Programming Windows: Meanwhile, in Cupertino (Premium)

In March 1996, Steve Jobs experienced a professional nadir when he appeared on stage at Microsoft’s Professional Developers Conference (PDC). The theme that year was “building Internet applications,” and the software giant promised that executive vice president Paul Maritz would present Microsoft’s strategy for integrating the PC with the Internet how Win32 and OLE developers could “extend their investments in this area.”

At the time, Jobs was on the receiving end of over a decade of failure: his firm NeXT had never attracted a meaningful customer base, forcing him to switch strategies again and again in turn. NeXT originally offered expensive workstation-class computers for the education market, but it was eventually forced to drop the hardware and adapt its advanced, object-oriented NeXTSTEP operating system into a cross-platform programming environment and application layer called OpenSTEP that would run on successful enterprise platforms of the day, including Sun Solaris and Microsoft’s Windows NT.

Despite its business failures, NeXT had some key technical assets, among them a rapid application development (RAD) called Interface Builder that developers could use with the Xcode integrated development environment (IDE), and a web application framework called WebObjects that could create dynamic, data-driven websites at a time when most of the World Wide Web consisted only of static web pages with light gray backgrounds and blinking text. WebObjects, like most NeXT products, was far too expensive and years ahead of its time.

It was also the reason Jobs appeared at PDC just one year after he had testily explained to “60 Minutes” that the company hosting the conference, Microsoft, “had no taste.”

“They don’t think of original ideas, don’t bring culture into their products,” he said, clearly in a testy moment as Microsoft was flying high while NeXT was not. “I don’t have a problem with their success. I just have a problem with the fact that they make really third-rate products. They are very pedestrian … Microsoft is just McDonald’s.”

Onstage at the Internet PDC, Jobs employed the fake humble persona that would later become quite familiar to fans and critics alike after his return to Apple. But on that day in March 1996, Jobs was an outsider, a has-been. And it’s unclear whether the Microsoft-centric crowd at PDC had any idea what to expect. I assume it was eye-opening to most. It certainly was to me, as I ordered a VHS tape of a nearly identical WebObjects presentation by Jobs via the NeXT website that year and I watched it again and again, transfixed. Both by his presentation style and by the futuristic web software he was shilling.

Unfortunately, WebObjects, as noted, wasn’t cheap. Potential customers who called NeXT by phone—the only way to discover pricing in the mid-1990s—were shocked to discover that a single license cost well over $50,000. There were few takers. For that, for OpenSTEP, and for anything else NeXT offered. And as 2001 wore on, it became increasingly clear that NeXT was heading towards insolvency.

And then the miracle happened.

Apple had been trying for years to create a next-generation operating system for the Mac that would include modern features like protected memory and preemptive multitasking. But that effort, codename Copland, had failed due to feature creep, a problem that would later doom Microsoft’s Longhorn. And so Apple took the unusual step of looking outside the company for a replacement, a decision that could only have been made by the inept Gil Amelio, who had succeeded Michael Spindler, an engineer, in early 1996.

Apple’s executive board considered options like Sun Solaris and even Windows NT, a possibility that Bill Gates welcomed eagerly. But then they turned their attention to tiny Be, a startup created by former Apple executive Jean Louis-Gassee. Be’s BeOS had everything Apple needed, including multi-processor support, multithreading support, preemptive multitasking, an advanced file system, and, best of all, advanced multimedia support and a modern user interface. But when a mid-level engineer at NeXT heard about Apple’s plans, he secretly contacted friends at Apple and asked them to consider NeXTSTEP instead. And just like that, it was a two-way race between Be and NeXT.

The history here is well-understood: NeXT won, and Apple acquired the firm for $429 million in cash, with Steve Jobs returning to the company he had co-founded, initially as a part-time technical advisor to Amelio. But the NeXT acquisition was, in effect, a reverse merger, a coup, and Jobs quickly put his top people in charge of Apple’s business divisions, with Avie Tevanian, the software engineer responsible for the Mach microkernel and NeXTSTEP put in charge of Apple’s software design. Jobs also brought back Jon Rubinstein—who would later go on to co-create the iPod and lead Palm during its ill-fated webOS era—to run the hardware side at Apple.

There’s a lot to this story, but of most interest here is Apple’s plans to modernize the Mac OS using NeXTSTEP technology.

“The new operating system is code-named ‘Rhapsody’ and will be based on the merging of technologies from Apple and NeXT Software,” Apple announced in January 1997 as part of series of announcement related to that week’s Macworld event. “As well as providing pioneering next generation technology, Rhapsody aims to provide strong backward compatibility for Mac OS software offering current customers a smooth migration path to the new OS.”

Among other things, Rhapsody would “allow developers to create new applications that leapfrog those of other ‘modern’ operating systems, such as Windows NT,” thanks to NeXT’s object-oriented Interface Builder and Xcode tools. It would feature “an advanced look and feel,” preemptive multitasking, protected memory, and symmetric multiprocessing capabilities, as well as a modern kernel based on Tevanian’s Mach. And it would utilize a fully integrated version of Java.

Apple would also continue to keep the classic Mac OS limping along in tandem with Rhapsody, until the latter was ready for primetime. Both would be optimized for the Power PC hardware Apple was already using, the company said. This strategy was “strongly endorsed” by “industry leaders,” none of whom were named Bill Gates, who had railed against Amelio’s naiveté in adopting NeXT and bringing back Jobs, a person Gates claimed had no technical skills at all.

Rhapsody was eventually bifurcated into three internal efforts. A project called Yellow Box was the closest technical successor to NeXTSTEP and was eventually renamed OS X. Blue Box, meanwhile, became the Classic environment that would allow Yellow Box/OS X to run classic Macintosh apps. And then a little-remembered Red Box project would bring the OpenSTEP/Rhapsody application environment to Windows PCs. Red Box was quickly canceled, but in the early days of Apple’s NeXT acquisition, Apple provided free copies of OpenSTEP to interested parties so they could get started on the coming platform. I briefly experimented with this previously prohibitively expensive product on a Dell laptop.

Over time, Steve Jobs insinuated himself more and more into Apple’s day-to-day operations, and he eventually consummated his coup in mid-1997 by ousting Amelio and becoming the interim CEO, or iCEO, of Apple that September. And for the next few years, Jobs kept the lights on at Apple via various hardware advances, like the iMac and iBook, while Tevanian and company prepped Mac OS X for release.

Jobs even orchestrated an agreement with Bill Gates and Microsoft in which the two companies’ long-standing patent infringement lawsuits were dropped, with Microsoft investing $150 million in Apple and saving it from financial ruin. The agreement was capped by an infamous Macworld Boston event in 1997 during which Bill Gates appeared, virtually, via a gigantic display, dwarfing Jobs onstage to the boos of the audience.

Jobs gave up some important technical and moral ground to Microsoft by agreeing to make Internet Explorer the default browser on the Mac. But that would prove temporary, and, more important to Jobs and Apple, Microsoft had agreed to continue developing new versions of Microsoft Office for the Mac, including versions for Mac OS X. This made the Mac viable to potential customers.

But Mac OS X, like NeXTSTEP before it, was a technical marvel. It would take four long years for Apple to transform Rhapsody into the Mac OS X it first shipped in early 2001. And the team still needed a new look and feel for the product so that it looked new and different and didn’t remind customers of NeXT’s failures. Serendipitously,  it found inspiration in the transparent Bondi Blue elements of the first iMac: Jobs was so taken with this look that the OS X team tried to implement it in software. And that resulted in the Aqua graphical user interface (GUI) used by the first version of Mac OS X. Jobs loved it. Aqua looked so good, Jobs said, “you’ll want to lick it.”

Over a succession of Macworld keynotes and other events, Jobs discussed the evolution of Rhapsody into Mac OS X. He announced the OS X branding and firs showed off its new Aqua UI at Macworld San Francisco in January 2000. And it is perhaps important to consider how new, different, and exciting this interface looked when compared to the more staid look employed by Windows 2000 and 9x/Me, which were the most recent Microsoft releases at the time. Thanks to its higher resolution and more scalable graphics, Mac OS X would likewise outshine Windows XP’s bitmap-based user interface, and XP wouldn’t even ship until late 2001.

Apple had five goals with Mac OS X, Jobs said. It would mark the beginning of a single OS strategy at Apple, with the Classic OS eventually disappearing. It would include state-of-the-art technical plumbing, thanks to NeXT’s Mach kernel (now called Darwin) and NeXTSTEP underpinnings.  “Killer” graphics. Integrated Internet standards support throughout. And a gentle migration strategy to get Apple’s 25 million Mac users to move forward to the new platform.

Jobs promised a 12-month rollout for Mac OS X, and he nearly hit that mark, as the somewhat incomplete first version of the product landed in March 2001. That said, Mac OS X was not, even then, ready for most users, and so new Macs in 2001 shipped with both Mac OS Classic and OS X in a dual-boot configuration, and the Classic version was still the default.

Architecturally, OS X offered its “killer” graphics via a layer consisting of Quartz, OpenGL and QuickTime multimedia capabilities. Quartz was based on PDF and provided anti-aliasing and compositing, a technique Microsoft would later promise for Longhorn but not deliver until 2006. OpenGL provided hardware-accelerated 3D graphics, another feature Windows wouldn’t see until 2006.

OS X would offer three major application programming interfaces, or APIs, above the graphics and kernel layers. The Classic API would allow classic Mac OS to run apps as-is, without modification. Carbon was an interim API that let developers add new OS X features to existing classic Mac OS apps; these updated apps were said to be “carbonized.” And Cocoa brought the object-oriented NeXTSTEP/OpenSTEP APIs to the Mac in Java and Objective-C forms.

The new Aqua UI sat on top of those other layers, offering users an interesting amalgamation of classic Mac OS and NeXTSTEP user interfaces, plus some new touches that were unique to OS X. “We call that new user interface Aqua because it’s liquid,” Jobs explained of the look and feel. Aqua would feature blue, transparent, and liquid-inspired interface controls, windows with rounded corners and shadows that always correctly displayed their contents while resizing and moving, mouse rollover effects, and transparency and translucency throughout.

Instead of a Windows-like taskbar, Mac OS X would feature a Dock like NeXTSTEP, now centered and at the bottom of the screen, for pinned and running application shortcuts. And it would of course feature mouse-over animations that thrilled the crowd. Windows would minimize with animations so stunning that Jobs slowed it down again and again so the audience could appreciate the technical sophistication.

“We wanted to give a much more powerful user interface to our pro customers,” Jobs said. “But at the same time, we wanted to make this the dream user interface for someone who has never even touched a computer before. That’s really hard to do.”

Despite being a minority player in a market dominated by Microsoft, or perhaps because of that, Jobs and Apple never wasted an opportunity to belittle their much more successful rival. At the Mac OS X and Aqua reveal in early 2000, he made fun of how Microsoft copied the Mac OS UI with Windows without mentioning that Apple had, before that, done the same thing to Xerox. But as Mac OS X and Windows advanced in lockstep throughout the 2000s, Apple’s digs against Microsoft would grow more and more antagonistic.

We’ll examine those events in the future. For now, it’s more important to understand the shot across Microsoft’s bow that Mac OS X represented at the start of the 21st century. And how Apple’s technical masterstroke would impact the development of Windows in the early 2000s. It was too late for Microsoft to stop what was happening with Windows XP—the company was still desperate to consolidate its NT- and 9x-based Windows versions at that time—but the software giant was determined to take a major step forward, and hopefully leapfrog Mac OS X, with Longhorn. The battle for the future was now underway.