Ellen Hancock, the Woman Who Turned the Tide#
By Wang Yue
Let's start the story in the 1990s. Ellen Hancock is the protagonist of this article and a female hero who is well-known for her experience at IBM.
Ellen Hancock served as the CTO of Apple
From 1966 to 1995, Ellen Hancock worked at IBM for a total of 29 years. In 1985, she became the Vice Chairman of IBM. From 1986 to 1988, Ellen Hancock served as the President of IBM's Communication Products Division, and in 1992, she was elected as a Senior Vice President.
In September 1995, she was persuaded by Gil Amelio, the CEO of National Semiconductor at the time, to join the company as an Executive Vice President. During her time there, she led a team to develop the CompactRISC architecture, which later became the precursor to the ARM7 series. Many people have forgotten about her and few can associate her with Mac OS X. However, she was actually the key figure who convinced Apple to abandon Copland and acquire NeXT.
Gil Amelio served as the CEO of Apple
As early as 1994, Gil Amelio had already found a new home for himself at Apple and became a member of the Apple Board of Directors. In February 1996, Gil Amelio resigned from National Semiconductor and became the CEO of Apple. In order to follow his boss's call, Ellen Hancock was once again persuaded to join Apple, which was in a crisis at the time. This was in May 1996, and why was it called a "crisis"? Let's go back to what happened earlier.
In the 1980s, John Sculley, who sold cola, became the CEO of Apple, and Steve Jobs was ousted from Apple. After all, cola and computers are not the same thing, so both hardware and Mac OS development projects at the company were increasingly hindered.
John Sculley and Steve Jobs, once as close as family, but ended up becoming enemies
Moreover, due to its inherent shortcomings, Mac OS did not have the features that a modern operating system should have from the beginning. Therefore, in 1987, plans to develop the next generation operating system began to emerge. The specific plan was to write the required functions of the new system on a stack of cards. Short-term achievable goals, such as adding color support (computers were still black and white at the time), were written on blue cards; long-term goals, such as multitasking, were written on pink cards; and long-term goals that could not be achieved in the foreseeable future, such as adding a purely object-oriented file system, were written on red cards.
With this approach, the Mac OS development team was divided into two groups. One was called the "Blue Group," with the goal of releasing an updated version of Mac OS in 1991. The other was called the "Pink Group," which worked simultaneously with the Blue Group and planned to release a completely new operating system in 1993.
On May 13, 1991, the Blue Group successfully completed its development tasks and released Mac OS 7 (commonly known as System 7). However, the Pink Group failed to produce anything practical and experienced consecutive delays. Additionally, due to a lack of personnel for the release of Mac OS 7, it was often necessary to transfer staff from the Pink Group to the Blue Group's development. Furthermore, Apple was focusing on collaborations with companies like IBM (Taligent project) rather than the Pink Group, which ultimately led to the demise of the Pink Group's project. Apple had hoped that the Taligent project, in collaboration with IBM, would develop a usable new system, but IBM eventually stopped playing with Apple, leaving Taligent without any results, which was quite frustrating for Apple.
Demonstration of Taligent OS
At this point, due to the inherent shortcomings of Mac OS (single-tasking, no memory protection), combined with the unlimited increase in Apple's own software and third-party software (during this period, Apple had already added QuickDraw, PowerTalk, QuickTime, and other software and technologies, each of which was larger than Mac OS itself), the problems with Mac OS finally erupted. In the 1990s, Mac OS was known for being unstable and frequently crashing, similar to the impression Windows 95 left on PC users, or even worse.
After the failure of the Taligent project, Apple tried more than ten different internal projects, most of which were short-lived. At this time, Windows NT was becoming mature. Seeing that the days were getting tougher, Apple began to reconsider the establishment of the next generation operating system. In 1994, after the release of Mac OS 7.5 (Mozart), Apple came up with a new plan to establish a completely new operating system named Copland (in memory of Aaron Copland, Mac OS releases were named after musicians, which was very different from the later use of cat names in Mac OS X). This project would have a completely new kernel with all the advanced features similar to the Windows NT kernel, and old software would run as separate processes.
This project had a tight schedule and heavy tasks. The plan was announced in March 1995, with a planned release in 1996. The version after Copland, Gershwin (in memory of George Gershwin), was expected to be released in 1997 and would rewrite all major components of Mac to fit the new kernel's various features.
Copland would use microkernel technology, only handling tasks and memory allocation. All other functions, such as the file system and hardware drivers, would run as services on top of the microkernel. All user interface functions of Mac OS would become an independent framework called the "Blue Box" (we will encounter this term again when we introduce Mac OS X in the future). All tasks would be independent, occupying separate memory, and could communicate with each other using IPC.
Anyone who has studied operating systems knows that microkernel was a buzzword at the time. A system could only be considered advanced if it was called a microkernel. At that time, there was also the famous Tanenbaum-Torvalds debate about the Linux system. However, it turned out that all successful projects that aimed to be microkernel-based systems abandoned their original designs (including NeXTSTEP and Windows NT) because such Mach-like microkernel systems often faced difficulties. Projects like GNU/Mach + Hurd, which have been in development for 20 years, still haven't succeeded. It's not easy to develop a microkernel system within a year.
Illustration of monolithic and microkernel architectures
Before the microkernel was developed, almost all development teams at Apple caused chaos. Everyone claimed that what they were working on was very important and they must join the Copland development team. Therefore, development teams for products like QuickDraw GX and OpenDoc became core components of the new system. Even development teams for things like user interface skins joined in the chaos. As a result, Copland became an unmanageable project. The plans became longer, the number of projects increased, but progress became less and less. Even if a product was developed, there were not enough testers. In 1995, an engineer pointed out that it was pure fantasy to release Copland in 1996, and it would be good if it could be released in 1997.
By 1996, Gil Amelio was already in power. He happily announced at the Apple Worldwide Developers Conference that the legendary Copland, which was the development version of Mac OS 8, would be released in the summer of that year, and the final version would be delivered to every user in the fall.
Matt Neuburg, an editor at TidBITS, had the opportunity to see this legendary system. To his surprise, the system could only open or close files at the time and could not edit text files. Even the text boxes in the entire user interface couldn't accept input. The system would crash randomly even if nothing was done, and the crashes could even cause file system damage. The Apple employees who participated in the demonstration had to constantly stand by and their job was to format the crashed computer disks and reinstall the system.
Demonstration of Copland on The Computer Chronicles
That summer, the zeroth beta version was sent to a small group of unsuspecting developers, scaring those fragile and inexperienced people half to death. Even within Apple, there were jokes that the official release date of Copland might have to be postponed to 2030.
Gil Amelio was anxious to get Copland back on track. As a good friend of Gil Amelio, Ellen Hancock arrived at Apple during this chaotic time. Her position was to be in charge of this chaotic project. She personally visited each team to understand the situation. After all, she had worked at IBM for nearly thirty years and relied on her exceptional judgment to reach the conclusion within 2-3 months that Copland was hopeless. Given Apple's current state, Copland would never be released, so it would be better to cancel it as soon as possible. In the short term, some useful achievements from Copland would be gradually integrated into the old Mac OS, and efforts would be made to quickly purchase a new system externally to meet Apple's needs.
It was her conclusion that ended Apple's five-year struggle and put the company back on track. The golden age of the PC had already passed, and Apple had a long way to go to turn things around. Gil Amelio supported Ellen Hancock's plan. In August 1996, Apple canceled the Copland project. The CD covers for the development preview version had already been made, and the addresses on each package had been printed, but the CDs were never produced.
From 1996 to 1998, Apple experienced its most chaotic years. In terms of business, there were rumors that Apple was going to be acquired by Sun. What's even more interesting is that in June 1997, Wired magazine published an article titled "101 Ways to Save Apple," which included a suggestion that the best way was for Apple to be bought by Motorola and become a division of Motorola, producing PowerPC series products. From today's perspective, these suggestions are quite ironic and amusing. During this period, Ellen Hancock's outstanding work successfully saved Apple.
Cover of "101 Ways to Save Apple"
Firstly, Ellen Hancock's internal policy was to continue the development work of Mac OS 7.5 and gradually incorporate the technologies from Copland into 7.5. At the same time, a large number of third-party system enhancement packages were purchased, including plugin management tools, hierarchical menus, and other technologies. Apple bought them and integrated them into the existing system. The entire old system was continuously updated while the new system was still under development. By 2000, Mac OS had reached version 9.0, retaining as many old users as possible. Additionally, the aforementioned Blue Box became part of the later Mac OS X system, supporting the running of classic Mac OS programs.
Externally, Ellen Hancock made a big move. She assisted Gil Amelio in searching for operating system technologies outside of Apple. Her experience with IBM and Microsoft's collaboration taught her that buying the rights to use an operating system had many issues, so the best plan was to buy the entire system. Therefore, Gil Amelio started negotiations with Be, which was highly regarded at the time, but the deal fell through due to price issues. In the end, Apple acquired NeXT. After Apple merged with NeXT, NeXTSTEP evolved into Rhapsody and eventually became Mac OS X. We will discuss these events in detail in the future.
After the acquisition of NeXT, Steve Jobs took over, and Gil Amelio was fired due to the significant losses during his tenure as CEO. Steve Jobs gathered trusted individuals (many of whom were former NeXT employees) around him and began a new era, while Ellen Hancock, who was associated with Gil Amelio, was marginalized in the personnel changes. Steve Jobs even called her a "fool" on many occasions. Ellen Hancock eventually resigned voluntarily in 1998. Afterwards, she started a business with Gil Amelio and Steve Wozniak, one of the founders of Apple, but it was never prosperous. Her glorious era had passed.
Gil Amelio summarized his work at Apple by saying:
Apple is like a ship with a hole in the bottom, leaking water, and my job is to get this ship pointed in the right direction.
Although Ellen Hancock, like Gil Amelio, didn't know how to plug the leak, it was her outstanding performance at Apple that not only steered the ship in the right direction but also brought in capable people to fix the leak.
This article was written by Wang Yue, a graduate student in the Computer Science Department at the University of Pennsylvania, a well-known developer of TeX, and an unknown developer of OpenFOAM.
Wang Yue is one of the people I admire most. This series of articles has been published for over ten years, and there is hardly a complete version available online. I hope to further improve it based on the existing content and make it known to more people.