Talk:Shell (computing)/Archive 2

Archive 1 Archive 2 Archive 3

"Shell (computing)" or "operating system shell"?

  OS   Appli-
cation
Command line
{+
command
line
interpreter
+}
operating
system
shell
GUI +
  not
called
a shell
All the above is user interface

The article is a confusion between two meanings: "operating system shell" (including CLI and graphical ones), and "command line interpreter" (including OS CLIs and application software CLIs). I think that the article should be technically moved to operating system shell, all lists of CLIs have to be merged with list of command-line interpreters, all information about application "shells" should be removed (possibly to command-line interface), and then, the page "shell (computing)" should redirect to a disambiguation page, where not only above meanings will be listed, but also Secure Shell and other things referred to as a "shell". "Shell (computing)" is inherently confusing and should not lead to an article. Incnis Mrsi (talk) 09:19, 12 March 2012 (UTC)

I agree that the list of command line interfaces should be move to an appropriate article i.e. list of commanf line interfaces. If you do that however, the list of GUIs ought properly to be moved to an appropriate article as well. I disagree that command line interface needs to be split off from this article per se. Also if you want this article to be about just operating system shells then this article ought to be renamed, but a disambiguation page ought to be left for computing shell to describe the myriad uses of the term. Op47 (talk) 13:29, 9 September 2012 (UTC)
I already wrote just above:
Which "myriad uses" beyond aforementioned three, and also shell account, did you find? So few items as 4 do not require a separate dab page. A dedicated section in shell will be enough, just like the infamous code (computer programming). Incnis Mrsi (talk) 16:28, 9 September 2012 (UTC)
  • Oppose rename. I think the article is very clear in stating what it's about, namely, "an interface for users of an operating system which provides access to the services of a kernel" and that they "fall into one of two categories: command-line and graphical." Nearly any article can be improved and this one probably could be as well. But what it doesn't need is a new title. Msnicki (talk) 17:06, 9 September 2012 (UTC)
    The consideration that an article "is very clear in stating what it's about" has nothing to do with WP: Article titles (although I do not agree with Msnicki in that "the article is very clear", anyway). Look on the IPython article, at the very beginning:
Why have we accept that Msnicki knows what is The Very True Shell in Computing better than authors of the article about IPython? Otherwise, a diagnosis is ready: the title is ambiguous. Incnis Mrsi (talk) 18:20, 9 September 2012 (UTC)
The attitude is unnecessary. We're all anonymous here, each entitled to an opinion and to be treated respectfully. We are expected to assume good faith. I appreciate that you'd like to make a change and if the consensus supports you, we'll do it even over my objection. That should be good enough. Msnicki (talk) 01:45, 10 September 2012 (UTC)
  • Oppose rename, but massively edit the existing article.
There is a vast literature out there on "shells" as command-line interfaces. There is an equal volume on GUIs, graphical desktops and other graphical user interfaces. The use of "shell" applied to such a graphical interface is almost unheard of. An obscure and obsolete version of OS/2 is the only one I recall.
This article is a nonsense. It presents both command line and GUI, as if the article topic was UI or HCI in general, and it gives equal weight to GUIs under this "shell" label. That's simply wrong. Although at least one graphical shell has used the already-widespread "shell" term for its innovative GUI, this is very much the exception. Claiming that GUIs are described as "shells" such that they deserve equal weight in this article specifically on shells is ridiculous, not to say unsourced. It's also riddled with basic technical errors, such as over-stating the amount of function that's provided by Windows Explorer.
A question of trivial renaming about disambiguators is not the point here (although Shell (computing) better suits our general policy on their technical formatting, as it allows the pipe trick). This article should usefully be re-written to remove most of the coverage of GUIs as being outside its scope. It might instead be renamed as "operating system user interface", which would at least be closer to the scope included at present. Andy Dingley (talk) 13:03, 10 September 2012 (UTC)
I do not insist on any particular title to move the article to; the only thing I perceive as an important one is making the title "shell (computing)" ambiguous. But I have yet several comments. First, the Andy's perception of "shell" (GUIs are outside its scope, CLIs are inside) is totally opposite to Msnicki's one (an interface for users of an operating system… fall into one of two categories: command-line and graphical). Note that the article about command-line interface already exists. Second, Andy is certainly not right that “the use of "shell" applied to such a graphical interface is almost unheard of”, see Windows shell replacement. Operating system shell or operating system user interface certainly is a valid topic which, in its explicit form, English Wikipedia still misses. BTW ru:Оболочка операционной системы literally translates as "operating system shell". Incnis Mrsi (talk) 14:45, 10 September 2012 (UTC)
It's not about what I or Msnicki think a "shell" means, it's about what's sourceable. This article is seriously lacking in sources, especially for equating GUIs with the term "shell". Once again, WP is not an acceptable recursive source for WP. Andy Dingley (talk) 15:24, 10 September 2012 (UTC)
So, suppose that we verified that a sufficient amount of sources with the word "shell" exist for all 3 cells on my diagram. What do you propose to do in this situation? If you agree with me that "operating system UI" must be separated from command-line interface, disregarding terminological issues, then I'll go forth with the semantic separation and then, you can discuss which title would be better for the article about OS UI, and how to mention it from the "shell" dab page. If you do not agree, then we shall continue arguing here. Incnis Mrsi (talk) 15:46, 10 September 2012 (UTC)
Obviously command lines are radically different to GUIs and both are important. So in terms of delivering good, clear, articles we should probably split the scope on that basis alone. A lightweight overall article for both would be worthwhile, but it's never going to be much more than a navigational aid.
I care very little about the article naming. But if it has to be called "shell", it does need to source that name. The article in its current state is a long way from that. Andy Dingley (talk) 16:58, 10 September 2012 (UTC)
Existence of the term "Windows shell" is corroborated by MSDN: http://msdn.microsoft.com/en-us/library/windows/desktop/bb773177%28v=vs.85%29.aspx
Also, at least one printed book refers to different (CLI and graphical) "kinds of shell" under MacOS. Incnis Mrsi (talk) 19:43, 10 September 2012 (UTC)
On this point, I would agree with you and disagree in part with Andy. Andy is correct in arguing that the basic claim that there are two kinds of shell should be sourced. But I think he goes too far to extent he seems to be arguing the claim is untrue. I think it is true. I think the two kinds are command line and graphical. I think that is how people understand the term and I say that even though I can remember feeling very uncomfortable with that in the 90s, especially as someone (check my edit history) who really likes command line shells. (Btw, setting aside arguments about WP:OTHERCRAP, I see no conflict between shell as it's understood in this article and the use of the term in connection with IPython. A shell is a layer that lets you manipulate what's beneath it. IPython looks like a shell to me.)
What changed my view, causing me to decide that shells now include graphical shells were two things. The first was realizing I had to concede that graphical is just the other obvious UI paradigm for building a thin layer around everything below; it's not necessarily a limitation on what it does. Also, arguing from WP:COMMONNAME, I think that is how people now use and understand the term, especially in the Windows world, where few users ever touch a command line, yet they do manipulate the system with the explorer and graphical admin tools, etc., and I do hear them call that a shell. Basically, the only thing they really give up are the programming constructs (probably of little interest to most end-users), replaced with multi-select. Msnicki (talk) 21:14, 10 September 2012 (UTC)
A layer that lets you manipulate what's beneath it has exactly one name, quite simple: user interface. There are two orthogonal classifications of it: by design and by purpose. There are two classes relevant to the word "shell": CLIs (a design) and OS interfaces (a purpose). Let us just separate one from another, and dump the ill-fated title into a dab page via anchored redirect. Do not invent unnecessary complications. Incnis Mrsi (talk) 22:06, 10 September 2012 (UTC)
No. Msnicki (talk) 22:57, 10 September 2012 (UTC)
So, if you insist that there exist some notion of the "shell" which is neither operating system shell (an undisputably valid concept), nor command-line interface (with the article already existing), then why is it not the entire "user interface"? Try to propose your definition, preferably not by explicit exclusion of all application non-CLIs. Incnis Mrsi (talk) 09:55, 11 September 2012 (UTC)
I've given you my answer. I'm satisfied with the article title as it is and I don't find your arguments persuasive. You don't have a consensus and you're not getting one, certainly not by continuing to argue as if you don't think anyone else's opinion could have merit. Personally, I don't think you know what you're talking about. I'm opposed to your proposal. That's pretty much all you need to know. Msnicki (talk) 14:47, 11 September 2012 (UTC)
  • Oppose: First, the thing is called "shell", not "operating system shell". Second, most (if not all) shells are not tied to any particular operating system. Third, the different between classic shells and programming languages interpreters is really very blurry, so enforcing the difference will lead to more confusion. Overall, the article could benefit from explanation of the problems, but not from this change of name. BTW, GUIs are sometimes called "shells". Some of them are even mostly referred to as shells, eg. Windows 3.1, Workplace Shell — Dmitrij D. Czarkoff (talktrack) 18:57, 11 September 2012 (UTC) updated 19:14, 11 September 2012 (UTC)
    Which "thing"? Give your definition, please. Incnis Mrsi (talk) 08:53, 12 September 2012 (UTC)
    Interactive command line interpreter. (ksh(8) – OpenBSD System Manager's Manual says: "ksh is a command interpreter intended for both interactive and shell script use".) — Dmitrij D. Czarkoff (talktrack) 16:33, 13 September 2012 (UTC)
    I did not ask for reliable sources. BTW, in this context a man page is useless because states something about what some program (such as ksh) is, but does not explain what the concept means in general. I ask for a comprehensive definition. But now, I lean to creation of a clean article from scratch and am not particularly interested in continuation of the dispute. Users, who are interested in Wikipedia making sense, will contribute to an article which makes sense. Those who are willing to keep a WP:DICTIONARY article, will keep an unsense. Incnis Mrsi (talk) 20:24, 14 September 2012 (UTC)
    BTW the "not tied" argument is completely absurd. A bicycle wheel unlikely is tied to some particular model of bicycle, but it is a wheel for a bicycle, not for a motorcycle, for example. An airstrip is not intended to use with a particular model of airplane, but it is used for airplanes, not sport cars. Cf: an operating system shell is not tied to a particular model of OS, but is the shell of an operating system, not of a video game. Incnis Mrsi (talk) 09:41, 12 September 2012 (UTC)
    Bicycle wheel in general is definitely not tied to any particular bicycle, but each bicycle wheel is tied to some bicycle. To contrary, shell isn't necessarily tied to operating systems. Python isn't operating system, but it has its shell. That said, Windows Media Center is a graphical shell, as well as some other non-OS software. It may be practical to have separate article on graphical shell, but separating CLI shells of OS from CLI shells of programming languages is at best impractical (though "plain wrong" sounds more appropriate to me). — Dmitrij D. Czarkoff (talktrack) 16:33, 13 September 2012 (UTC)
    I said what I said: "not tied", as an argument against the term "operating system shell", was absurd. Its absurdity does not depend neither on Python nor on Windows Media Center, only on the incorrect assumption that a term "X’s Y" implies a dependence of Y on some particular instance of X. Also, I spoke about conceptual separation. Dmitrij D. Czarkoff apparently does not understand what means "enforcing the difference" in Wikipedia. It does not mean that any thing must be classified either as A or as B, but not both. It actually means that the statement "Y is A" must have a well-defined predication A, not "maybe A, maybe B. Existence of the current article hinders the purpose of internal links to clarify linked terms. Each link should mean either command-line interface or operating system shell, but now there is not way to link to the OS shell concept without adding a terminological confusion. Unlikely one can demonstrate a link bound to "shell (computing)" where both CLI and OS UI, in one article, represent a desired target. Incnis Mrsi (talk) 20:24, 14 September 2012 (UTC)
Do you think dismissing other editors' opinions and insisting they don't "understand" stuff is likely to persuade them to your view? I don't. You've got unanimous opposition. Isn't it time to let this go? We decide by consensus and sometimes it goes against you. That's just how it works. Msnicki (talk) 20:41, 14 September 2012 (UTC)
Incnis Mrsi, you could probably find more sense in my response upon reading it. — Dmitrij D. Czarkoff (talktrack) 00:10, 15 September 2012 (UTC)
Dmitrij D. Czarkoff introduced three arguments. First one, potentially, is valid with appropriate definition, but such definition is not yet provided. Second one I demonstrated to be a grammatical fallacy. Third one I repudiated on structural grounds (Wikipedia can and does classify concepts, but does not aim to classify each phenomenon). On the other hand, noone of Msnicki, Dmitrij D. Czarkoff and Kvng did not attempt to repudiate our crucial argument about inherent ambiguity of the title "shell (computing)". As a reasonable person can notice, the support-the-mash side is losing by arguments, but does not perform so poorly in numbers, scoring three votes (I do not count Andy Dingley, of course, because he actually favoured the separation). Unfortunately, too many against me and Op47. Incnis Mrsi (talk) 06:08, 15 September 2012 (UTC)
Incnis Mrsi, this article isn't ambiguous, it covers the single coherent topic (though it omits one of the members). — Dmitrij D. Czarkoff (talktrack) 10:56, 15 September 2012 (UTC)
  • Oppose Create a Computing section in Shell to disambiguate other computing topics that include shell in their title. No need to have separate articles for GUI and CLI shells. Command line interface needs to acknowledge this article. List of shells is repeated in Unix shell. There's plenty of cleanup to be done but the basic structure seems to be sound. --Kvng (talk) 21:57, 11 September 2012 (UTC)
    Separate articles on the TUI–GUI divide, you mean? Then, you understood the proposal incorrectly. Try to read it again. Incnis Mrsi (talk) 08:53, 12 September 2012 (UTC)
I think I understand the proposal. My salient comment (last sentence) is that the organization does not need to change. The rest of my comments are suggestions for alternate means of achieving the goals of the proposal -—Kvng 14:35, 14 September 2012 (UTC)
  • Comment (I didn't carefully read all comments above, so feel free to ignore this comment as well.) During the Windows 3.1/95 era (their) GUIs where known as shells, today not so much anymore. I've never heard REPLs being referred to as "interactive shells". A shell is always a user interface, but a user interface is not always a shell. There might be other uses of "shells" in computing, so renaming this to Shell (operating system) would do little harm, although I don't think it necessary either. In general it's completely pointless arguing over small issues like the title or lead section of articles that are in as dismal a state as this one. —Ruud 22:31, 11 September 2012 (UTC)
    There is no internal links bound to your red title, but a plenty of links bound to "operating system shell", hence one should not invent completely new titles. Give your definition of an "[interactive] shell", please. I could make article much better had I know what is it about. As I stated a half-year ago, I see nothing here but a confusion and mixing-up two valid but distinct senses. Incnis Mrsi (talk) 08:53, 12 September 2012 (UTC)
    A shell is is the user interface to the operating system (or kernel, if you prefer the walnut metaphor). In the narrow sense that may include COMMAND.COM in DOS, sh in Unix and the taskbar and Explorer in Windows 95. In the broad sense it may even include (the user interface) of many of the utilities shipped with an operating system (e.g., FORMAT.COM, mkfs, the "Format Disk" window). It generally does not include all the user interface services an operating system may provide to 3rd party applications. —Ruud 17:38, 12 September 2012 (UTC)
Definition of 'shell'

shell: n.

        [orig. Multics techspeak, widely propagated via Unix]

        1. [techspeak] The command interpreter used to pass commands to an operating system; so called because it is the part of the operating system that interfaces with the outside world.

        2. More generally, any interface program that mediates access to a special resource or server for convenience, efficiency, or security reasons; for this meaning, the usage is usually a shell around whatever. This sort of program is also called a wrapper.

        3. A skeleton program, created by hand or by another program (like, say, a parser generator), which provides the necessary incantations to set up some task and the control flow to drive it (the term driver is sometimes used synonymously). The user is meant to fill in whatever code is needed to get real work done. This usage is common in the AI and Microsoft Windows worlds, and confuses Unix hackers.

        Historical note: Apparently, the original Multics shell (sense 1) was so called because it was a shell (sense 3); it ran user programs not by starting up separate processes, but by dynamically linking the programs into its own code, calling them as subroutines, and then dynamically de-linking them on return. The VMS command interpreter still does something very like this.

Eric Raymond's Jargon File

  • Oppose. 'Shell' commonly refers to a CLI shell simply because throughout most of computing history text was king, however the 'shell' is nonetheless a more generic concept. There are already articles for the Command-line interface, Operating system shell, and Graphical user interface - each distinct from each other and this article. This article should provide a broad overview on the subject of 'shells' and provide the general context for how the other articles fit into that greater context. This article needs cleaning up, not a merger.
    Sowlos (talk) 17:46, 16 September 2012 (UTC)
  • Oppose. The WP:COMMONNAME for these things is "shell" not "operating system shell", even if shell does need to be disambiguated from eggshells, seashells and all other kinds of shells. Msnicki (talk) 10:56, 20 April 2014 (UTC)

Merger proposal

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


Operating system shell article duplicates quite a lot of content already well presented within the Shell (computing) article (particularly the Shell (computing) § Graphical shells section), introducing a potential confusion at the same time. As the Shell (computing) article provides a better structure and carries a more universal title, Operating system shell should be merged into it. Let's discuss. — Dsimic (talk | contribs) 05:18, 20 April 2014 (UTC)

Argh, there's another article on this? I've only barely been sold on the need for an article here at all; we certainly don't need two. I do think that at least some of that content can be salvaged when it's merged here. Chris Cunningham (user:thumperward) (talk) 10:04, 21 April 2014 (UTC)
Given that it's unlikely that there will be opposition in principle to a merge, I've gone ahead and carried it out. Some of the imported content is low-quality, but it actuially meshes quite well with what we had here and helps to fill out short sections. There's still a lot of work required here though. Chris Cunningham (user:thumperward) (talk) 10:23, 21 April 2014 (UTC)
Umh, too much crap went into the destination article. Sorry, but I'll revert the merger changes and do it carefully by hand later today. — Dsimic (talk | contribs) 10:36, 21 April 2014 (UTC)
Urgh. I fail to see why you bothered wasting everyone's time asking for input, in that case. Chris Cunningham (user:thumperward) (talk) 11:09, 21 April 2014 (UTC)
Per WP:MERGE I guess. That's why we should wait for opposition until 05:18, 27 April 2014 (UTC). — Dmitrij D. Czarkoff (talktrack) 11:23, 21 April 2014 (UTC)
Exactly, that's how mergers are to be performed and there should be some time for people to comment, as Czarkoff pointed out. So, let's leave it as-is at least until the end of April. — Dsimic (talk | contribs) 17:12, 21 April 2014 (UTC)
Dsimic, the merge proposal apparently meets no opposition. I believe you can implement it if you want to do so. — Dmitrij D. Czarkoff (talktrack) 22:49, 1 May 2014 (UTC)
It seems so. I'll try to do that in the next day or two. — Dsimic (talk | contribs) 23:27, 1 May 2014 (UTC)
Sorry for the delay, currently the time is a little bit tight on my side, and I'll merge the articles on Sunday. — Dsimic (talk | contribs) 03:15, 3 May 2014 (UTC)
  • Support – after much deliberation... Best regards, Codename Lisa (talk) 11:33, 29 May 2014 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Shell vs. Desktop environment

  • I'm slightly confused here: what's the difference between an operating system shell and a desktop environment? Is a shell a certain section of the D.E. that solely deals with management of files or are the words shell and D.E. synonyms? --67.70.21.13 05:17, 6 Mar 2005 (UTC)
    • This could be answered from a specific reading of both articles:
      • Desktop environment: a desktop environment (DE) offers a complete graphical user interface (GUI) solution to operate a computer.
      • Operating system shell: ircThe shell of an operating system is a program that presents an interface to various operating system functions and services.
    • So, while a DE provides a means of a GUI to the user, it need not concern itself with manipulating many aspects of the OS, while a shell is intended for this purpose. To answer your question; a shell and DE overlap, but I wouldn't say they were synonymous or had the same purpose. HTH Dysprosia 05:25, 6 Mar 2005 (UTC)
    • Enlightenment about the "kernel"/"shell" dichotomy can also be found at kernel. Uncle G 12:14, 2005 Mar 6 (UTC)

IMO, desktop environments like GNOME or KDE SC, provide an entire plethora of programs. I think the main problem is the distinction between the GUI of any program, which is composed of GUI widgets and works as part of a Windowing system and special programs like GNOME Panel, GNOME Shell, which can best be termed as "graphical shells". Some people call them UX. User:ScotXWt@lk 11:02, 14 June 2014 (UTC)