Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Is OS X UNIX?

Taft

Taft

macrumors 65816
Original poster
Jan 31, 2002
1,319
0
Chicago
I wrote this in response to the bickering over wether or not OS X is in fact Unix in the forums here and at other sites. I direct this writing especially towards Tadpole.

The problem with labeling OS X a non-Unix is that the definition of Unix is very fuzzy. A tiny history of Unix is needed to spark this discussion. Unix did in fact originate at AT&T Bell Laboratories and developed along a single, unified path until 1975 when Version 6 was released. (Reference 1) This is where the first major split occurred. Developers at Berkeley began extending Unix (BSD) while AT&T continued releasing their own Unix's. An interesting point to make here is that every version of Unix released by AT&T after Version 6 contained elements of the BSD Unix developed at Berkeley. That means when System V.4 was released in 1988 by AT&T it contained many of the extended features that 4.4 BSD contained (released in 1993).

Contrary to statements I have seen about OS X inheriting from System V.x, it certainly did not. Darwin (OS X's base--including its kernel) was originally based on OpenStep 4.x, but was soon moved onto a Rhapsody base. (Reference 4) Rhapsody is based on BSD Lite2 which, in turn, is based on 4.4 BSD from Berkeley. This means that Darwin currently has no roots in AT&T's System V.x but is instead a derivative of BSD.

None of this addresses the issue of wether we can label OS X as Unix. For that, we must consider the current attempts at standardization. There have been many attempts at this, each having varying success. (Reference 2) All of these attempts occurred while AT&T still held a trademark on the name Unix.

Considering that a single party had control over the Unix trademark, these official standardization attempts were a bit futile: AT&T or whoever held the trademark could call anything they wanted to Unix and deny that right to anyone they wanted. Not to say that this type of arbitrary decision ever happened but the the owner of the trademark has that same right today.

Currently the trademark is held by the Open Group. This group is currently working on standardizing Unix for the world and allow companies to test and apply for the Unix label for their distributions. (Reference 3) Open Group only allows those distributions that conform to their standards use the Unix trademark and its logos. The standards they have laid down are a combination of the BSD, OSF/1, and System V.x Unix definitions. These standards are closer to the System V.x definition, however.

This can be demonstrated by Sun's various distributions. SunOS is Sun's BSD derivative operating system while their Solaris OS is based mostly on System V.4. (Reference 1) As you can see from the Open Group's web page (Reference 3), SunOS is NOT a registered Unix product while Solaris is.

The Open Group also maintains the POSIX standard that, while related to the Unix standard, is not the same as Unix.

So what does this all mean? Should we consider Mac OS X to be Unix? How about Linux? Or SunOS for that matter?

My personal inclination is to say that they are ALL Unix. That does not mean that everything that is compatible with linux will run on Solaris version 8. In fact, a quick test of Solaris compatible, non-gui based software on linux will show quite a few incompatibilities. The same would hold true with SunOS.

But the standard set down by Open Group should not be used as the end all definition of Unix. Developers at Berkeley did some amazing things with the original Unix released by AT&T. Many (if not most) of these things were later integrated into the original Unix (like virtual memory, the C shell, job control, and TCP/IP networking). I'm not sure that it is in the best interest of the Unix community to shut certain derivatives of Unix out while allowing others in. At the very least, another less constricting definition should be created to include those Unix derivatives that have viable uses as Unix systems.

Not to say that what the Open Group is doing is useless. They are attempting to create cross-platform compatibility standards. This is a very good cause, but does not strike at the heart of what is really a UNIX system.

I consider linux and Mac OS X Unix simply because most of the properties of Unix exist on these systems. I can take most shell scripts and run them on either OS X or Solaris and command line programs will generally compile on both systems. This is not always true, but I'm sure differences can be found in Unix compliant distributions by Compaq and Sun. Calling OS X a non-Unix operating system is senseless nit-picking.

I hope this provides clarity and inspires constructive comments or corrections.

Thanks

Matthew


References

1 Frisch, Aeleen Essential System Administration. O'Reilly & Associates, Inc., 1995.

2 Martin, Vernard There can be only one! A summary of the Unix standardization movement. ACM, 2001. http://www.acm.org/crossroads/xrds1-3/unix-standards.html

3 Register of Open Branded Products. Open Group, 2002. http://www.opengroup.org

4 Darwin FAQ Rob Braun, 2000. http://www.darwininfo.org/faq.shtml
 
the ultimate family tree

http://perso.wanadoo.fr/levenez/unix/

hey fellas, i dug up this huge ass mother of all family trees. the site is kind of slow, its french afterall, but its in English though.

I'll have more on this in a bit, I want to consult the nerd herd in my lab some.
 
Re: Maxrool

Sure a person could argue for DOS being a UNIX. And I take your point.

But what arguments and similarities could you gather to support the claim. My point is that the differences between those operating systems that are generally referred to as UNIX are mostly superficial. Sure there are different design principles at work in some areas of the OS (such as the kernel). But the fundemental ways in which they work are hugely similar. The same cannot be said of Classic Mac OS, DOS, NT, BeOS, etc... The fundemental design of these OS's is very different.

The reason I would call Mac OS X a UNIX is that nearly everything below the GUI level works in a way that is completely familiar to anyone that has ever used Solaris, or BSD, or Linux. I can use a UNIX administrator's handbook to figure out how to control and manage a very high percentage of functions in OS X. Everything just works like any other UNIX works. Try administering a DOS machine with a UNIX handbook!

This isn't a very scientific way to define what is or is not a UNIX, but I think it suffices. I just don't know what would be accomplished by discluding Mac OS X, Linux or any other clearly UNIX-like system from UNIX-hood.

Matthew
 
Although the exact definition of UNIX is a little fuzzy, the process that a vendor must complete to call a product "UNIX" is very clear. I think it would be interesting to find out how close OS X is passing the compliance test suite. I don't imagine it could be that far away.

BTW - in reference to the earlier post, some of you may be surprised to know that the design of DOS was heavily based on what UNIX was at the time.
 
tadpole according to your "family tree", mac osx is indeed a form of unix. or a "flavor" of unix if you prefer.
 
My own definition of UNIX/Linux

Well this is non scientific as hell but the way *I* determined that OS X is UNIX, more specifically, Linux-enough for me, is that I can install apps like Fink. Can't do this with XP, can't do it with OS 9, CAN with Red Hat's Linux distribution, so I say OS X is Linux-enough.

BTW, nice post - more like an article - thanks for the time and effort. (complete with a bibliography - did you do it for school?)
 
Originally posted by gbojim

BTW - in reference to the earlier post, some of you may be surprised to know that the design of DOS was heavily based on what UNIX was at the time. [/B]
Click to expand...

Actually I remember reading that somewhere, but I could not locate it or I would have made reference to it.

This is what prompted my original comment.
 
UNIX

One of the biggest differences between UNIX and another OS is that it uses files for everything. There are no drivers for UNIX. When you hook up a tape drive and you want to access it your reference it by using a file that points to it..

example:

/dev/rmt/1n

1n is a file that points to a device. UNIX uses simple files with rwx privileges to do everything.

OS X is built on UNIX ( the BSD version) ... it has similar file structure and it handles processes in a similar way. Its a "blend of UNIX" ... I would say.
 
With the exception of Steve Jobs, why did Apple go with NeXT and not Be?

I've read very few things about BeOS, but from what I have read, it seems like it was quite the OS. I wonder that OSX would have been like if it were based on Be's technology, and not NeXT's.

Why is Unix better than Be?
 
NeXT vs. Be

From my understanding of Be, there would have been more hurdles to get over in order to accomplish everything OS X want to do.

I loved Be, but I believe in this case Apple went in the right direction. Be was/is a cutting edge OS. Very low latency at the Kernel level, a database style filesystem, etc...it was just a different way of going about an operating system. I think this was why it was passed up when looking at Mac OS's future.

OS X needed to accomplish many things: it needed to offer a *ton* of backwards compatibility to legacy applications, it needed to offer a path to the future, and it needed to offer an ability to extend to server capabilities.

Be OS offerred one of these things as well as NeXT did, probably better. That category was a path to the future. Be had more forward looking elements than any other OS I can think of. However in the other two categories, it was lagging.

The backward compatibility would have been hampered by the very things that make Be so great. NeXT is based on UNIX so the classic environment was not a new concept (hear of MOL?). This made NeXT a good choice in this category (though not necessarily better than Be).

Also, from my understanding multiple API sets were a more viable option on NeXT for things like Carbon. Carbon made porting apps very attractive to big name developers (MS unfortunately comes to mind). This was definately needed for the next generation Mac OS to succeed. Remember Copland? Developers don't take kindly to rewriting their apps from scratch. As a developer, I can understand why.

Another plus for NeXT is that so much UNIX freeware was either available or easily ported to a truely UNIX system. Projects with the awesome power and potential of Fink would not have been feasible under Be. The Gimp, Abiword and the rest of the free office software, command line utilities and programs, network utilities...UNIX has a lot of software ripe for the picking.

Finally, on the server side, UNIX wins hands down. It does for a few reasons, but the biggest is that it has been tested. For years. With a lot of success. Be couldn't compete in these regards. Also, NeXT had some pretty cool LAN software (NetInfo is really a good thing, though it can be confounding to first time users). These allow small networks to run off of OS X servers very easily and competitive to similar MS initiatives (like Active Directory).

I know Be has a lot going for it as well, but NeXT really was the best choice for OS X. I was really sad when I heard Palm was killing Be OS after buying Be out. :( Then again, I was really sad when I heard Be was dropping Mac support. I just hope we see some Be innovations start to slip into mainstream OS's. Maybe there will be a ressurection?? Are you listening Palm...

Matthew
 
Thanks!

Thanks Matthew!

Here's another question.... Why is it that Apple had to resort to purchasing an OS instead of developing their own? I recall reading that they had begun work on a successor to the "classic" MacOS, but dropped it in favour of purchasing NeXT. Why? Does anyone know any sites that tell the history well?
 
BTW

In response to the length and bibliography and whatnot...

Its pure boredome. I quit my job and have been off for a week and a half. I'm really looking forward to starting my new job tomorrow.

I do like the forums here, though. I've had a good time scanning the forums here for the last week and a half. Good stuff.

Matthew
 
Mac OS history

Do a search for 'Copland'.

That was Apple's major attempt at a overhauled OS...around 95-97 I think. There were a lot of reasons it failed. A big one was a lack of proper legacy support. But that was also during one of Apple's darker times...the Amelio/Clone era. A lot of people moved from the platform then. Rough.

Here is a site that shows the history of Apple in general, though it doesn't focus on the technical issues much.

http://www.apple-history.com

Matthew
 
Mac OS X is UNIX according to the Open Group

This is a response to a "Is MacOS X UNIX?" flamewar on another BBS. This guy went nuts because I used the phrase "Pure, Certified UNIX by the Open Group". I have cut and paste it here for your review:


I thought I would give it one more shot. I would rather have the facts instead of insults.

Why don't you boys come back when you can show me its name in a list, instead of playing your silly games?


I will start from the beginning and show from the Open Group website what can be called UNIX and how it is certified as such.

First, here is the list of UNIX vendors from the Open Group website:

Platform Vendors Supporting the Single UNIX Specification:
Acer; Amdahl; Apple; AT&T GIS; Bull; Convex; Cray; Data General; Compaq; Encore; 88 Open; Fuji Xerox; Fujitsu Ossi; Hal; Hewlett-Packard; Hitachi; IBM; ICL; Matsu****a; Mips ABI; Mitsubishi; Motorola; NEC; Novell/USL; Oki; Olivetti; OSF; PowerOpen; Precision RISC; Pyramid; SCO; Sequent; Sequoia; Sharp; Siemens-Nixdorf; Silicon Graphics; Sony; Sparc International; Stratus; Sun Microsystems; Tadpole; Tandem; Thompson/Cetia; Toshiba; Unisys; Wang Labs.

So we see that Apple is listed as a Single UNIX Specification vendor. From the osOpinion article and the Linux News correction, we know that Apple was not listed before May 2001. We know that Apple does not distribute or support A/UX (only runs on 68k Macs from the 80s) or IBM's AIX (which only runs on ANS machines from the early 90s). So we know that the product that Apple is being listed for is Mac OS X and its core Darwin. No one would ever suggest that Mac OS 9 or its ancestors are UNIX or even UNIX-like (Junk-like in my opinion compared to Mac OS X).

What is the Single UNIX Specification and how are you "certified" as such, so you can use the UNIX trademark with your product? Once again, from the Open Group website:

The Single UNIX Specification is supported by the X/Open UNIX brand, which in turn is supported by a verification program. The X/Open brand provides the guarantee that products adhere to the relevant X/Open specification. Systems that provide the Single UNIX Specification interfaces can be X/Open UNIX branded as proof to the marketplace. The Single UNIX Specification is the programmer's reference to the portability environment provided on X/Open UNIX branded systems.


So we see that in order to say you support the Single UNIX Specification, you must be able to pass a verification program. If you pass, you receive the X/Open UNIX branding. Think of it like Underwriter Laboratories -- that UL symbol that should be on all your electrical appliances. In order to be certified by UL, you must pass a verification program so you can use their logo. Notice the sentence, "Systems that provide the Single UNIX Specification interfaces can be X/Open UNIX branded as proof to the marketplace". Branded as proof, certified, to the marketplace.

The following paragraph states:

It is important for application developers to realize that in committing to the brand, the vendor is obligating themselves to conform. X/Open conformance verification suites are vendors' ways of measuring their implementation against the specification, providing a guarantee that they have implemented the specification correctly.

Once again, if you are a Single UNIX Specification vendor and verified as such by the Open Group X/Open UNIX brand's verification program, Open Group is guaranteeing that you have a UNIX product. So you are either UNIX or not just like pregnancy. If you deviate from the specification, you are no longer certified as UNIX by the Open Group.

Think that pretty much raps it up for the facts as I see them. If you would have read the page that I referenced several times, this information would have been available to you. This would have alleviated your need to hurl insults at me. Your hang up on the phrase "Pure, Certified Mac OS X" is most likely my fault. A more accurate statement, on my part, would have "Pure, Verified, Guaranteed, Branded as X/Open UNIX, Proven Single UNIX Specification Mac OS X". I was being vague saying just "Pure and Certified" -- feel justified in celebrating this revelation.
 
i'm pretty sure they bought next because steve convinced them to

the only way to get steve back was to buy his company and use the os... he made them

dont ask me why steve was asked back... he was only supposed to "help" the company, but he eventually was reinstated as iceo

i love apple & its history :)
 
more haphazard research

http://bhami.com/rosetta.html

its a comparision of functionality. i emailed the open group guys to get the slim from them where OSX fits in all of this, I'll post as soon as I here back from them.

night.
 
It will be interesting to see what response tadpole gets from open group.

Without trying to extend/start any flame wars, I think soy_bomb may be missing or ignoring an important piece of information.

According to his link, Apple is listed as a vendor supporting the Single UNIX standard, and in fact Apple is a member of Open Group. However, if you look on the list of registered products for UNIX 93, 95 and 98, there are no listings for Apple. Also, Apple is not promoting OS X as UNIX, rather it is UNIX based or UNIX like.

It appears they have not certified yet, if they ever intend to do so.
 
How far away is OS X...

...from passing the Open Group tests?

I'm not really sure. But I have heard talk of a difference in the threading implementation of OS X that might make compliance with the Open Group standards difficult. Can anyone verify this?

Matthew
 
Copeland

Theres a slight oversight there.
Copeland was not dumped. Copeland was released under the name Mac os 8. The one that everyone gets confused with is Pink. Pink was started shortly after System 7.0 and was suppose to be a complete overhaul of the system from the ground up.

Pink was killed in favor of Copeland because copeland did not require a complete rewrite of the OS and was compatable with exsiting hardware of the time (Pink was not). Thise led to one of the biggest reasons why Apple selected NeXT.

There were several OSes that Apple looked at for a sussor to Copeland. NeXT, BEos, NT,AiX. It came down to NeXT and Aix because Pink was taken (partialy) from AuX (Apple Unix). These ment that alot of the codeing done to make the GUI work was portable to the new OS without alot of work.

The Next kernel was also better then the others and Apple wanted to be Kernel based. Pink's Kenel was very closely related to Mach (that NeXT used).

Rapsody was released under the name Mac OS X server (v. 1.0). This was NeXTStep with the mac GUI thrown on top. IT even had classic! (MACOS.app)
To this day OS X is much like NextStep that you could get a book on NextStep and use it for os X.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back