics and IBM Time Sharing System (TSS/360) were there first.
I believe that I can provide references for compatibility from OS/360 through z/OS. Howeveer, there are some edge cases that are not compatible. Shmuel (Seymour J.) Metz Username:Chatul (talk) 00:07, 2 November 2011 (UTC)