Steve Jobs and the History of Cocoa, Part Two

Editor's note: In Part One, Simson Garfinkel and Michael Mahoney explain why Cocoa and Mac OS X aren't nearly as revolutionary as they are evolutionary, beginning with Apple's genesis in the 1970s. Now they pick up the story with the Star Trek project and bring you to the current iteration of Mac OS X.

Star Trek, Be, and a New Hope

NeXT wasn’t the only burgeoning operating system company to have its hopes dashed. Another such company was Apple itself.

In the summer of 1992, a group of 18 engineers -- 14 from Apple, 4 from Novell -- moved into an office in Santa Clara, Calif., on a secret mission: port the Macintosh operating system to Intel-based hardware. The project, dubbed “Star Trek,” was designed to beat back the threat of Microsoft by giving PC makers an easy-to-use operating system they could install on their own hardware.

Better than Windows 3.1 and with the whole suite of Macintosh applications, Star Trek could have beaten Windows 95 to the market by more than a year and turned Apple into the dominant market player.

The Star Trek group pulled off its mission, delivering a working operating system by Oct. 31. But shortly thereafter, the Star Trek group ran into problems. One reason was politics: a key supporter of the project, Apple Vice President Roger Heinen, was hired away by Microsoft. Then Dell Computer told Apple that Dell would not be able to pay any money for Apple’s product, since Dell was already purchasing a copy of Windows for every computer that Dell shipped -- whether or not Windows was actually installed on those systems.

Reportedly, yet another reason Apple got cold feet was that Apple was right in the middle of transitioning from the Motorola 68040 microprocessor to the PowerPC-based systems -- and Star Trek running on an Intel 486-based system ran faster than Macintosh System 7 running on a PowerPC.

Apple killed Star Trek in 1993, just as another operating system project was moving into high gear. Called Copeland, this project was to build a next-generation Macintosh operating system that overcame the Mac’s legendary problems with stability and extension conflicts. Apple’s goal was to bring advanced features like memory protection and pre-emptive multitasking to the Macintosh operating system while maintaining a high degree of application compatibility.

Copeland sucked tremendous money, resources, and time out of Apple and ultimately delivered … nothing. After years of promises and untold millions (billions?) spent on the project, Apple shelved its attempts at building its own next-generation operating system and started looking outside the company.

Alas, there weren’t a lot of choices for a consumer-grade operating system. After abortive talks with Sun, Apple settled on an innovative, fast, exciting operating system called BeOS from a company called Be, Inc.

Apple reportedly announced it was in serious discussions with Be, which at the time was manufacturing a sleek, blue computer tower equipped with two PowerPC microprocessors. By December 1996 rumors were flying that an Apple buyout of Be was imminent but Apple wasn’t offering Be enough money. Then suddenly, Apple stopped talking to Be and started talking with its founder, Steve Jobs.

Within a few weeks, the terms of the deal were clear. Apple purchased NeXT for $400 million. Apple would use the NeXTSTEP operating system as the basis of Apple’s next-generation Macintosh OS. Jobs, meanwhile, would be hired as Apple’s “interim” CEO.

At the January 1997 Macworld trade show, Jobs addressed the Macintosh faithful for the first time since his ouster 12 years previous. He said his engineers would merge the Macintosh and NeXT operating systems into a unified operating system called “Rhapsody” that would be based on the OpenStep software. Every computer Apple was selling at the time would run this new operating system. And he promised it would be available for purchase in two years.

Getting NeXTSTEP to run on Apple hardware proved to be quite simple -- NeXT’s engineers had a lot of experience in porting their operating system to new hardware platforms -- but getting it to run on Apple’s complete product line proved to be an insurmountable task.

According to insiders, one of the problems was a lack of discipline among Apple’s hardware designers. Because Apple hardware never had to run any operating system other than Apple System 7 and System 8, many of the interfaces were not particularly clean or even documented. Largely as a result of the move to Mac OS X, Apple now sells far fewer kinds of computers today than it did three years ago.

Nevertheless, schedules slipped. The engineers at Apple tried very hard to make the Rhapsody deadline Jobs had set forth, but things kept cropping up. One of them was Java, which started getting very popular in 1997. NeXTSTEP was written in Objective-C, but in 1997 Java looked like a new language that could eliminate the disparity between the Macintosh operating system and Windows.

So Apple adopted Java whole-hog, and spent countless man years making it so that Rhapsody programs could be written in either Objective-C or Java, just to be safe. Another cause of delay was the underlying difficulty of integrating Macintosh System 7 with NeXTSTEP. Apple’s customers made it clear they would not accept an operating system in which all the old Mac programs ran in a single window of a “virtual Macintosh” -- they wanted the old Mac programs to be first-class players on the Mac desktop. This proved to be a considerably harder task than it first appeared.

A year later, when the original deadline came around, it was clear that the Rhapsody computer system was not ready for prime time. Rather than release a product nobody would like, Apple pulled back and started working on its next “next-generation” operating system.

Aspects of the NeXTSTEP operating system proved unsuitable to the world of Macintosh. The NeXTSTEP Application Kit was given a facelift to become more “Mac-like.” Unable or unwilling to cut a deal with Adobe for Display PostScript, Apple decided to remove it from the operating system and replace it with a next-generation drawing subsystem called Quartz. (The Quartz system actually borrows heavily from Apple’s QuickDraw GX system, but uses Adobe’s Portable Document Format as an interchange file format. Nevertheless, Quartz implements PDF without a line of Adobe code -- and thus without a cent of royalties due to Adobe.)

The deadline for the Rhapsody system came and went. Finally, Apple announced it would bring out one more iteration of the original Macintosh operating system -- OS 9 -- and then begin the transition to the new OS X operating system in the summer of 2001.

Mac OS X, Ready or Not

Many Macintosh purists insisted that Mac OS X was not a real Macintosh operating system. But that just begs the question -- “What is a real Macintosh operating system?”

The original Macintosh operating system could run just a single program at a time on a computer with 64K of ROM and 64K of RAM. (The 128K Mac was Apple’s second-generation Macintosh computer.) It didn’t support up or down arrow keys because Apple’s dogmatic engineers insisted that anything done with a keyboard could be done better with a mouse.

In fact, Mac OS X is very much a Macintosh operating system. It may have a Unix kernel, but it’s a Unix kernel sitting on top of an Apple HFS+ file system, complete with creator codes and resource forks. Mac OS X still supports AppleTalk, Apple’s easy-to-administer, automatically configuring networking system.

What’s more, Mac OS X really does run all of those old Macintosh applications -- which is what the "Classic" Macintosh environment is all about. If you double-click on an application program that runs on older Mac OS computers, a Mac OS X system will launch a copy of Mac OS 9 within Mac OS X. When you activate this application, the Macintosh computer will have much of the look and feel of a Mac OS 9 computer. It’s weird, but you can run those old apps quite well in Classic mode.

Building Cocoa Applications

This spring, O’Reilly & Associates will publish Building Cocoa Applications -- A Step-by-Step Guide. This book is based on literally years of experience with the NeXTSTEP, Macintosh, Unix, and Cocoa environments. In it, we try to show users a step-by-step approach to building realistic applications. And we try to do it while having a lot of fun.

Our book does not assume prior knowledge with the Macintosh or any other window-based operating environment, nor does it assume knowledge of object-oriented programming. In fact, a previous version of the book was actually used in several schools to teach object-oriented programming back in the 1990s.

Building Cocoa Apps begins with introductory but important material on the Macintosh operating system and object-oriented programming. Then the rest of the book concentrates on building three major applications. The first, Calculator, is built in Chapters 5-8. The second, MathPaper, is built within Chapters 10-14. The third and final major application is GraphPaper, built in Chapters 16-21.

There are numerous additional simple applications built throughout the book to demonstrate features of Cocoa and Mac OS X. You can build all of these applications right along with us. We provide simple but complete instructions on how to do whatever is necessary to build these applications from scratch.

If you are not sure whether our book is for you, you can download the Calculator, MathPaper, and GraphPaper applications from the O’Reilly Web site. Run the programs, then read the code. If you have always wanted to write programs with super-slick user interfaces, now is the time to start.

Michael Mahoney is Professor and Chair of the Computer Engineering and Computer Science Department at California State University, Long Beach.

With so many OS screwups, it's hard to understand why the Apple folks are always so arrogant.

"Pride goes before destruction and a haughty spirit before a fall."

If ONLY this were successful! I would have loved to had seen this.

With so many OS screwups, it's hard to understand why the Apple folks are always so arrogant.

With all due respect my friend, I have detected no arrogance from Apple folk in either of these posts, only from our Microsoft/Linux friends. (And please excuse me if your reference was to corporate arrogance, but there also I think there are equal portions.)

There is much to like about Macs, but as long as the OS only runs on Apple hardware, it will never pose a threat to Windows. Sometimes I think that Apple's main strategist is actually a Microsoft employee.

I would have agreed a few years back, but if Apple is first a hardware company, then it can never port its OS to Intel, for as the clone experiment proved, we will switch to a cheaper box in an instant.

Yep, you nailed it.

Sun has the same problem. Sloaris on Intel with industry standard drives and interfaces would be more than enough for most of their customers (ask Google) but that would destroy their business. Scott M. must be sweating these days. There are only so many customers that need fault tolerant 64 processor machines ...

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.