Talk:Unix/Archive 1

From Wikipedia, the free encyclopedia

[edit] Category for Commands

Should we go through all those UNIX commands (cp, as et al.) and add them all to a Category? I was rather suprised to see all those listed in Wikipedia as articles; it seems more appropriate to host them in Wikibooks: (or leave them at the many online man page mirrors on the Internet).

However, if we're going ot have them here (and I don't see any VfDs on them) then it seems they should be linked into a Category. Ne c'est pas? JimD 02:56, 6 February 2006 (UTC)

Question: which ones? Do we draw the line at Linux programs? What about emacs? Should we restrict entry to Fifth Edition CLI tools, cutting out vi? --maru (talk) contribs 04:59, 6 February 2006 (UTC)

[edit] Pipes

Somehow, some reference to Pipe (computing) seems compulsory. Is there a good reason why it is not included? --4lex 10:35 13 Jul 2003 (UTC)

You are very welcome to include it yourself, just find an appropriate place to mention it. See also: Wikipedia:Be bold in updating pages. --Ap Sun Jul 13 20:31:37 UTC 2003

The Impact section ought to be moved to the top of the article, ahead of history, and someone needs to go over it carefully to make sure it catches everything Unix did; pipes should go in there. Have ken/dmr's original ACM article in hand when doing this, that paper does a first-class job of listing what was novel and key. The section could also use some copy editing. --Noel 11:27, 17 Aug 2003 (UTC)

[edit] History issues

Someone should get ahold of a copy of the early history of Unix article written by ken/dmr that appeared in the Unix BSTJ issue, and check the history section of this page for errors. I fixed many of the worst ones, but I'm still dubious about a few things. Also, I temporarily deleted this line from the end of the history section:

The development of Linux was set upon this back-drop, and one can understand why Linux became so popular (and continues to be so popular) in the face of legal wars over the ownership of Unix.

because it can be read in a way that's not true; the GNU project (which fostered Linux) was started to develop freely available source in the 1980's, long before even Novell bought Linux. The issue was whether code was public domain or not. If someone rewrites it carefully, to say what really happened, and clearly, something like it could go back, but I don't have the energy right now. --Noel 11:22, 17 Aug 2003 (UTC)

Three corrections: First, Novell bought Unix, not Linux. Second, the GNU Project did not foster Linux, having had its own kernel project (HURD); rather, Linux among other Unix-like systems (such as NeXT) made use of GNU's compiler and other software. Last, GNU software is not in the public domain, which would mean that it was not covered by copyright; rather, it is copyrighted but licensed for free source redistribution. --FOo 23:00, 21 Sep 2003 (UTC)

Ooops, you're right, Novell bought Unix, not Linux - that was just a thinko/typo on my part. As for the "public domain", I was being lax in my terminology (I do know all about the CopyLeft); I really just meant "publicly available" (without any definitive specific meaning).

I will however debate the other point, about "GNU didn't foster Unix". Did the "publically available" compiler, utilities etc give Linus (and the many people who've added to Linux) an incentive? I would think they did; even if he didn't know about it when he first started, Linux certainly became part of that whole "open source" conglomerate (which GNU did get rolling) pretty quickly, and I doubt it would have become what it is today without it. I wonder how much Linus knew about GNU when he started? --Noel 22:56, 24 Sep 2003 (UTC)

I'm not sure whether Linus knew about the HURD when he started Linux, though I seem to recall reading that he had not. He came from a Minix background. Linux was begun in '91, when gcc was already getting wide use on proprietary Unix systems as an alternative to commercial compilers. (The first commercial system to use gcc as its native compiler was the NeXT, prototyped in '88 and released in '90.) So gcc and the GNU toolset was probably an obvious choice, if not a forced move.

Linus knew full well what the HURD was - Linux was intended to allow people to run the GNU utilities without having to own UNIX because HURD was taking forever to be released. Thats the whole reason for Linux to exist - the HURD had not been finished.

(Just want to point out that in the original Usenet message announcing the Linux kernel, Torvalds said it "won't be all big and professional like the HURD".) --Maru 05:38, 23 August 2005 (UTC)

I took your expression "the GNU project (which fostered Linux)" as meaning that GNU set out to support Linux in its early stages of development. I'm not sure if that's the case; the GNU people were focused on their own kernel. But you seem to have meant, rather, that GNU software was useful to Linux in starting up. That's certainly true.

Incidentally, the term "open source" was created by ESR and Perens in the late '90s IIRC -- some time after Linux was already becoming popular -- as a "marketing term", an explicitly less political-sounding alternative to the GNU term "free software". --FOo 03:56, 25 Sep 2003 (UTC)

Yes, I did not mean "fostered" in the sense that "GNU set out to support Linux in its early stages of development"; rather, I meant that the whole GNU/FSF movement was a hospitable home to Linux, and helped it develop (much as foster parents help a child who comes into their home).

Then we're in agreement there. --FOo

If nothing else, I think the earliest GNU/FSF work predated Linux - I recall the early stages of the GNU work from when I was still physically at Tech Square, and I'm sure we'd never heard of Linux at that point. It's possible that Linus had already started work in a corner somewhere, independently, but someone would have to compare the exact timings.

"Someone" has already -- Wikipedia. GNU was launched in 1983, Linux in 1991. As I mentioned above, Unix users were using other GNU software such as gcc (and emacs) years before Linux. The GNU kernel, HURD, was also already in development. --FOo

As to "free software", I agree that was poor word choice on RMS' part; people were too likely to misunderstand his real goal. He never meant "free" as in "no charge" (even in the earliest stages, he was reasonably realistic about the economics of the real world :-), he always meant free as in "no painful restrictions" - especially in making the source fully available, for people to fix problems, customize, enhance, etc. I suspect he'd be the first to agree that "open source" is a far better, and more descriptive, term for his real goal. --Noel 12:48, 25 Sep 2003 (UTC)

I did not say that "free software" was a poor word choice. I don't even think so. RMS did not mean "free" as in "no painful restrictions", but rather "free" as in "liberty". He created GNU and coined "free software" in a conscious political response to what he perceived as the growing enclosure by computer companies of software created and developed freely in academia. Recall, that was in the early '80s. "Open source" doesn't enter into it until over 15 years later, when a number of free-software advocates were trying to sell the idea to venture-capitalists during the Internet boom. It wasn't intended to be more descriptive so much as less political, to sell free software to people leery of "freedom" talk. --FOo 13:33, 25 Sep 2003 (UTC)

I created an entry for Oryx/Pecos, which is a specialized flavor of Unix that Bell Labs developed to run AT&T's System 75 PBX switching equipment. Oryx/Pecos is still a "living" operating system, in that it is in use in current switching equipment that's actively supported by Avaya (which only recently abandoned Oryx/Pecos in favor of Linux).

It has a place somewhere in the timeline of Unix development, but I haven't been able to pin down the precise period during which it was developed and initially put into use in the real world. Obviously this would coincide with the introduction of AT&T's System 75, but I can only guess at this being the late 1970s. A web search for "Oryx/Pecos" produces numerous references to a 1985 AT&T/Bell Labs technical document, but the operating system would have been in use several years prior to that. Heiff 08:10, 17 May 2006 (UTC)

[edit] Last non-commercial version

Comment about Version 8 being the last non commercial version of Unix needs to be checked. Xenix was based on Version 7 so was Unisoft's first version of Unix for the Motorola 68000. I believe that version 6 was the last non commercial version of Unix.

Eddie Bleasdale eddie@netproject.com

I would say that 10th Edition was non-commercial(only sold/given to a few universities), and it was the last True Unix release. --Lost Goblin 18:03, 2005 Jun 21 (UTC)

I Was under the impression, that Versions 8-10 were not publicly available (that is outside the Bell Labs). Does anyone have different information? Would be very interested! --M-e-leypold 22:24, 25 July 2005 (UTC)

[edit] "UNIX" vs. "Unix"

Anonymous user from 24.64.223.205 just went through the page and replaced "Unix" with "UNIX" throughout. This seems unnecessary. Both spellings are widely attested (even in Unix manuals!), but the original is "Unix" -- which also seems to be more popular among those who actually use Unix. --FOo 11:19, 3 Oct 2003 (UTC)

I agree that Unix is more popular and that it should probaly be Unix, however the last para (or so) says that UNIX is the ingrained name, so I'm not sure whether to leave or revert... --Dysprosia 11:28, 3 Oct 2003 (UTC)

I think that Unix should be preferred, like in the Jargon file and elsewhere. --Shallot 13:39, 12 Dec 2003 (UTC)

The trademark is UNIX. http://www.unix-systems.org/ uses UNIX consistently. The UNIX Programming Environment by Brian W. Kernighan uses UNIX. So, it seems to me that the formal spelling is UNIX, but there sure is a lot of use of the alternative Unix as well when you do an Amazon search for books with UNIX in the title, or just sample the newsgroups by doing a Google.com search for UNIX. A hard call! --Bevo 18:34, 12 Dec 2003 (UTC)

For reference, here is the relevant excerpt from the Jargon File:

Some people are confused over whether this word is appropriately `UNIX' or `Unix'; both forms are common, and used interchangeably. Dennis Ritchie says that the `UNIX' spelling originally happened in CACM's 1974 paper The UNIX Time-Sharing System because "we had a new typesetter and troff had just been invented and we were intoxicated by being able to produce small caps." Later, dmr tried to get the spelling changed to `Unix' in a couple of Bell Labs papers, on the grounds that the word is not acronymic. He failed, and eventually (his words) "wimped out" on the issue. So, while the trademark today is `UNIX', both capitalizations are grounded in ancient usage; the Jargon File uses `Unix' in deference to dmr's wishes.

And from FOLDOC:

"Unix" or "UNIX"? Both seem roughly equally popular, perhaps with a historical bias toward the latter. "UNIX" is a registered trademark of The Open Group, however, since it is a name and not an acronym, "Unix" has been adopted in this dictionary except where a larger name includes it in upper case. Since the OS is case-sensitive and exists in many different versions, it is fitting that its name should reflect this.

I think that those rationales make sense. Also, it's probably not particularly advisable to use a trademarked string in just about every sentence that may or may not be in accord with the trademark holder's wishes... --Shallot 18:58, 12 Dec 2003 (UTC)

A note with regard to Unix v. UNIX - since the most recent events in the SCO v. IBM Linux Lawsuit, a rough convention exists; UNIX has been used to signify that the operating system in question is compliant with the Open Groups' Single UNIX specification standards for UNIX operating systems, while Unix has been used to signify that an OS contains (or did once contain) code from the ancestral code base. On a related note *nix has come to signify operating systems that work like UNIX, but aren't by code and Open Group certification. This is mainly a Groklaw usage, but it makes it easier to write about the various flavours.

In case it needs to be explicated, we seem to have picked up on the same convention, except that we don't use *nix but Unix-like. --Joy [shallot] 18:53, 18 Jun 2005 (UTC)

As Examples, AIX (IBMs UNIX), is both a Unix and a UNIX. The BSDs are Unix (because they include or were built upon ancestral code - see USL v. BSDi), but most are not UNIX as they haven't been certified. Linux isn't Unix (presuming innocence of the defence in SCO v. IBM), because it has no ancestral code, but some (very few) versions have been UNIX, as they conformed to Open Group specifications. Linux is a *nix. Confusing, isn't it? --SkArcher 23:05, 28 Jun 2004 (UTC)

I lean greatly in favour of UNIX as it is the correct term (as mentioned above referencing the trademark) and this is an encyclopedia! I believe Unix is more popular in the UNIX community because they get slack of writing the correct name. It is worth the extra effort. If this is decided, then I also think that the redirect should be Unix->UNIX, not the other way around as it is currently to keep things uniform. --Avochelm 10:26, 9 Mar 2005 (UTC)

Calling it "the correct term" is just begging the question, since obviously we want to give the article the correct title. The question here is, which is the correct term? This is especially tricky since several different things are being referred to. The Groklaw distinction between genetic-Unix and trademark-Unix (or UNIX) is worthwhile ... however, there plenty of people who use the word "Unix" to encompass Linux and other systems which are neither. It is probably not our place to call their usage "incorrect", although we may certainly say that it is not the usage preferred by the Open Group.

As it stands, the most common usage among Unix users is "Unix". I hold that we can and should safely continue to use this expression (and certainly not the affected jargon "*nix" or "*n*x") for this very reason. --FOo 16:07, 9 Mar 2005 (UTC)

What is your source for writing 'most common usage among Unix users is "Unix"'. besides do you mean applicaion users or system admins or system programmers? as an old UNIX hacker I would say that UNIX is correct because it was used in the first release of the manual PROGRAMMER'S MANUAL by K. Thompson and D. M. Ritchie ] (November 3, 1971) and has been used consistently by those with more than a passing interest in UNIX for decades. the problem comes in that to be consistent it should have been unix because everyone knows that using shit keys slows down typing and unix programmer are very lazy when it comes to the effort of typing. looking at dennis m. ritchie bell lab pages it seems to me that dmr was using UNIX in the techical documents for the propeller heads and Unix in his presentations to management etc. so i guess that it dependes on which end of the pipe one is looking at UNix as to which is correct. --Philip Baird Shearer 21:14, 10 July 2005 (UTC)

It's not an acronym, and dmr wishes it were Unix, so that is correct. --ssd 12:45, 26 July 2005 (UTC)

Indeed - AT&T adopted the uppercase "UNIX" long after it was called Unix. You cannot say anything published at a Bell or AT&T web site is as Thompson or Ritchie wishes, because those companies enforce(d) certain formatting (for various/commercial reasons), namely "UNIX". ¦ Reisio 14:33, 8 October 2005 (UTC)

To sum up - anything that is "Unix" or "UNIX" is Unix - it is made UNIX only if it is a commercial product that endeavors to please the trademark holders, or if that format of uppercase letters is specifically and consistently used as a name (as AT&T adopted). AT&T's Unix would appropriately be referred to as Unix, except when referring to a specific named version which they call UNIX. A UNIX is still Unix, but a Unix is not necessarily UNIX. Clear? MMmmkay. ¦ Reisio 14:24, 8 October 2005 (UTC)

Good summary, I agree. Just remember that it's not an acronym and it'll all make sense. --DylanW 07:23, 23 January 2006 (UTC)

Nevermind my last comment about it not being an acronym . . . I just read more about it and learned for the first time that it's a modification of an acronym. Hmmmmmmm . . . now I don't know what to think. --DylanW 09:38, 23 January 2006 (UTC)

The wiki can do small text just fine, along with the registered trademark symbol. Clearly UNIX® is correct. 24.110.60.225 06:16, 5 March 2006 (UTC)

What is UNIX® see http://www.unix.org/what_is_unix.html which states "The Open Group holds the definition of what a UNIX system is and its associated trademark in trust for the industry. " Johnmarkh 18:23, 19 April 2006 (UTC)

[edit] "Classic" / "canonical" UNIX commands?

This seems to be just a listing of common commands in *NIX distributions these days. "Classic" to me would imply coming from old AT&T, SYSV, or BSD Unices, but this list includes GNU innovations like less, bash, tcsh, and the modern ssh. Maybe the section should be renamed.

I agree, and have done so. It is now Canonical UNIX Commands. --User:Jeffwarnica, 22:13, 26 Sep 2004

I've since removed bash and zsh, they're late 80s stuff and not really considered canonical. Tcsh is a fair bit older than both and written by an original csh guy, it should probably stay. I also dropped sam and less. I'll keep SSH because it's become fairly canonical over the years. --Joy [shallot] 19:00, 18 Jun 2005 (UTC)

I think some of those choices sound rather arbitrary, specially I think it's unfair to leave sam out having in mind that it was developed at the Labs and it's to this day used by many of the original Unix developers(Ken, Dennis, Brian and Duff are all known sam users, and I'm sure there are more) also I would add the rc shell as it's the official successor of /bin/sh. --Lost Goblin 21:47, 2005 Jun 20 (UTC)

I have been a user of Unix from 1977 to now, and I've never heard of sam. It apparently only runs under GUIs with mice, so I don't see how it can possibly be 'canonical', regardless of how popular or how good it is. --Macrakis 22:28, 20 Jun 2005 (UTC)

Sam was developed during the early 80's at Bell Labs for the first bitmaped Unix terminal and has been used by most of the Unix team since then, it's roots go back to sed, which is the original Unix text editor. I think it's much more unix-worthy than Emacs with was developed outside Unix and only later on ported to Unix, and it has never fit with the way Unix works. --Lost Goblin 11:50, 2005 Jun 21 (UTC)

If the sam article is correct, sam wasn't developed for Unix, either, but for Plan 9. Although I am a longtime and heavy Emacs user, I would agree that Emacs is not a "classic" Unix command. First of all, it isn't really a "command" but an interactive system. Second, it is cross-OS (I have used it on ITS, Tenex, TOPS-20, various Unixes, and Windows). As for sam being rooted in sed, I don't think that's relevant -- after all, Perl is rooted in sh and awk, but is hardly "classic". --Macrakis 15:56, 21 Jun 2005 (UTC)

The sam article was correct but unclear (I fixed it now), sam was originally writen for the DMD 5620 windowing terminal which ran Unix in the early 80's, Plan 9 was not started until years later, but as the Unix team at bell labs moved to Plan 9, sam was ported to the new system, ironically most sam users this days(like Tom Duff use an X port of the Plan 9 version. Also sam is a more clear descendant of ed than perl of awk and sh, Ken Thompson the original author of ed was very involved in the design of the sam command language. Lost Goblin 17:56, 2005 Jun 21 (UTC)

I used pretty much the same reasoning as Macrakis when considering sam. I really don't want to get into an Emacs bashing game, but fact is that it has existed since 1976 on OSs older than Unix (while Plan 9 is younger than Unix) and that it is much better known than sam. And you can't really include vi without including emacs :) --Joy [shallot] 17:03, 21 Jun 2005 (UTC)

This dialog only highlights the need to classify properly. There is at any given time a command set which is considered by many users to be "classic", but this changes with time and doesn't really correspond to any historical reality. For an encyclopedia article we should identify the historical development of the toolset, at least from the Bell Labs V7 UNIX onwards. UNIX V1. in PDP7 assembler, already had ar, as, cat, cp, df, du, ln, ls, mail, od, su, wc and who, all of which are in everyday use in modern Unixes. Fourth edition in '74, by which time UNIX had been rewritten in C, introduced pipes, which more even than the hierarchical filesystem, gave UNIX its characteristic flavor and geek-friendly reputation [1]. V7 brought us K&R C, uucp and Bourne shell. sccs and named pipes were introduced at Bell Labs as part of Programmer's Workbench {PWB) [2]. Somewhere along the line "make" appeared... --Tony Sidaway|Talk 12:28, 21 Jun 2005 (UTC)

Absolutely. The definitions need to be clearer. --Macrakis 15:56, 21 Jun 2005 (UTC)

The word "canonical" also needs to be qualified better, almost as much as "classic" does. --Joy [shallot] 17:03, 21 Jun 2005 (UTC)

[edit] UNIX vs Linux/FreeBSD/etc

Perhaps a section from someone more knowledgable than myself should be added regarding the differences between UNIX and UNIX-like operating systems. From my experience in both UNIX and Linux environments, UNIX-like operating systems seem to not follow UNIX standards as strictly as Linux does! What inner-architecture advantage (be it kernel, file system, etc) does UNIX have over its clones that still makes UNIX popular among companies and organizations who operate their data systems on mainframe computers? --69.234.183.71 19:42, 17 Mar 2005 (UTC)

AFAIK the most important difference is usage of the Unix trademark. It would be reasonable to call the modern BSD-derived platforms Unix; they are after all Unix-derived. But they don't call them Unix. Because using the Unix trademark requires certification and money. Why bother? Of course there's also the argument that commercial stuff (like Solaris) is more scalable and supports the big iron better, but I don't have any experience with assessing that claim. --Anonymous 10:42, 5 Jul 2005 (UTC)

Solaris is indeed more scalable and does support big iron better; that is empirically demonstrable. Solaris's SMP core natively scales up to hundreds and thousands of processors. Are there any standard Linux distributions that can run reliabily on more than an 8-way system? --unsigned

Actually, Linux is replacing the UNIX® blood line in business. Even the dollar figures, a rather poor way to measure things, are moving in Linux's direction. Standard Linux distributions run reliably on 32 or 64 processors, limited only by an optimization related to the number of bits in a long. Up to 1024 processors have been used. 24.110.60.225 06:24, 5 March 2006 (UTC)

There is a difference between "reliably won't panic" and "equally scalable"; not panicing is pretty simple. Scaling reliably requires making sure that the kernel mutex locks are all sufficiently fine grained enough, and the multithreaded kernel sections wide and robust enough, that you can actually use those extra processors. Having 64 processors which are all waiting on the same I/O mutex at the same time is a great way to waste someone's money.

Linux is a lot better than it used to be, but the commercial UNIXes have been doing large system many processor stuff for a lot longer, with a lot more of the brainpower who actually understand high end SMP system behavior and code optimization (there aren't enough of those people, and big companies tend to snap them up).

There is a lot of Linux in industry, and its gaining market share. Those are true. For most applications, most system roles, it's completely adequate and functionally equal to the competition. For truly big iron... even if Linux came as an option for a Sun F25K instead of Solaris, I wouldn't use it for my 64 processor database. I've tested, specced out, and been involved in the order process for thousands of systems over the last 5 years; the last large order I did was several hundred 2 CPU Linux boxes, for a bunch of applications that scale horizontally every day of the week and twice on sunday. But given the choice between debugging and implimenting Oracle 9i RAC on 40 servers for a world-class database, or a F25K and Solaris... Sorry, charlie.

Maybe next year. More likely 2010; the amount of work it takes to make the kernel that scalable is a lot, and Linus is somewhat averse to going down that road when not that many total systems are in that corner of the space, last I heard. Georgewilliamherbert 07:30, 5 March 2006 (UTC)

[edit] "Unremarkable"

4.228.216.247 removed the following sentence from the start of the History section:

As is often the case with developments that go on to become highly popular and influential, the beginnings of UNIX are unremarkable.

I agree that it's basically opinionizing (editorializing if you prefer). In fact, the statement is POV. Personally I don't find the beginnings of Unix at all unremarkable. The concept of a kernel plus tools appears to have been present from an early date, as was the then-revolutionary concept of a hierarchical filesystem, apparently borrowed from MULTICS. The "can do" atmosphere at Bell Labs at the time also undoubtedly was remarkable. --Tony Sidaway|Talk 13:35, 18 Mar 2005 (UTC)

Either the statement is blatent POV with little or no suporting evidence in which case it needs to be removed no matter how long it has been here, or it is a quote from another source on unix history in which case it has merit but needs to be removed because it is plagerism (or credited proerly). --ssd 05:16, 23 Mar 2005 (UTC)

Or you just don't want to accept the fact that, when Unix started out, it really WAS no big deal, and was just written so that the principal authors could continue playing their computer games.

This is NOT to diminish the tremendous contributions that Unix has made, but its origins, whether you like it or not, were humble. Not exactly an animal's manger, but similar. --Atlant 14:59, 23 Mar 2005 (UTC)

First, I don't see how calling something's beginnings unremarkable really adds anything to the article. Secondly, it is obviously a very POV opinion, true or not. Is the opinion credible? If so, credit it. If not, it's worthless and needs to go. --ssd 03:42, 13 May 2005 (UTC)

I agree, the whole statement(starting with As is often...) does not add any value to the article and should be removed Lost Goblin 20:42, 2005 May 14 (UTC)

[edit] SUPER-UX and UNICOS?

No mention of SUPER-UX and UNICOS?

Which should be included why? --Maru 05:38, 23 August 2005 (UTC)

Never heard of SUPER-UX, but UNICOS is semi-significant -- it was one of the standard operating systems on a lot of Cray machines back in the day. --Haikupoet 04:15, 30 August 2005 (UTC)

Well, even so, I doubt it's worth inclusion unless there was some novel hardware/software feature. Strike me if I'm wrong, but Cray machines were all mainframes, and Unix has never been particularly strong on mainframes. (I think z/OS and other proprietary OSs have always had the vast lion's share of the market.) --Maru 04:29, 30 August 2005 (UTC)

Hm, I don't know that I'd consider a Cray a mainframe per se -- I think of a mainframe as essentially being a superserver, meant for pushing data around, whereas the Crays were supercomputers, more meant for high-speed number crunching. That is not to say that there isn't some overlap, but apparently for a while before Cray's demise UNICOS was actually the primary OS for their later systems. Haikupoet 04:44, 30 August 2005 (UTC)

The Cray systems were generally called supercomputers which has traditionally been held in a different category than mainframes. Specifically the term, as applied to Cray computer was mostly focused on their vector processor-based systems. (Read the linked articles to compare the terms). Regarding the orginal observation: I wouldn't suggest that we make this article an exhaustive list of every version of UNIX that's ever been. UniCOS has its own article which points back to Unix-like which, in turn, points back to us; and a search on UNIX and Cray will find it at the top. JimD 02:51, 6 February 2006 (UTC)

[edit] Bash shell

Joy, I put in the proper link to the bash shell. The original one was a redlink. Please test the link before you revert my edit. --Gbeeker 15:14, 19 August 2005 (UTC)

I reverted an edit that had produced a red link. Neither is red any more, so this is all moot. --Joy [shallot]

[edit] Unix Filiation Image

The new image posted 4:35, 27 August 2005 by User:Camje lemon doesn't show AIX or IRIX. Can these be added back into the image? --Gbeeker 03:32, 29 August 2005 (UTC)

I'm going to revert back to the old image, which includes AIX. Perhaps someone with image editing expertise can merge the 2 images. Gbeeker 17:37, 14 September 2005 (UTC)

The current diagram comport some errors (Unix Time Sharing System to HP-UX, and so, without Unix System III and V for example.), and in my opinion AIX or HP-UX didn't have a real role, i don't remember they invented concepts used by others UNIX systems.
Can you revert my diagram while waiting for other opinions ?
Thanks. CAMPESATO Jean-Baptiste / camje_lemon 09:07, 15 September 2005 (UTC).

In my opinion, both AIX and HP-UX have a significant presence in the current Unix market. Many research firms state that the top selling Unix systems are IBM's pSeries (AIX), HP's HP-UX, and Sun's Solaris. Each Unix belongs in an an article on modern Unix in an encyclopedia. If you like, please make the corrections you feel are appropriate to the existing diagram. I welcome other comments on the issue, but for 2 weeks no-one has commented, so I felt it ok to revert to the more inclusive diagram. I cite a source from the current article.

AIX, FreeBSD, HP-UX, Linux, Solaris, Tru64. UNIXguide.net. Retrieved on March 17, 2005.
And a Press release about IBM pSeries. Gbeeker 11:45, 15 September 2005 (UTC)

First; The current diagram have some errors, and it's in catalan,
Second; The current diagram is less precise and complete than mine contrary has what you said: where are the various BSD' branches? Where are System III / V? How do you want to display HP-UX or AIX or Solaris without System III and V?
Thirst; Why XENIX is stop in this diagramm ? XENIX is always in development (And we say XENIX, and not Xenix OS, and now OpenServer). SunOS is not Standford University. And so... (What's Univel in UnixWare ?),
Fourth; sure they have "significant" presence. that doesn't mean they influenced the development very much.
What role they played please?
I think that it's better to show a diagram of the UNIX having a role in this fabulous history, rather than those having a commercial presence.
CAMPESATO Jean-Baptiste / camje_lemon 15:19, 15 September 2005 (UTC)

It would be great if you or someone with image editing experience could make the changes you are suggesting. The current image shows the BDS family of OSs on the top, and the System III and System V family decendants on the bottom. As far as a role they play, the article is called "Unix", and not "the historic signigicance of unix systems". Anyway, I'd say that the LVM system and ODM created in AIX is historically significant, and HP-UX's tie to the highly dependable HP hardware is significant. As I mentioned before, market share is also significant, and AIX, HP and SUN are still very much the top vendors. Gbeeker 17:26, 15 September 2005 (UTC)

We cannot speak about Unix by neglecting his history. The purpose of a chronological diagram, is to tell history, not to number commercial UNIX.
For the LVM i don't know but:
I repeat, it's not because a manufacturer sells his UNIX, this UNIX is important. I always wait for proofs of important concepts invented by HP-UX and wich contributed to the other operating systems.
If by magic you read what I said, how do you want to correct so many errors?
Please, Let the others to read what I told to answer.
Bye. CAMPESATO Jean-Baptiste / camje_lemon 19:46, 15 September 2005 (UTC)
PS:[My diag]

[edit] Random thoughts on merging with Unix-like

The necessary course of action is a little more complicated than a simple merger -- I think there's a place for both articles, honestly, though Unix-like needn't be much more than a stub. The real problem is that despite the Open Group's attitude towards the trademark issue, the issue of what is and isn't Unix is a very muddy issue. In other fora there has been a distinction between "genetic Unix", "branded Unix", and "Unix-like" which seems to be the most valid way of making the distinction. For those unfamiliar:

Genetic Unix refers to any Unix-like system with a connection to the original AT&T codebase. That would include the BSDs and (naturally) all Ancient Unix variations as well as most commercial Unices.
Branded or Trademark Unix refers pretty much exclusively to commercial variants of Unix, but that's largely a market issue as the branding is theoretically available to any entity willing to pony up for the testing. The important thing to note is that a Branded Unix is not necessarily derived from a Genetic Unix codebase -- IBM's z/OS, I believe, carries the brand, for example, though its Unix layer is only a small part of the OS.
Unix-like refers to workalikes, systems such as Linux, Minix, QNX, and GNU Hurd. They for the most part behave like Unix, sometimes even using the same ABI, but are based on non-AT&T code bases for the most part. Dennis Ritchie is one of those who considers such systems to be Unices in fact if not in brand, but this could be a controversial point.

Any attempt to merge the two articles would require acknowledging these three strains of the Unix world and how they interact and even overlap. Haikupoet 04:48, 31 August 2005 (UTC)

I have added the above, rewritten a bit to use Eric Raymond's original terminology, to Unix-like. Haikupoet 04:25, 1 September 2005 (UTC)

[edit] Splitting away history

The article is already quite large, and if Unix-like merged, it will be too large. I think, History of Unix should be extracted into a separate article --MvR 12:55, 30 September 2005 (UTC)

Good idea. &amp;brvbar; Reisio 16:02, 30 September 2005 (UTC)

[edit] UNICS and eunuchs

The article says: "Following bad puns of Unics (homophone of eunuchs) being a castrated Multics, the name was later changed to Unix." However, as I read it in other places the pun was intended. Does anybody know more about this? --Qwertyus 00:06, 27 October 2005 (UTC)

Bad puns can be intentional. And I think the original original name was Eunics (I'm taking this from Andrew Tanenbaum). --Maru (talk) Contribs 00:19, 27 October 2005 (UTC)

I'm pretty certain it was Unics, no e -- that's what most of the sources cite, including the Jargon File IIRC. The accepted etymology is indeed a pun on Multics -- if you add an e to the word it has no obvious meaning anymore. Haikupoet 04:26, 27 October 2005 (UTC)

I have no doubt that the pun on Multics was intentional, but the pun on eunuchs? Qwertyus 10:05, 27 October 2005 (UTC)

[edit] UNICS vs UNIX

Hello,
I would like to say that I saw a lot of people who spelt the name UNICS but this name doesn't exist:
A lot of people called the PDP-7 version UNICS, but this version was called Ken's new system (Or something like). UNICS was proposed for name by Peter Neumann (Apparently. In 1970, so ~6 months after the PDP-7 version if it's correct) but it was never used and it would seem that the modification to UNIX came from Kernighan (But he said it's not him... Nobody remember who modified the name). The OS is called UNIX since the first edition.
To discuss.
Cordially,
CAMPESATO Jean-Baptiste / camje_lemon 22:34, 28 October 2005 (UTC).

What're we supposed to discuss? Whether saying it was ever called "UNICS" is okay? You just said it was called UNICS. ¦ Reisio 23:48, 28 October 2005 (UTC)

I said no version had been called UNICS. And even Dennis Ritchie doesn't remember to have used UNICS. Can we say UNICS if this name is only one intermediary which was not used? (It's about this question I'll like to discuss ;).) CAMPESATO Jean-Baptiste / camje_lemon 07:31, 29 October 2005 (UTC).

This is how I see it:

the jump from "Multics" to "Unics" makes much more sense than straight to "Unix"
Kernighan supports "Unics" coming first, and that it was his suggestion [3]
Ritchie supports Kernighan coining the term [4]
I have not yet found any respectable source (interview, claim by anyone that matters) that supports Peter Neumann having coined the term, and his verbose homepage mentions nothing about it [5]
it was most commonly "Multics", not "MULTICS", so it therefore makes sense that it was "Unics", not "UNICS"
Ritchie supports that it originally morphed into "Unix", not "UNIX", but that Bell didn't want to change it back after a typesetting incident [6]

So, as I see it: Multics -> Unics (Kernighan) -> Unix (nobody recalls) -> UNIX (typesetting stumble)

I'll continue to look into Neumann, but so far I've found nothing conclusive...just a lot of articles (that seem to be duplicates of one article) with no sources cited spouting the same vague claim. These claims often include that Kernighan switched "Unics" into "Unix", but Kernighan claims nobody remembers who did that; so the source of the information used in these articles is that much more questionable. Kernighan could of course be lying or misremembering a lot, but I doubt that. :p Ritchie would also have to have been not paying attention if Kernighan didn't coin it.

¦ Reisio 10:53, 29 October 2005 (UTC)

Thank you for the Kernighan's Interview. I'll continue to search too; When I mailled Ritchie (1month ago) he said that he doesn't remember ever using "UNICS", he said they called the system "Ken's new system" (PDP-7 Version) and he thinks that Kernighan came with UNIX (Since 1st edition).

Finally in all the cases it would seem that UNICS is an intermediary of UNIX but that no version had this name. I hope we'll find exactly the real history of the name.

For information I read that Peter Neumann invented "UNICS" in "A Quarter Century of Unix" (Peter H. Salus).

Now I see the history like that: 1969 Release of the PDP-7 Version with the name "Ken's new system" ---> 1970 Kernighan created UNICS but this name wasn't used ---> ?1970? Somebody modified UNICS in UNIX ---> 1971 Release of the First Edition with the name UNIX.

Cordially, CAMPESATO Jean-Baptiste / camje_lemon 11:43, 29 October 2005 (UTC)

Ah, I didn't think there was any doubt that officially (after it was far enough along it really started getting to other people) it'd always been 'Unix' or 'UNIX'. ¦ Reisio 15:54, 29 October 2005 (UTC)

Yeah I'm okay since the begining of the discution about there is only UNIX, but I began this discution because I saw a lot of persons who are using UNICS (See on google) instead of UNIX (and to clarify the history of the name). CAMPESATO Jean-Baptiste / camje_lemon 19:05, 29 October 2005 (UTC).

I think you people are taking this way too seriously and are getting confused because I don't think the folks at the Labs cared that much about carefully naming everything. Lost Goblin 21:53, 29 October 2005 (UTC)

I think you should realize that this is text, and any emotion you see in it you're attributing to it yourself. ¦ Reisio 18:46, 31 October 2005 (UTC)

[edit] Major commercial Unixes in the 2000s

I added HP's HP-UX to the list of major commercial Unixes still surviving in the 2000s, since it consistently ranks in the top three shares of the server market (along with AIX and Solaris). I also re-ordered the list of manufacturers alphabetically to avoid the appearance of bias. Unfortunately, that puts HP first in the list, and I currently work for HP, which is problematic. Perhaps they should be listed alphabetically by OS name (putting AIX first), or ranked according to factual data such as creation date. The whole paragraph has NPOV problems, lacking both sources and objective criteria for the surviving OSes. Bradd 23:29, 8 November 2005 (UTC)

[edit] Unix architecture

Having read this whole talk page, I'm left wondering if it wouldn't be better to rename UNIX architecture as Unix architecture and expand it where it is. If not, then I definitely think there needs to be an explanation somewhere of what attributes Unix-like systems have that make them like Unix. Is it the file system? Is it the kernel/user space distinction? Is it pipes? Is it all of the above and more? Tarheelcoxn 00:12, 21 December 2005 (UTC)

I think for now that'd be fine. We should probably get stuff organized in this article before we try and branch out. ¦ Reisio 13:34, 21 December 2005 (UTC)

Rename it as what? Sorry, that didn't make sense to me, can you explain? Thanks. Georgewilliamherbert 21:28, 22 December 2005 (UTC)

Rename UNIX architecture as Unix architecture. Tarheelcoxn 01:33, 23 December 2005 (UTC)

Does anyone still think Unix architecture should be merged here? Ideogram 02:29, 3 June 2006 (UTC)

[edit] mentioning the B programming language

Comparing the UNIX article with the UNIX course book from my school, I have seen that the programming language B is not mentioned in this article.

It was a language invented by Thompson at Bell labs (before 1974), written not to have to re-write UNIX from scratch in assembler every time a new machine came out. B was a simplification of the BCPL language, which was based on CPL.

By the way, I don't believe this is true. I've never read it anywhere. It might be true for C. Qwertyus 11:38, 7 January 2006 (UTC)

Later, Dennis Ritchie would have developed the C language as a successor for the B language. Afterwards, Thompson and Ritchie re-wrote UNIX in C. Bernard François 11:50, 27 December 2005 (UTC)

See B programming language, C programming language. Qwertyus 08:06, 28 December 2005 (UTC)

What I meant is that it should be mentioned in the text, because the B programming language and the early development of UNIX are related. I didn't mention it myself, because I'm not a UNIX history expert (I just read that information in my course book, so I thought it might contribute to the completeness of wikipedia...). If someone who can confirm the bound between the B programming language and UNIX reads this, please add this information to the article. Bernard François 07:13, 7 January 2006 (UTC)

B led to C, and Unix was rewritten in C. But AFAIK, no significant part of the OS was ever written in B. IMHO it's not important enough to mention it explicitly here. Qwertyus 11:37, 7 January 2006 (UTC)

[edit] The chart is wrong

The chart shows SunOS as a descendant of BSD. This is correct.

But then it shows Solaris as an another BSD derivative. This is not correct, Solaris was a switch from BSD family to the System V family.

A chart is harder to change than some text, thus I leave this note here. 70.23.236.53 15:20, 3 January 2006 (UTC)

Solaris versions also have a SunOS version number IIRC. So it is a littlle more complex. Rich Farmbrough 02:01 25 March 2006 (UTC).

[edit] UNICS capitalization

Sources: The Jargon File entry on Unix Andrew Tanenbaum's tangental account of the origins of UNIX, including his time at Bell Labs Unix History page quoting Peter Salus' A Quarter Century of UNIX; Reading, MA: Addison-Wesley, 1994

UNICS was an acronym (UNIplexed Information and Computing Service). Acronyms get capitalized. UNIX is a pseudo-acronym. But UNICS was, in its short lifetime, an acronym.

Of these, the Jargon file, going back to 1976, is probably the most authoritative source. It was regularly updated by principal sources contemporaneously with the activities.

Georgewilliamherbert 22:44, 11 January 2006 (UTC)

UNICS was an acronym (UNIplexed Information and Computing Service).

So was Multics, but they didn't capitalize it.

The Jargon File entry on Unix

It makes no sense that they'd go from "Multics" to "UNICS" to "Unix". Going to take more than an uncited remark on a page a twitchy guy like ESR is in charge of to trump that. Even if that text is straight from 1975, that's about five years time in which the formatting could've been screwed up. Consider Ajax; coined just under a year ago and already being plastered as "AJAX" all over the place.

Andrew Tanenbaum's tangental account of the origins of UNIX, including his time at Bell Labs

Tanenbaum had nothing to do with it, and that page uses "MULTICS" consistently, which is plainly wrong.

More sources above in #UNICS vs UNIX. ¦ Reisio 01:56, 12 January 2006 (UTC)

All the contemporaneous sources used UNICS. You admitted up in the #UNICS vs UNIX debate that Multics was usually not, but sometimes, capitalized. The 1970s sources I have seen, and the 80s and 90s sources when I was becoming active in the industry, used all caps as a general rule. UNIX was also generally all caps in those timeframes, before the X/Open snafu. I am a primary source as to the usage from the mid-80s onwards. People looked at you strangely when you used Unix back then, and I generally didn't see MULTICS as anything but all caps.

You keep arguing it makes no sense - to you. That is unfortunately irrelevant. Multics and MULTICS were both in use at the time preceding the Unix original research. There's no record disputing that UNICS was an acronym. UNIX versus Unix is a different question: we know from primary sources that it was originally typically lowercase, until the all caps typesetting experiment.

Tanenbaum is a secondary source, who was at Bell Labs reasonably soon after the invention of UNIX, and worked with the group in question. That is not nothing to do with it.

Calling ESR a twitchy guy is irresponsible research. ESR was the custodian of the UNIX / Internet shared cultural heritage dictionary for quite a long time. The written Jargon file is a major source of reference material into common usage and was a contemporary source back to its origin.

Peter Salus, whose material you haven't seen fit to rebut, goes back about that long himself, and has been researching the history of UNIX and the Internet in depth for the last decade. While not perfect, he has generally gone to great lengths to source material properly, get primary sources to talk in interviews, and document in depth.

None of the sources above rebut the sources I quote.

You cannot go around substituting what makes sense to you for what contemporary and nearly contemporary sources say about the event. The case that Unix is properly mixed case is well grounded, though it was more typically all caps in the 80s and early 90s. The evidence on MULTICS is consistent when it's contemporaneous and only trends towards mixed case again when the common usage of UNIX started to shift back towards predominantly Unix.

Based on the available contemporary sources, you have no case. Georgewilliamherbert 02:48, 12 January 2006 (UTC)

All the contemporaneous sources used UNICS.

Let's see them, then.

You admitted up in the #UNICS vs UNIX debate that Multics was usually not, but sometimes, capitalized.

Naturally - we wouldn't be having this discussion if everyone stuck to correct formatting. :p

The 1970s sources I have seen, and the 80s and 90s sources when I was becoming active in the industry, used all caps as a general rule. UNIX was also generally all caps in those timeframes, before the X/Open snafu. I am a primary source as to the usage from the mid-80s onwards. People looked at you strangely when you used Unix back then, and I generally didn't see MULTICS as anything but all caps.

Well then kudos to people, then, because nowadays "Multics" dominates. Now we just need to clear up this "UNICS" issue.

You keep arguing it makes no sense - to you. That is unfortunately irrelevant. Multics and MULTICS were both in use at the time preceding the Unix original research.

Then all we need is a source saying whoever coined "Unics" meant it in caps or not. The only close one I have handy is this one which uses "Unics", not "UNICS".

There's no record disputing that UNICS was an acronym.

It wouldn't matter to this discussion if there were, because I'm not disputing "UNICS" being an acronym or not. I have hinted, however, at the fact that not all acronyms (including the one "Unics" is derived from) are standardly presented in all caps.

UNIX versus Unix is a different question: we know from primary sources that it was originally typically lowercase, until the all caps typesetting experiment.

Yes, it is a different question - why are we talking about it?

Tanenbaum is a secondary source, who was at Bell Labs reasonably soon after the invention of UNIX, and worked with the group in question. That is not nothing to do with it.

When the three primary people involved can't even say who switched it from Unics to Unix, that doesn't leave much hope that people who came roughly five or more years later - like Tanenbaum - will be privy to that class of info.

Calling ESR a twitchy guy is irresponsible research. ESR was the custodian of the UNIX / Internet shared cultural heritage dictionary for quite a long time.

...which is why the largest section at Eric S. Raymond is about criticism of him.

I can't rebut what I haven't seen; the page you listed is just a blurb compiled by who knows at a university in Georgia.

The book isn't online fully; it's a book. It's hard to provide a URL into printed text material. I provided the biblio entry info. Georgewilliamherbert 02:21, 13 January 2006 (UTC)

None of the sources above rebut the sources I quote.

I think they rather do.

You cannot go around substituting what makes sense to you for what contemporary and nearly contemporary sources say about the event. The case that Unix is properly mixed case is well grounded, though it was more typically all caps in the 80s and early 90s. The evidence on MULTICS is consistent when it's contemporaneous and only trends towards mixed case again when the common usage of UNIX started to shift back towards predominantly Unix.

Then we're in agreement that "Multics" and "Unix" (neither all caps) are best? - What's the problem with "Unics", then?

Actually, no, we aren't in agreement that "Multics" and "Unix" are best, either. The common usage in the late 70s through mid 90s was more all caps than not. The original Unix was mixed case, but they started capitalizing it as a general rule over 30 years ago. What's correct depends on your interpretation of whose usage matters more. The people who started capitalizing it are also the originators, and their all caps standard was the standard for literally two decades, until X/Open became a problem. People looked at you funny if you typically used mixed-case fifteen years ago.

With the usage of Unix, it has contemporaneously shifted, and there is good evidence that the mixed case to predominantly all caps shift was after it was originated, in the 74 article, when they got too happy with the macro package. Looking to original intent, and current usage, one can argue coherently for mixed case. Looking to how it was commonly used for 20 of its 35 years, it should remain all caps.

Concluding that either way is the only proper way is clearly wrong. It's not that simple. There may be a justifyable case that Wikipedia should be consistently one way or the other, except for alternate usage explanations. You may have noted, I am not reverting your UNIX->Unix -ification.

But, do I agree with a statement that mixed-case is correct in an absolute sense? No, you're grossly simplifying what is in fact a complicated issue. Georgewilliamherbert 02:21, 13 January 2006 (UTC)

Based on the available contemporary sources, you have no case.

Everyone's entitled to an opinion. Shall we go ahead and proceed to Wikipedia:Requests for comment? ¦ Reisio 03:41, 12 January 2006 (UTC)

We can do that, or someone can attempt to contact Brian Kernighan (preferably) or Ken Thompson or Dennis Richie and ask them. That presumably would be somewhat more authoritative than even the results of a Rfc. Georgewilliamherbert 02:21, 13 January 2006 (UTC)

I got my thumb out, and have fired off a query to Brian Kernighan, Ken Thompson, and Dennis Richie (with a cc to Peter Salus). I don't know if they will have the time or interest to respond, but if any do that should be more authoritative than any more rounds of this debate. Georgewilliamherbert 03:40, 13 January 2006 (UTC)

Indeed. ¦ Reisio 05:23, 13 January 2006 (UTC)

We have what is probably a canonical answer, though it's not exactly what either of us supposed. Paraphrasing Kernighan and Richie (Thompson hasn't responded), UNICS was never written down contemporaneously, only a verbal name; it was UNIX/Unix by the time it reached any sort of official paper form. So the official answer would be, I guess, it's indeterminate and irrelevant. I will write up the answers in more detail as time allows. Georgewilliamherbert 20:46, 22 January 2006 (UTC)

[edit] Let a thousand pluralizations bloom.

'Unices' has a long and storied history, especially amongst those who are fond of Latinate pluralization. The only exception for ix -> ices I can find in English is 'fix/infix'; I don't see why 'unix' should be another exception. --Dogcow

Well, many during the 1980s insisted that while the plural of Unix was Unices, the plural of VAX was VAXen. Consistency wasn't the point, humor was. IIRC, Armando Stettner was one of the leading proponents of this. -- Gnetwerker 07:07, 21 February 2006 (UTC)

[edit] Components

A previous edit (reverted, not by me) tried to make the point that Unix was a "kernel". I disagreed, and wrote a section to prove it. Perhaps over-long, but as Mark Twain said, I didn't have time to make it shorter. Comments here, of course. -- Gnetwerker 08:11, 21 February 2006 (UTC)

[edit] Length, Section Order, New Pages

I was WP:BOLD and moved the long list of Unix versions out of (essentially) the middle of the article, and moved the important Impact section somewhat higher. (I did not change or delete any actual text). What I would really like to do is move most of the list of Unix versions to a separate article and shorten the See Also section. I think the article is overly long and structured in a cumbersome way. Comments? -- Gnetwerker 08:08, 5 March 2006 (UTC)

[edit] Portability (was "Why?")

RE: [7]

Because saying UNIX wasn't design as to what was claimed. And Plan 9 was. Reverting back. ems 19:03, 6 March 2006 (UTC)

Yes, one of the primary goals of V6 UNIX and beyond was portability. Plan 9 came along after OS portability was old hat, and while it was portable, this was not a primary goal. Please don't revert that change without further explanation of your position. -- Gnetwerker 20:11, 6 March 2006 (UTC)

Don't make a fool of yourself, UNIX was never designed to be any of the things claimed. Most of those came in the rewrite to C, and not in the designing peirod. The design ideas helped alot for portablity, but not totally (hence making UNIX's problems "to deep"). Maybe we should look up the word design? ems 08:31, 7 March 2006 (UTC)

Please try not to be rude and to actually read the answer. Dennis Ritchie, wanting the OS to be portable, started rewriting it beginning in 1971[8], first in a descendant of BCPL ("B"), then in C. The notion that an OS is "designed" as a unitary idea is an idea that could only be held by someone who has never designed one. In any case, one of the original Unix papers emphasized its portability, and this[9] document from 1976 also confirms it. And, in terms of looking like a fool, I was using and writing OS code for Unix in 1977 -- were you? -- Gnetwerker 08:49, 7 March 2006 (UTC)

I fail to see how it is being rude. I just said don't be foolish. I never called you a fool or anything. UNIX was designed in the late 60s, the foundermental design was set since then. By the late 70s most of the Bell Labs researchers (Ken, Dennis, Rob, David, etc) already started working on rebuilding UNIX from the ground up. Ritchie noted that the Plan 9 system is not based on the UNIX system, which pioneered multitasking computing on small machines. "It's reworked from the ground up, and it has no UNIX code," he said. [10] He added that Plan 9 is not a successor to the UNIX system, which pioneered real-time, multitasking computing. "It's a different system, not a reworking of the UNIX operating system, and it has no UNIX code." [11] ems 12:16, 7 March 2006 (UTC)

Let me add some detail to this:

In 1973, V4 Unix was complete, written predominantly in C, and was thusly "the first portable operating system";
Around 1976, the Interdata 7/32 port is completed;
In 1979, V7 Unix is ported to the VAX;
In 1980, Microsoft announced Xenix, a port of Unix to an Intel 8086;
In 1982, SunOS (for the 68k) ships, and HP-UX ships;
In 1990, Plan 9 is announced for the first time;

This last point is wrong. There was internal release since the 80s. ems 12:30, 7 March 2006 (UTC)

I myself ported Unix to the NatSemi 32016 and to the Intel i960 during these years.

So of course, a statement (in the Unix article or anywhere else) indicating that Plan 9 was the first OS "designed to be portable" is essentially non-sensical, as early Unix design documents (cited above) point to a need for portability, as UNIX was widely ported 10 years before Plan 9's creation, and because Plan 9 itself was descended from Unix. Indeed, it could be claimed that after V7 Unix, no significant OS was designed or implemented without portability (VAX/VMS was done by then, and other than pre-NT Windows, though it could be argued that this was not really an OS in the normal sense). In short: Unix was the originator of the concept and the first demonstration of the feasability of OS portability. IMHO, no other position is defensible. -- Gnetwerker 09:28, 7 March 2006 (UTC)

:Indeed it was portable, and was ported. But it still didn't have portability, and certainly wasn't designed to be like the article claims. Also portablity is the only incorrect statement, most of them are incorrect for UNIX, but true for Plan 9 and mordern Unixs. ems 12:33, 7 March 2006 (UTC)

Huh? -- Gnetwerker 15:35, 7 March 2006 (UTC)

Indeed, huh? What the hell was I thinking. ems 18:35, 9 March 2006 (UTC)

Let me just link to this instead of ... (note the dates) ems 21:57, 9 March 2006 (UTC)

If you had read my previous comments more carefully, you would have seen that I cited this article above, in my comment from 08:49, 7 March 2006 (UTC)[12]. For anyone too lazy to follow the link, it shows that Dennis Ritchie wanted a machine (Interdata) in 1976 in order to port UNIX. The document refers to "portlable Unix" as a goal, and hopes it might have broader appeal. Is this argument over now? -- Gnetwerker 22:00, 9 March 2006 (UTC)

Retrieved from "http://en.wikipedia.org../../../u/n/i/Talk%7EUnix_Archive_1_b184.html"