In 1989, mid-level managers at Apple sat down and planned the future course of Mac OS development. Ideas were tossed out at random, and written down on index cards for storing. Ideas that seemed simple enough to implement in the short term, like adding color to the user interface, were written on blue cards, while more advanced ideas, like an object-oriented file system were written on pink cards.
Development of the ideas contained on both sets of cards was to proceed in parallel, the team working on blue releasing an updated Mac OS in the 1990-91 time frame, and pink an entirely new OS around 1993. The blue team delivered what became known as System 7 only slightly late in 1991, but Pink suffered from second-system effect and continued to slip into the indefinite future. Eventually Apple semi-abandoned it by spinning it off to form Taligent.
This left the Mac OS in a bad position. Originally intended to support a single user running a single application on a non-networked machine with a floppy disk for storage, many parts of the operating system simply did not "scale" well. In particular the architecture of QuickDraw made it very difficult to introduce multitasking into the system, and the file system was particularly inefficient when used on hard drives.
System 7 would have been an opportune time to address some of the "low level" problems in the Mac OS, but instead consisted of a huge set of "bells and whistles." With the previous System 6 at only 500k or so it may have been possible to run several "virtual Macs" to provide robust multitasking, but a basic System 7 install took well over 1MB of RAM in an era where 2MB was common and 4MB was the most the majority of the Mac line could handle. A path forward was no longer simple, as users would be unwilling to give up on features to get there.
Several attempts were made by various teams to address these issues and start work on an updated system, but invariably they ran afoul of internal politics and turf wars. John Sculley, Apple's CEO during this period, largely ignored managing of the company itself while he concentrated on sales and marketing, and as a result the engineering departments were essentially out of control. Several projects made significant progress, only to be ended personally by Sculley after some other department manager complained and Sculley ended the feud by cancelling both.
With System 7.5 released in the autumn of 1994, Apple management finally decided that the decade-old Macintosh system software (Mac OS) had run its course, and an entirely new operating system, with more advanced features, would be needed for the platform to compete with upcoming releases of Microsoft Windows. Soon word was out that Apple was working on a new system known as Copland, to be followed by an even more advanced system known as Gershwin.
The basic idea for Copland was to run the Mac OS on top of a microkernel. Older applications would be run inside a "box" that made it appear as if they were on a Macintosh all to themselves, while newer programs would have direct access to the features of the microkernel, including a new networking system, file system, and high-speed interapplication communications. Particularly important to the system was the use of QuickDraw GX, which was able to run existing QuickDraw applications in a multitasking system, as well as offered new features to developers of new software.
The trick was to make all of this fit into existing Macs. Running a number of copies of System 7 would simply not work due to its large size. Copland attempted to address this by introducing a large number of shared libraries and a fiendishly complex memory management scheme. It was intended that Copland would only be perhaps 50% larger than the existing Mac OS (now well into the 2MB range), a reasonable goal given the ever-increasing amount of RAM being incorporated into the average machine.
Another key feature of Copland was that it would be completely PowerPC "native." System 7 had been recompiled on the PowerPC with great success, but the system still relied on the processor looking like a member of the Motorola 68000 family. In particular the interrupt handlers in the Mac OS had to be emulated, requiring an expensive call into the OS to translate these to the PowerPC's much simpler system. Removing this limitation would allow Copland-native applications to run much faster, as much as 50%, with no special effort on the part of the developers.
Parts of the future Copland system were demonstrated at Apple's Worldwide Developers Conference in May 1994, notably an early version of the new file system. Apple also promised that a beta release would be ready by the end of the year, for full release in early 1995. Throughout the year, Apple released a number of mock-ups to various magazines showing what the new system would look like, and commented continually that the company was fully committed to this project. By the end of the year, the developer release was nowhere in sight.
At WWDC '95, Apple's new CEO, Gil Amelio, talked exclusively about Copland, repeatedly stating that it was the only focus of Apple engineering. Now referred to as System 8, he announced that it would really ship to developers only a few months later at the end of the summer, with a full release planned for late fall. Oddly, very few demos of the system were actually shown. Instead what was demonstrated were various bits of technology and user interface that would go into the system, for instance a new file management dialog. Little of the technology of the core system was demonstrated, including the file system that had been shown a year earlier.
Then an interesting thing happened. After a number of people at the show complained about the lack of sophistication of the microkernel, notably the lack of multithreading, Amelio came back on stage at the end of the show and announced that they would be adding that to the feature list. This implied that the system was nowhere near ready, as a feature like this is so "core" to a kernel that it would be impossible to add it so close to the shipping date. Amelio had been brought in, among other reasons, to help clean up the feature creep that had destroyed project after project during the early 1990s. Realizing that something serious had to be done, Amelio lured Ellen Hancock away from her management position at IBM to take over engineering as Chief Technology Officer and try to get Copland development back on track.
In August 1995, a small number of "preview" releases were sent out to selected developers. The system was so unstable that it was impossible to develop on, and would frequently kernel crash spontaneously on its own. In November 1995, a beta version was released 50 key developers. Though more stable, this version required two computers linked via serial connection to run, which made it impractical for real world use. Amelio realized that the market knew Copland was floundering and in May 1996, Apple delayed the target delivery date to mid-1997. This wasn't too much of a surprise to people who had been watching the debacle unfold. In fact one of the groups most surprised by the announcement was Apple's own hardware team, who had been waiting for Copland to allow the PowerPC to truly shine.
Among the reasons given for the cancellation was the slow pace of development, and that many technical problems remained to be solved. Essentially, Apple had concluded it could not feasibly complete the project on its own, nor produce a product that could viably compete with Microsoft and its then-new Windows 95 operating system, which was rapidly gaining acceptance.
In the aftermath of the cancellation, Hancock mentioned an interesting fact that illustrates the problems within Apple at the time. For many years Apple had talked about Copland and Gershwin as a one-two punch, with Copland getting developers and users onto a modern platform, and Gershwin surpassing anything on the market in terms of power. But when she started at Apple in mid-1996, she quickly discovered that no real plan existed, and there was not a single person in the company who was working on anything related to Gershwin. The project existed entirely in marketing materials.
It was Hancock, more than anyone, who decided to kill Copland. After only a few months, she realised that the situation was hopeless, and given current development and engineering it would never ship at all. Her suggestion was to continue development of the existing Mac OS in order to improve its stability, while looking outside the company for a new OS to build future Macs on.
Having canceled its own development of an all-new operating system, Apple sought a new strategy. Amelio began to search for an alternative operating system, to acquire and merge with Apple's own legacy platform. The two leading contenders to provide Apple with the technology it needed were firms founded by Apple alumni: NeXT Computer, led by Steve Jobs, and Be Inc., led by Jean-Louis Gassée.
- Jobs's firm, NeXT Computer, had developed an advanced computer workstation and multi-tasking operating system, known as OpenStep. The machine had been aimed at the higher education market, but had never been a commercial blockbuster.
- Gassée's firm, Be, Inc., had a similar mission, although their product was still in earlier development stages. Their computer, known as the BeBox, had a high level of graphical performance -- also an Apple strongsuit -- together with strong multi-tasking performance, which Apple needed.
- The third contender was Solaris, an Unix by Sun Microsystems, popular on servers and workstations. This option was strongly defended by Hancock, but rejected by Amelio.
Amelio considered both firms closely, and widely published news accounts in the fall of 1996 speculated that an Apple purchase of Be was imminent. Ultimately, however, Amelio balked at Gassée's asking price, and chose instead to purchase NeXT for USD $427 million in December 1996.
NeXT not only offered Amelio the software and technologies that Apple needed to succeed, but the unique opportunity to bring Steve Jobs, one of Apple's co-founders, back into the fold. Jobs had been dismissed from Apple in 1985 by John Sculley, and remained an iconic figure in Silicon Valley.
The NeXT development team immediately began work on the Rhapsody project, which would eventually become Mac OS X. In July of 1997, Amelio was forced to resign, and soon thereafter, Jobs was appointed "interim CEO." Jobs took drastic steps to restore the company to profitability. He terminated the Mac OS "licensing" program, which had opened up the platform to "clones," and cancelled a number of development projects, including HyperCard, Newton, OpenDoc, Pippin, and QuickDraw 3D.
- "The Long and Winding Road", MacWorld, August 31, 2000
- For the good of the company? Five Apple products Steve Jobs killed by Casey Johnston, Ars Technica. 2011-08-25.
- Apple Flops: Copland at Cupertino.de (Deutsch, archived 2002-01-08)
- Copland and NeXT by David K. Every at iGeek (2002-09-10, archived 2004-02-23)
- Copland was a failure by David K. Every at MacKiDo (1999)
- The Copland Crisis by Owen W. Linzmayer at MacSpeedZone (1999)
- Copland, Where Are Thee? by Don Crabb at MacTech (1995-12, archived 1999-04-22)
- Copland by BG at The Long View (2011-02-26)
- Copland (operating system) at Wikipedia