Talk:Unix/Archive 2

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Case-sensitivity

Wahkeenah recently twice inserted some text about case-sensitivity and Unix, and two of us have reverted him. The first insertion of the text was blatant PoV; the second blunted this a bit (by changing to a now-out-of-place but at least NPOV headline), but the text was still misguided.

Yes, the average user on the average Unix sees a more case-sensitive environment than they may be used to, and that may be troubling to a new arrival.

(Wahkeenah, based on your edits here and at vi, I assume you're a new arrival to Unix?)

Hang on, I agree it should be UNIX and not Unix. The owner of the trademark says so. See there web page at http://www.unix.org/trademark.html where it says "UNIX is a registered trademark of The Open Group". Drkirkby 09:11, 23 October 2006 (UTC). I don't know if its possible to rename the page UNIX rather than Unix, but to be consistant with owners of the trademark, it should be UNIX and not Unix. All this talk about case sensitivily of files names, terminals have only upper case characters is irrevant. Yes, UNIX commands are case-sensitive, but the name should not be. Drkirkby 09:30, 23 October 2006 (UTC) Also, to add more weight to the case for UNIX, we all know it was developed by Ken Thompson, Dennis Ritchie, and Douglas McIlroy. In the book, "THE C PROGRAMMING LANGUAGE" by Ken Thompson & Dennis Ritchie, it is refered to as UNIX and not Unix. See for example chapter 8 of the 1st edition, which has the title "The UNIX System Interface". So given the original developers of the operating system refered to it as UNIX, the current owners of the trademark refer to it as UNIX, what possible arguement is there for this page being called Unix? I think we should respect the original developers. Can someone from Wikipedia change the page to UNIX from Unix? Drkirkby 09:30, 23 October 2006 (UTC)

But the case-sensitivity isn't a function of the Unix operating system. For files, it's a function of the file system. ufs and most Unix-oriented file systems are case sensitive but others (such as pcfs when using 8.3 filenames) are not. For commands, it's a function of the particular shell you're using combined with the file system (for the verbs) and the utilities (for the qualifiers). But case-sensitivity isn't really an inherent part of Unix; the MacOS/X GUI certainly goes some distance towards proving that.

Wahkeenah, if you want to include text about case-sensitivity, take some time, develop your proposed text here on this talk page and expand it until you're including enough of the facts. We can help you create a balanced, NPOV presentation. But right now, it feels to me like you're still trying to bash Unix because of a characteristic that simply hasn't become "native" to the way you think.

Speaking for myself, having worked in environments both case-sensitive and case-insensitive, I prefer case-sensitivity in filenames and symbol names, don't care about it in commands, and would usually prefer it not exist in qualifiers. But after years of dealing with this, I can operate in whatever modes the particular OS, shell, and file-system throw at me. I find it a lot easier to deal with ufs's case-sensitivity than my PC's NTFS set of how-ever-many-characters it is that thou shall not include in file names. (Oh yeah: \ / : * ? " < > | Or is that a limit of the PC UI? ) But beauty is in the eye of the beholder, right ;-) ?

Atlant 17:19, 1 May 2006 (UTC)

I stand by my original complaints. Wahkeenah 22:56, 1 May 2006 (UTC)

Fortunately, Wikipedia is not a platform for you to voice complaints. Whether or not the complaint is valid, it is not for you to add your complaint to this article. I stand by Atlant's reversion of your addition. -lethe ^{talk +} 23:54, 1 May 2006 (UTC)

I agree with Lethe and Atlant. I've managed, literally, thousands of *nix boxes and have never heard of anyone who had a problem with case-sensitivity (other than an occasional, initial realization when a novice user first finds the importance). It's definitely not a popular complaint. Simply saying "I stand by my original complaints." is not very helpful. Wikipedia is not a soapbox. This is the same reasoning we reverted similar 'complaints' you logged regarding the vi editor in its article. -- The Deviant 12:39, 3 May 2006 (UTC)

Case sensitivity was largely tied to the arrival of text-I/O peripherals that could support both cases, unlike most DP equipment of the early 1960s. Unix certainly was different in this respect than most other OSes of the time, but I don't think it was actually the first to support both cases as distinct at the filename and command level. Note that the early Unix terminals were model 33 Teletypes, which supported upper-case only, and the terminal driver mapped the upper-case ASCII codes to and from lower-case ASCII! (Upper-case was specified via an escape sequence.) Experienced Unix users generally have no complaints about case sensitivity; it's pretty much a complaint of Unix novices who have previous experience using case-insensitive systems. Anyway, I agree that the Wikipedia is not a forum for airing complaints. A neutral mention of case-sensitivity as a characteristic property of Unix would be appropriate, but not a discussion of whether it is a good idea. — DAGwyn 09:18, 1 September 2006 (UTC)

Uhh probably irrelevant, but is it officially UNIX or Unix? — Elwayfan01 02:18, 5 October 2006 (UTC). Officialy it must be UNIX, so the official trademark owner, the Open Group, use the term UNIX - see their web page. http://www.unix.org/

• Talk:Unix/Archive 1#.22UNIX.22_vs._.22Unix.22
• Talk:Unix/Archive 1#UNICS_vs_UNIX

¦ Reisio 06:55, 5 October 2006 (UTC)

Nextstep, openstep, mac os x, freebsd

Isn't darwin based on a freebsd core? I thought nextstep's influence was mainly in the UI. Ideogram 12:23, 31 May 2006 (UTC)

I guess darwin descended from nextstep which took parts of BSD? Ideogram 12:48, 31 May 2006 (UTC)

NeXTSTEP begat OPENSTEP, OPENSTEP begat Rhapsody (OS), Rhapsody (OS) begat Mac OS X. None of these operating systems included a "pure" version of FreeBSD, but there are some elements of BSD in all, especially in Rhapsody, if my memory serves me correctly. Dysprosia 07:16, 1 June 2006 (UTC)

OS X has had a frequent exchange of components with the modern BSDs, with Apple donating code to FreeBSD directly IIRC as well as using code from FreeBSD and OpenBSD. This is utility and library code, not kernel code. --FOo 09:01, 1 June 2006 (UTC)

There's actually a big hunk of NetBSD code in there too, so I'm told, evident if you run strings on the various binaries. --moof 09:44, 1 June 2006 (UTC)

For what it's worth, the whole family tree thing is a little confusing. Are we talking about the kernel, the libraries, the tools, or the UI? When I first saw the tree, I assumed it was describing kernel evolution, in which case CMU's Mach was conspiciously absent from the OS X history. Branciforte3241 05:14, 16 August 2006 (UTC)

The Unix family tree is indeed somewhat complicated, especially once the different development branches started adopting pieces from each other. Rather than add complication, it might be best to leave it as it stands. — DAGwyn 09:33, 1 September 2006 (UTC)

dot-com crash and DEC and DG

Strictly speaking DEC and DG died before the dotcom crash, didn't they? Ideogram 14:08, 1 June 2006 (UTC)

Yes. — DAGwyn 09:30, 1 September 2006 (UTC)

The old GNU / Linux debate

Not to drag this in here, but there's a reasonably well-established consensus on its own article that this OS should be referred to as "Linux" generally and "GNU/Linux" on those occasions where the FSF or Debian are being discussed. This article currently reverses this and rather labours over the topic anyway. In the process of cleaning it up I'm changing the default reference to "Linux" in line with the operating system article. Chris Cunningham 12:06, 21 July 2006 (UTC)

2038

On the other hand, a Unix-based system in a restaurant may calculate your check based on the cost of a meal in 1970 dollars-a pleasant thought, although an unlikely result.

Vandalism? 155.247.166.28 17:16, 21 July 2006 (UTC)

The BSD lawsuit in the 90's.

This article mentions nothing about the USL v. BSDi lawsuit. That lawsuit basically allowed for a free Unix, namely BSD, and paved the way for Linux to fill the void during those years of uncertainty over BSD. Anybody willing to add a blurb about that? Oleg Alexandrov (talk) 17:12, 27 July 2006 (UTC)

That (and the text in the article) is misleading; the lawsuit was not concerned with whether a Unix look-alike could be made freely available in general, but rather with whether IP rights had been infringed by one particular product. I don't think that the spread of Linux was much due to concern over the status of BSD, although that may have been a factor in Linus's decision to roll his own version. — DAGwyn 09:28, 1 September 2006 (UTC)

2038

This paragraph is in a section about the history of Unix. It does not belong there, since the 2038 problem isn't a part of the Unix history, it's a problem that will appear in the future. —msikma <user_talk:msikma> 19:24, 28 July 2006 (UTC)

It looks like whoever originally inserted the 2038 paragraph did so as a slightly whimsical addition to the Unix timeline -- it's a "future history". Being a slightly whimsical guy myself, I think it's a fine place for it to exist, but recognize others may disagree. But it seems more out of place in its new home in the Impact section than it did in the History section. My recommendation would be to return it to History, but if it really annoys enough people, I'd say just remove it entirely instead of moving it to a section where it's even more out of place. --NapoliRoma 04:29, 2 August 2006 (UTC)

PoV in illustrations?

Is it just me, or is the choice of illustrations a bit slanted? Three Unix command line screenshots, and then, hey wow, Linux is all graphical and stuff! Shouldn't there be at least one Unix GUI illustration? After all, Unix GUIs predate the existence of the Linux kernel by maybe a decade, and (oddly enough) the same GUIs available today for Linux are also available for Unix. --NapoliRoma 04:35, 2 August 2006 (UTC)

Please feel free to supplement those images with work of your own! --FOo 05:45, 2 August 2006 (UTC)

general clean-up

When I stumbled across the Unix article, I was struck by several errors and omissions, and also by awkward wording in a few places. As a major participant in Unix history through the 1980s (e.g. member of the original POSIX committee), I am able to correct much of this, and so I did. I tried to maintain a neutral POV and not step on anybody's toes. Feel free to add more tweaks, but if I changed an assertion somewhere it was because I have first-hand information about it. — DAGwyn 09:05, 1 September 2006 (UTC)

Last edit

I didn't mean that the adjectives were ridiculous, I meant that sixty links in a row was ridiculous. I say we take off and nuke the whole section from orbit. Chris Cunningham 00:05, 7 September 2006 (UTC)

Creator?

Some debate over whether UNIX was created by individuals, or by Bell Labs, see this edit [1]. Anyone know the answer? I'm leaning towards its original state (I did revert the edit back to here), reckon it ought to stay there unless concensus / sources can be used to decide. --Oscarthecat 17:46, 10 September 2006 (UTC)

First, you'd need to define what it means for UNIX to have been "created by individuals" or "created by Bell Labs". Is this the distinction between it being an official project of Bell Labs (where, presumably, somebody in Bell Labs decided "let us create an operating system for the PDP-7" and tasked Ken Thompson with doing so) and it being a project where some Bell Labs researchers worked on the project without being officially tasked with doing so?

According to some notes Dennis Ritchie had

The charter for the [PDP-11 UNIX] project, and the reason the machine was obtained, was to develop a document editing and formatting system. The original notion was to use UNIX as a development tool only, and have the editing system run stand alone.

and according to his later comments on those noted, "The earlier work on the PDP-7 just happened in the course of doing research."

Ken and Dennis's CACM paper on Unix says, in the "Perspective" section:

Perhaps paradoxically, the success of the Unix system is largely due to the fact that it was not designed to meet any predefined objectives. The first version was written when one of us (Thompson), dissatisfied with the available computer facilities, discovered a little-used PDP-7 and set out to create a more hospitable environment. This (essentially personal) effort was sufficiently successful to gain the interest of the other author and several colleagues, and later to justify the acquisition of the PDP-11/20, specifically to support a text editing and formatting system. When in turn the 11/20 was outgrown, the system had proved useful enough to persuade management to invest in the PDP-11/45, and later in the PDP-11/70 and Interdata 8/32 machines, upon which it developed to its present form. Our goals throughout the effort, when articulated at all, have always been to build a comfortable relationship with the machine and to explore ideas and inventions in operating systems and other software. We have not been faced with the need to satisfy someone else's requirements, and for this freedom we are grateful.

so I'd say that PDP-7 Unix was originally created by Ken Thompson when he was working at Bell Labs (i.e., it wasn't an Official Bell Labs Project, although they did pay for it, in that Ken was working in the CS research group at the Labs when he developed it, and they paid for his time), PDP-11/20 UNIX was arguably a Bell Labs project, as were the 11/45 version and the PDP-11/70 and Interdata 8/32 version - the last of those was Version 7 Unix, and everything considered Unix or Unix-like these days is a clear descendant of V7.

On the other hand, I suspect there are Bell Labs projects and Bell Labs projects, and that Unix wasn't a project like, say, one of the Electronic Switching System projects or Operational Support Systems software projects, i.e. there was a lot less management involvement, and it might've been management saying "OK, we need a text processing system for patents and the like, and here's what we need it to do; keep us up to date on your progress". Guy Harris 18:44, 10 September 2006 (UTC)

The slightly ambiguous phrase "originally developed" is satisfying enough for this encyclopedic article rather than "created", as is the use of the phrase "Bell Labs employees" to avoid the who or what arguing. I'm more interested in whether McIllroy deserves less credit as was implied with the previous edit. McIllroy became a major participant in the sytem, but I don't think he *originally* developed it. I think that's why the person tried to use the verb "create" in their edit, regardless of what poor choice of a word it was to use. --72.92.129.182 21:36, 10 September 2006 (UTC)

MacOS reference (challenged)

I see somebody added a mention of "Unix-like MacOS" and somebody else added a factuality challenge for this. I think at the place where that occurs, there should instead be a sentence along the lines: "Today, Linux and Mac OS X are the operating systems that most resemble Unix." Also, these systems should be mentioned in the "2000 to present" section. The obvious issue is how to say these are essentially "Unix" when they (at least Linux) aren't directly descended from the trademarked product. — DAGwyn 23:26, 12 September 2006 (UTC)

Was he challenging whether OS X is Unix (it's not in any of the registered product lists for systems that have passed one of the Open Group test suites), or whether it's "the most notable" Unix adopted by a commercial vendor (whether one would call Apple - or even NeXT - a "startup" at this point is another matter)? I suppose if that paragraph really means "Unix", and not just a "Unix-like system", if it challenges the former it also implicitly challenges the latter. (I don't see a challenge of whether it's "Unix-like"; that phrase isn't used in the challenged statement, unless I'm missing what that statement was.)

As for how to say something is "Unix" when not descended from the AT&T source code, "it's Unix if it passed one of the Open Group test suites" and/or "it's Unix-like if it looks 'enough like Unix'" might do. "Unix-like" bothers me a little bit, as I (and you :-)) remember when a lot of "Unix-like" systems didn't try to be directly compatible with AT&T Unix of any sort, but had a similar feel to some flavor of AT&T Unix, but, these days, it's probably good enough to describe systems that haven't passed an Open Group test suite but "look like Unix". Guy Harris 08:50, 13 September 2006 (UTC)

My concern is that somebody who hears about "Unix" and looks it up in the Wikipedia ought to see some links to its major current incarnations, which if I'm not mistaken are pretty much Mac OS X and Linux, with official Unix products such as Solaris Operating System (which I still use) now much less commonly encountered. If the whole article seems to be talking about a largely defunct product line, it will convey the wrong impression about the current availability of things "close enough" to genuine Unix for the purposes of many computer users. — DAGwyn 21:24, 14 September 2006 (UTC)

The thing is, when people say "Unix" they generally are using it to refer to those historic Unixen leading up to and involved in the Unix wars. Making references to modern compatible systems in the history section would seem to be leading the reader. Of course such systems need to be mentioned, but not in the timeline section. We have a "Unix-like" article, after all. Chris Cunningham 00:17, 15 September 2006 (UTC)

Congratulations! You're not running Eunice. --moof 07:42, 15 September 2006 (UTC)

Family history image

I removed this text that had been added by 69.248.146.151:

This image contains error(s). Solaris_Operating_System (SunOS 5.x) is a SVR4 Descendant, not BSD.

The text is referring to the accuracy of Image:Unix.png. I don't know which is correct. — Loadmaster 19:37, 18 October 2006 (UTC)

The editor is correct: Solaris 2 is a sort of chimera, but it's essentially SysV with BSD bolted on as opposed to earlier SunOS ("Solaris 1") releases which were definitely BSD derivatives. I'd suggest removing the image altogether to be honest, it's not pretty and it's far too small to discern anything at thumbnail size. Chris Cunningham 20:23, 18 October 2006 (UTC)

An accurate diagram would be a graph, not a tree; SVR4, for example, incorporated a significant amount of code from BSD and SunOS 4.x, and Solaris 2 and beyond inherit from that.

I'd vote for getting rid of the image, too. Guy Harris 21:20, 18 October 2006 (UTC)

A "Unix family tree" made some sort of sense back when there were just a few nodes and those were mostly branches. Now that there are dozens of nodes and code has been moving back and forth among most of the recent Unix-like systems, there doesn't seem to be much value in such a graph. — DAGwyn 22:18, 20 October 2006 (UTC)