How Apple Created Hackintosh, Part I

At a developer conference in June 2005, Apple announced that OS X will transition from PowerPC to Intel processors. The first Intel Macs (sometimes nicknamed “Mactels”) were introduced several months later, in January 2006. The announcement generated significant interest in the PC world, and everyone started wondering: Is OS X going to run on my PC? The then-current OS X 10.4 Tiger was in many ways superior to Windows XP, and while Windows ruled the business world, Macs were very strong in the media creation department.

Soon after the conference, Apple shipped custom development systems to interested parties who were willing to pay $999 for the privilege. The systems came with an Intel version of OS X 10.4.1 which refused to run on any standard PC, even though the Apple development systems were extremely close to a vanilla PC.

No Intel OS X for you

No later than August 2005 (still months before the official release), the first hacked versions of OS X Tiger for x86 appeared. This in turn spawned an entire strange subterranean cult of Hackintosh, populated largely by people whose time is worth very little and who put a lot of effort into creating an entirely untested and unsupported system.

But how did Hackintosh appear on the scene so quickly? The person most responsible for that was none other than Steve Jobs, and to understand why, we have to rewind another 20 years back.

In 1985, Steve Jobs left Apple (how much that was voluntary not is a matter of perspective). Later that year, Jobs founded NeXT (later NeXT Computer), a maker of hip workstations and a very unusual operating system called NeXTSTEP. Initially NeXTSTEP only supported NeXT’s own 68000-based workstation launched in 1988, the NeXTcube. In the early 1990s, NeXTSTEP was ported to SPARC, HP PA-RISC, and Intel’s x86 architecture.

While not an Apple product, NeXTSTEP was in many ways an embodiment of Apple’s “Think Different” mantra. The Mach microkernel was combined with much of BSD kernel and userland, with a very modern, object-oriented programming interface based on Objective-C (when NeXTSTEP was developed, C++ wasn’t mature; Apple also funded the addition of the Objective-C front end to GCC). NeXTSTEP became popular with developers because it enabled rapid development of complex software. Notable developed-on-NeXT software includes HTML/WWW and DOOM.

NeXTSTEP 3.3 Desktop (1994)

By about 1994, NeXTSTEP ran on many generic PCs. In 1995, NeXTSTEP turned into OpenStep, without much change in the underlying OS. In February 1997, Apple acquired NeXT (and Steve Jobs) due to a desperate need for an OS more modern than the classic Mac OS, which was rather similar to Windows 3.x in far too many ways.

OpenStep 4.2 Desktop (1997)

Apple got to work and in 1997/1998, released developer versions of Rhapsody OS, essentially an Apple-ized NeXTSTEP. For many years, Rhapsody DR2 (Developer Release 2) from 1998 was the last Apple desktop OS which ran on x86.

Apple Rhapsody DR2 Desktop (1998)

By the time Rhapsody turned into OS X and was released first as Mac OS X Server 1.0 (1999) and Mac OS X 10.0 Cheetah (2001), only PowerPC Macs were supported. Except not really. All through the early 2000s, Apple supported a mostly-open-source project called Darwin. This was essentially OS X without the Apple GUI and applications. And Darwin did support x86 all along—including AMD and VIA processors.

There are mythical stories about how one man ported OS X to PCs, but they’re just that, stories. The OS started as a cross-platform one (with last release in 1998, see above), and Apple maintained the x86 support in Darwin all along. The only mystery is why anyone would be surprised by Apple switching to Intel despite the rumors of Project Marklar going back to 2002 and the obvious evidence of Darwin x86 support which was right on Apple’s web site staring everyone in the eye.

As of November 2017, Apple still provides the last Darwin 8.0.1 bootable ISO if one knows where to look (and here are the release notes). This is a stripped down OS X 10.4 Tiger which runs on generic PCs (there’s a version for PowerPC Macs, too). Among other things, Darwin contains a BIOS boot loader which can start the xnu kernel from a HFS+ partition (and Apple provides the source code, too).

Darwin 8.0.1 installer (2005)

To recap, Darwin 8.0.1 in April 2005 was Apple’s last Darwin binary release. In early June 2005, on top of Darwin 8.1.0, Apple had this:

First Intel developer release of OS X (2005)

It is obvious that if Apple shipped the first Intel OS X release to developers in June 2005 and Intel Macs were released in January 2006, the Intel version of OS X was quite far along by the time anyone outside Apple saw it. Certainly it does not feel like a 1.0 release—the PowerPC edition of OS X had been in wide use for four years by then, the OS had internally been running on Intel for a while, and for the transition period, Apple provided Rosetta, a dynamic PowerPC to x86 application translator.

As mentioned at the beginning of this article, the developer version of OS X 10.4.1 (and 10.4.3 in the Fall of 2005) did not run on standard PCs, even though the DTS was 99% a standard PC. For the most part it was because Apple simply didn’t care about hardware that Apple didn’t ship, but there was an added “safety” to ensure that OS X wouldn’t run on non-Apple PCs. But of course the 99% part was crucial and that’s why it only took a few weeks for Hackintosh to show up. More about that in Part II.

This entry was posted in Apple, Intel, PC history, Pre-release. Bookmark the permalink.

10 Responses to How Apple Created Hackintosh, Part I

  1. ender says:

    I remember following a tutorial to get OS X running inside VMWare. IIRC, it involved installing OS X inside PearPC, then booting that image in Workstation.

    https://eternallybored.org/osxvmware.png

  2. Andreas Kohl says:

    There was a liason between NeXT and IBM from 1988 to 1990. It resulted in ports of NextStep (not to confuse with NeXTSTEP) for IBM’s AIX/6000 and AIX PS/2 operating systems. NextStep is the same graphical environment and development system (including the Stepstone Objective-C compiler) without NeXT’s kernel and base operating system components. For more information see IBM announcement letters 290-072 (http://www-01.ibm.com/common/ssi/rep_ca/2/897/ENUS290-072/) and 290-171 (http://www-01.ibm.com/common/ssi/rep_ca/1/897/ENUS290-171/).

  3. Neozeed says:

    I should add that Darwin 0.1/0.3 which match OS X Server 1.0 and 1.2, that they were only provided in binary form for the PowerPC, actually could be compiled on Rhapsody DR2, and the resulting 5.5 kernel runs just fine. I have one VM with close to 6 months uptime.

  4. bsd says:

    Also see “The Mac Observer Interview with Jordan Hubbard” [1] where they discussed this (“We talked about the early development of Mac OS X derived from NeXT and even the earliest BSD origins.”)

    [1] https://www.freebsdnews.com/2017/02/23/tmo-background-mode-interview-open-source-developer/

  5. Michal Necasek says:

    NeXT later had a similar arrangement with Sun, working on OpenStep for Solaris.

  6. Michal Necasek says:

    Darwin really is one of the weirdest operating systems on the planet, it’s such a mix of… well, everything. But in many ways I think it’s the nicest desktop Unix out there, and most Mac users probably have no idea!

  7. Yuhong Bao says:

    Interestingly, Darwin/x86 (pre-2005) used call gates: https://opensource.apple.com/source/Libc/Libc-391/i386/sys/SYS.h.auto.html
    Call gates are hard to support on x86-64 due to them not disabling interrupts, and Solaris used a hack using #NP to support them. When x86-64 with SYSCALL/SWAPGS was being developed, AMD was mostly focused on Linux that did not use them. (OS X now uses SYSENTER/INT on 32-bit)

  8. Michal Necasek says:

    A very special flower indeed, that was nicely put in the interview.

    I can’t but wonder if the departure of the likes of Jordan Hubbard and Bertrand Serlet from Apple is directly related to the deteriorating quality of OS X/macOS.

  9. >I should add that Darwin 0.1/0.3 which match OS X Server 1.0 and 1.2, that they were only provided in binary form for the PowerPC, actually could be compiled on Rhapsody DR2, and the resulting 5.5 kernel runs just fine. I have one VM with close to 6 months uptime.

    Continuous uptime, continuous-except-for-host-machine-reboots uptime, or total uptime?

  10. @Sean McDonough continuous. I haven’t rebooted either in months.

    ~ # uptime
    15:41:02 up 146 days, 10:05:17, load average: 0.17, 0.16, 0.15

    [darwin:~] root# uptime
    7:31AM up 146 days, 9:59, 2 users, load averages: 0.72, 0.80, 0.82
    [darwin:~] root# hostinfo
    Mach kernel version:
    Kernel Release 5.5:
    Sun Apr 30 10:53:53 SGT 2017; root(rcbuilder):kernel-7/BUILD/RELEASE_I386
    Copyright (c) 1988-1995,1997-1999 Apple Computer, Inc. All Rights Reserved.

    Kernel configured for a single processor only.
    1 processor is physically available.
    Processor type: pentium (Intel Pentium)
    Processor active: 0
    Primary memory available: 512.00 megabytes.
    Default processor set: 32 tasks, 56 threads, 1 processors
    Load average: 0.89, Mach factor: 0.13

    If this formats correctly… the first one is the ESX host, the second is the VM.

Leave a Reply

Your email address will not be published. Required fields are marked *