Darwin in the background

Here we need to make another detour towards the Darwin operating system. When Apple bought Jobs’ NeXT company in 1997, the transaction naturally included the NeXTSTEP operating system and its variant created in cooperation with Sun Microsystems and called OpenSTEP. The NeXTSTEP operating system was also supposed to become the basis for Apple’s new operating system, which was one of the reasons Apple bought Jobs’ NeXT. The attractive and perhaps underappreciated charm of NeXTSTEP at the time was its multi-platform capability; the system could run on the Intel x86 platform as well as on Motorola 68K, PA-RISC, and SPARC, i.e., virtually all processors used by desktop platforms at the time. It was also possible to create distribution files containing binary versions of the program for all processor platforms, known as fat binaries.

The legacy of NeXT thus served as the basis for the development of a new operating system called Rhapsody, which Apple first unveiled at a developer conference in 1997. This system brought a number of changes compared to previous versions of Mac OS, the most notable of which, in our view, were the following:

•The kernel and related subsystems were based on Mach and BSD.

•subsystem compatible with the previous Mac OS (Blue Box) – later known as the Classic interface

•extended implementation of the OpenStep API (Yellow Box) – later developed into Cocoa.

•Java virtual machine

•window system based on Display PostScript

•interface based on Mac OS, but combined with OpenSTEP

Apple planned to transfer most of the software frameworks from Mac OS, such as QuickTime, QuickDraw 3D, QuickDraw GX, and ColorSync, as well as file systems from the original Apple computers, Apple Filing Protocol (AFP), HFS, UFS, and others, to Rhapsody. However, it soon became apparent that this was no easy task. The first developer release (DR1) in September 1997 was followed by a second DR2 in May 1998, but there was still a lot of work to be done. The first developer preview (Developer Preview 1) came a year later, in May 1999, and by then the system was called Mac OS X. A month earlier, Apple had spun off the server version, Mac OS X Server 1, which it officially released, as well as an opensource version Darwin, thereby fulfilling (to a large extent, which was heavily criticized and debated) the condition of releasing the source code of the system, which uses other open source components that require this and which Apple included in its system when it was based on the Mach and BSD kernels.

Darwin is essentially Mac OS X without a graphical interface and without a number of proprietary libraries, such as FairPlay music file security. You can download it; since later versions are only available as source files, not binary versions, you can compile them and run them as an operating system on a wide range of processor platforms. Darwin will serve two purposes for Apple in the future: it will be a constant reminder that porting the Mac OS X operating system to another processor platform will not be so difficult as to be impossible. And it will be a response to objections that Apple software is closed and proprietary, an impression that Apple will later give rise to, especially in Europe. In America, where it is more widespread in education and Darwin is commonly used on many school servers, awareness of the openness and use of standard components in Apple software is much greater. Darwin is still the core of every Mac OS X system and has a fairly large group of contributors to its open source development, with this development feeding back into the core of Mac OS X.

The first stable version of Mac OS X 10.0, nicknamed Cheetah, was released in March 2001, four years after the start of development of Rhapsody, which was expected to be easily converted for use on the Apple platform. This irony caused the company a number of problems, as it forced its users to use the unsatisfactory and unpromising Mac OS platform for those four years.

Darwin thus became the basis for the operating system within the Purple 2 project. At a time when it was unclear whether Apple would opt for ARM processors, which it had helped design, or Intel, which it had just started using in desktops, this was a very wise choice, as it allowed the processor platform to be changed without much pain, as Apple was doing with PowerPC and Intel. In addition, it was a compact and proven system to which an interface (API) needed to be added—in this case, Cocoa Touch, a touch-optimized OpenSTEP API with a library for working with mobile phones.

Ultimately, a design was created that divided the system into four abstraction layers:

•system kernel layer

•core services layer

•media layer

•Cocoa Touch touch interface layer

Why was this important and noteworthy? Jobs believed that a mobile phone must respond perfectly to user requests. If the user presses a button, the phone must respond. It must clearly confirm that it has accepted the user’s input, and the best way to do this is by performing the desired function. One of the developers demonstrated this approach to Jobs on a Nokia phone running Symbian, where the phone responded too slowly when a selection was pressed. The user swiped a name in the list and accidentally called another name. This was frustrating for Jobs, and he didn’t want to see anything like that on his mobile phone. The operating system had to prioritize processing the user’s selection, and the Cocoa Touch touch interface had the highest priority in the system. Only after that did the other layers of the system have priority. If the user made a selection or input, something had to happen to reassure the user that everything was proceeding smoothly. Another argument for this approach was the “bouncing icons” in desktop Mac OS X. When the user launched a program from the system dock, nothing visible usually happened for a while until the program was loaded from the disk into the computer’s operating memory. Users would click on the icon repeatedly because they didn’t know that the program was already loading into memory. The developers solved this by making the icon bounce until the program was fully loaded into memory. In the mobile version, it was necessary for the system to respond immediately to any user input in a similar way.

This approach subsequently became so ingrained in the mobile system that even individual functions within Cocoa Touch are processed in the system with different priority classes so that the user has the best possible impression of the phone running smoothly.

At that time, Apple was not yet seriously considering the possibility of running third-party applications on the phone. In fact, it was not even desirable at that time. The upcoming operating system was, of course, fully supportive of preemptive multitasking, memory protection, and other advanced features of modern operating systems, which was in contrast to other operating systems at the time that struggled with memory protection (Symbian), multitasking (Palm OS), or both (Win CE). However, Jobs primarily viewed the upcoming mobile phone as a device intended for consuming music supplied by Apple. Third-party applications would only slow things down, and Jobs realized that a number of details would need to be worked out, such as the distribution system, so even though mobile OS X natively supported the ability to run additional applications in the background, Apple artificially limited this option. When the iPhone hit the market, only “jailbroken” phones, i.e., phones that had been stripped of this protection, could install emerging third-party applications. Long after the iPhone’s launch in January 2007, Jobs assumed that developers would create exclusively web-based applications and that only Apple would create native applications.

However, in the summer of 2006, the development of the mobile version of OS X was still in a completely unsatisfactory state. Although the basic porting of the system was completed in record time by a team of only two engineers, the integration and coordination of the individual elements of the mobile phone interface was dire. Calls dropped, the software often crashed, and battery life was unreasonably low. While 200 people were working on the project in September 2005, that number quickly grew to 500 in two parallel teams, but it was still not enough. A serious disadvantage was the secrecy in which Apple worked: new people could not be recruited through public recruitment, but only through recommendations, often via intermediaries. For example, the testing part of the software team was largely virtual, with prototyping and testing carried out by people who communicated primarily by email and did not even know they were working for Apple for a long time. That’s how secretive it was.

There is a story about one of the user interface development project managers who believed he was working on interface changes in the Symbian system licensed by Samsung – that’s how strict the secrecy was. He worked remotely and had a contract with a law firm representing the “customer.” One day, his supervisor asked him for his cell phone number, saying that a message with recommendations for the interface had arrived on the desk of the top boss, who had a habit of asking the people working on the task directly for details that interested him. When the developer working from Europe received a late-night call from an American number and the caller began asking why he had not recommended one of the options due to legal issues, the developer pointed out that this method infringed on Apple’s patents and that he doubted Jobs would sell the rights. There was a moment of silence on the line, and then the caller asked with a laugh, “You have no idea who’s calling you, do you?”

Since spring 2006, Apple had intensified its recruitment of developers, but still had limited capacity available, which contrasted with Steve Jobs signing a cooperation agreement with Cingular in July 2006. Cingular was to invest substantial funds in expanding its mobile network and supporting data services, as well as investing a significant amount in supporting and selling the new Apple phone. In return, it was to receive a share of the money earned from orders placed on the iTunes Store and five years of exclusivity on the US market for the sale of the phone. Along with this, Cingular committed to working with Apple on the development of a visual voicemail system. Apple was betting that users would want to use voicemail in the same way they used email, and that, along with email and a web browser, it would be one of the most interesting new features added to the phone and iPod. Cingular committed to launching the mobile phone and investing millions of dollars in its promotion, but no one in the company’s management had actually seen the mobile phone itself. Such was the belief in Apple’s success and Steve Jobs’ persuasiveness, even though Cingular had already had one bad experience with the Rokr project, which was led by Motorola.

The contract with Cingular was supposed to remind Jobs of the seriousness of the situation. Apple needed a new selling point, otherwise the iPod and iTunes revenues would slowly disappear, replaced by mobile phones equipped with MP3 players. When even another wave of recruitment failed to improve the situation in the software development team, Jobs took an unusual step: he transferred the developers who were working on the new version of the Mac OS X 10.5 Leopard desktop operating system to the team developing the mobile version. This kick-started its development, but delayed the regular release of new versions by six months. When Jobs explained this event a year later at a meeting with investors, he emphasized that it was his decision, and he was rewarded with thunderous applause. The courage with which he prioritized the development of the mobile system paid off, but it also meant that many users of the Mac desktop platform would be unable to shake the impression that computers were now taking a back seat to Apple’s mobile platform.

In November 2006, at the launch of Cingular Music, Cingular CEO Robert Hyatt responded to journalists’ questions about whether he was planning to launch the iPhone with Apple by saying that he knew nothing about it and had never seen such a device. He wasn’t lying. It wasn’t until December 11, 2006, that Steve Jobs showed a prototype of the iPhone to Cingular representatives at the Four Seasons hotel in Las Vegas. It was already a relatively stable prototype that Jobs had been using personally for several weeks, which, incidentally, resulted in the need to replace the plexiglass cover of the display with real glass. The plastic cover scratched easily, which Jobs hated, so he pushed for the search for “gorilla glass.” However, even when the phone was unveiled in January 2007, it was not clear whether it would be possible to replace the glass in mass-produced phones, as drop tests had not been completed.

By mid-December, Jobs was convinced that the company was capable of resolving all potential problems with launching the product and that it would be possible to unveil the phone at MacWorld in January, the most appropriate and dignified date for presenting this sensational new product eagerly awaited by fans. Although development was far from complete, he decided on another trick to keep the attention of an already tense audience and market: he would introduce the product but not immediately put it on sale with the announcement, as had been the custom until then. Instead, the company would allow itself a six-month delay, citing the necessary approval process by the US regulator, the FCC.

The context of this story

Darwin in the background

Table of contents