There were now two competing projects vying to become the iPhone—a “bake-off,” as some engineers put it. The two phone projects were split into tracks, code-named P1 and P2, respectively. Both were top secret. P1 was the iPod phone. P2 was the still-experimental hybrid of multitouch technology and Mac software.
If there’s a ground zero for the political strife that would later come to engulf the project, it’s likely here, in the decision to split the two teams—Fadell’s iPod division, which was still charged with updating that product line in addition to prototyping the iPod phone, and Scott Forstall’s Mac OS software vets—and drive them to compete. (The Human Interface designers, meanwhile, worked on both P1 and P2.)
Eventually, the executives overseeing the most important elements of the iPhone—software, hardware, and industrial design—would barely be able to tolerate sitting in the same room together. One would quit, others would be fired, and one would emerge solidly—and perhaps solely—as the new face of Apple’s genius in the post-Jobs era. Meanwhile, the designers, engineers, and coders would work tirelessly, below the political fray, to turn the Ps into working devices in any way possible.
The Purple People Leader
Every top secret project worth its salt in intrigue has a code name. The iPhone’s was Purple.
“One of the buildings we have up in Cupertino, we locked it down,” said Scott Forstall, who had managed Mac OS X software and who would come to run the entire iPhone software program. “We started with one floor”—where Greg Christie’s Human Interface team worked—“We locked the entire floor down. We put doors with badge readers, there were cameras, I think, to get to some of our labs, you had to badge in four times to get there.” He called it the Purple Dorm because, “much like a dorm, people were there all the time.”
They “put up a sign that said ‘Fight Club’ because the first rule of Fight Club in the movie is that you don’t talk about Fight Club, and the first rule about the Purple Project is you do not talk about that outside of those doors,” Forstall said.
Why Purple? Few seem to recall. One theory is it was named after a purple kangaroo toy that Scott Herz—one of the first engineers to come to work on the iPhone—had as a mascot for Radar, the system that Apple engineers used to keep track of software bugs and glitches throughout the company. “All the bugs are tracked inside of Radar at Apple, and a lot of people have access to Radar,” says Richard Williamson. “So if you’re a curious engineer, you can go spelunking around the bug-tracking system and find out what people are working on. And if you’re working on a secret project, you have to think about how to cover your tracks there.”
Scott Forstall, born in 1969, had been downloading Apple into his brain his entire life. By junior high, his precocious math and science skills landed him in an advanced-placement course with access to an Apple IIe computer. He learned to code, and to code well. Forstall didn’t fit the classic computer-geek mold, though. He was a debate-team champ and a performer in high-school musicals; he played the lead in Sweeney Todd, that hammy demon barber. Forstall graduated from Stanford in 1992 with a master’s in computer science and landed a job at NeXT.
After releasing an overpriced computer aimed at the higher-education market, NeXT flailed as a hardware company, but it survived by licensing its powerful NeXTSTEP operating system. In 1996, Apple bought NeXT and brought Jobs back into the fold, and the decision was made to use NeXTSTEP to overhaul the Mac’s aging operating system. It became the foundation on which Macs—and iPhones—still run today. At Jobs-led Apple, Forstall rose through the ranks. He mimicked his idol’s management style and distinctive taste. BusinessWeek called him “the Sorcerer’s Apprentice.”
One of his former colleagues praised him as a smart, savvy leader but said he went overboard on the Jobs-worship: “He was generally great, but sometimes it was like, just be yourself.” Forstall emerged as the leader of the effort to adapt Mac software to a touchscreen phone. Though some found his ego and naked ambition distasteful—he was “very much in need of adulation,” according to one peer, and called “a starfucker” by another—few dispute the caliber of his intellect and work ethic. “I don’t know what other people have said about Scott,” Henri Lamiraux says, “but he was a pleasure to work with.”
Forstall led many of the top engineers he’d worked with since his NeXT days—Henri Lamiraux and Richard Williamson among them—into the P2 project. Williamson jokingly called the crew “the NeXT mafia.” True to the name, they would at times behave in a manner befitting a close-knit, secretive (and highly efficient) organization.
P1 Thing After Another
Tony Fadell was Forstall’s chief competition.
“From a politics perspective, Tony wanted to own the entire experience,” Grignon says. “The software, the hardware… once people started to see the importance of this project to Apple, everyone wanted to get their fingers in it. And that’s when the epic fight between Fadell and Forstall began.”
Having worked with Forstall on Dashboard, Grignon was in a unique position to interface with both groups. “From our perspective, Forstall and his crew, we always viewed them as the underdogs. Like they were trying to wedge their way in,” Grignon says. “We had complete confidence that our stack was going to happen because this is Tony’s project, and Tony’s responsible for millions upon millions of iPod sales.”
So, the pod team worked to produce a new pod-phone cut from the mold of Apple’s ubiquitous music player. Their idea was to produce an iPod that would have two distinct modes: the music player and a phone. “We prototyped a new way,” Grignon says of the early device. “It was this interesting material… it still had this touch sensitive click wheel, right, and the Play/Pause/Next/Previous buttons in blue backlighting. And when you put it into phone mode through the UI, all that light kind of faded out and faded back in as orange. Like, zero to nine in the click wheel in an old rotary phone, you know, ABCDEGF around the edges.” When the device was in music-playing mode, blue backlighting would show iPod controls around the touch wheel. The screen would still be filled with iPod-style text and lists, and if you toggled it to phone mode, it’d glow orange and display numbers like the dial of a rotary phone.
“We put a radio inside, effectively an iPod Mini with a speaker and headphones, still using the touch-wheel interface,” Tupman says.
“And when you texted, it dialed—and it worked!” Grignon says. “So we built a couple hundred of them.”
The problem was that they were difficult to use as phones. “After we made the first iteration of the software, it was clear that this was going nowhere,” Fadell says. “Because of the wheel interface. It was never gonna work because you don’t want a rotary dial on the phone.”
The design team tried mightily to hack together a solution.
“I came up with some ideas for the predictive typing,” Bas Ording says. There would be an alphabet laid out at the bottom of the screen, and users would use the wheel to select letters. “And then you can just, like, click-click-click-click—‘Hello, how are you.’ So I just built an actual thing that can learn as you type—it would build up a database of words that follow each other.” But the process was still too tedious.
“It was just obvious that we were overloading the click wheel with too much,” Grignon says. “And texting and phone numbers—it was a fucking mess.”
“We tried everything,” Fadell says. “And nothing came out to make it work. Steve kept pushing and pushing, and we were like, ‘Steve.’ He’s pushing the rock up a hill. Let’s put it this way: I think he knew, I could tell in his eyes that he knew; he just wanted it to work,” he says. “He just kept beating this dead horse.”
“C’mon, there’s gotta be a way,” Jobs would tell Fadell. “He didn’t just want to give up. So he pushed until there was nothing there,” Fadell says.
They even filed for a patent for the ill-fated device, and in the bowels of Cupertino, there were offices and labs littered with dozens of workin
g iPod phones. “We actually made phone calls,” Grignon says.
The first calls from an Apple phone were not, it turns out, made on the sleek touchscreen interface of the future but on a steampunk rotary dial. “We came very close,” Ording says. “It was, like, we could have finished it and made a product out of it.… But then I guess Steve must have woken up one day like, ‘This is not as exciting as the touch stuff.’”
“For us on the hardware team, it was great experience,” David Tupman says. “We got to build RF radio boards, it forced us to select suppliers, it pushed us to get everything in place.” In fact, elements of the iPod phone wound up migrating into the final iPhone; it was like a version 0.1, Tupman says. For instance: “The radio system that was in that iPod phone was the one that shipped in the actual iPhone.”
Hands Off
The first time Fadell saw P2’s touch-tablet rig in action, he was impressed—and perplexed. “Steve pulled me in a room when everything was failing on the iPod phone and said, ‘Come and look at this.’” Jobs showed him the ENRI team’s multitouch prototype. “They had been getting, in the background, the touch Mac going. But it wasn’t a touch Mac; literally, it was a room with a Ping-Pong table, a projector, and this thing that was a big touchscreen,” Fadell says.
“This is what I want to put on the phone,” Jobs said.
“Steve, sure,” Fadell replied. “It’s not even close to production. It’s a prototype, and it’s not a prototype at scale—it’s a prototype table. It’s a research project. It was like eight percent there,” Fadell says.
David Tupman was more optimistic. “I was like, ‘Oh, wow, yeah, we have to find out a way to make this work.’” He was convinced the engineering challenges could be solved. “I said, ‘Let’s just sit down and go through the numbers and let’s work it out.’”
The iPod phone was losing support. The executives debated which project to pursue, but Phil Schiller, Apple’s head of marketing, had an answer: Neither. He wanted a keyboard with hard buttons. The BlackBerry was arguably the first hit smartphone. It had an email client and a tiny hard keyboard. After everyone else, including Fadell, started to agree that multitouch was the way forward, Schiller became the lone holdout.
He “just sat there with his sword out every time, going, ‘No, we’ve got to have a hard keyboard. No. Hard keyboard.’ And he wouldn’t listen to reason as all of us were like, ‘No, this works now, Phil.’ And he’d say, ‘You gotta have a hard keyboard!’” Fadell says.
Schiller didn’t have the same technological acumen as many of the other execs. “Phil is not a technology guy,” Brett Bilbrey, the former head of Apple’s Advance Technology Group, says. “There were days when you had to explain things to him like a grade-school kid.” Jobs liked him, Bilbrey thinks, because he “looked at technology like middle America does, like Grandma and Grandpa did.”
When the rest of the team had decided to move on multitouch and a virtual keyboard, Schiller put his foot down. “There was this one spectacular meeting where we were finally going in a direction,” Fadell says, “and he erupted.”
“We’re making the wrong decision!” Schiller shouted.
“Steve looked at him and goes, ‘I’m sick and tired of this stuff. Can we get off of this?’ And he threw him out of the meeting,” Fadell recalls. Later, he says, “Steve and he had it out in the hallway. He was told, like, Get on the program or get the fuck out. And he ultimately caved.”
That cleared it up: the phone would be based on a touchscreen. “We all know this is the one we want to do,” Jobs said in a meeting, pointing to the touchscreen. “So let’s make it work.”
Round Two
“There was a whole religious war over the phone” between the iPod team and the Mac OS crew, one former Apple executive told me. When the iPod wheel was ruled out and the touch ruled in, the new question was how to build the phone’s operating system. This was a critical juncture—it would determine whether the iPhone would be positioned as an accessory or as a mobile computer.
“Tony and his team were arguing we should evolve the operating system and take it in the direction of the iPod, which was very rudimentary,” Richard Williamson says. “And myself and Henri and Scott Forstall, we were all arguing we should take OS X”—Apple’s main operating system, which ran on its desktops and laptops—“and shrink it down.”
“There were some epic battles, philosophical battles about trying to decide what to do,” Williamson says.
The NeXT mafia saw an opportunity to create a true mobile computing device and wanted to squeeze the Mac’s operating system onto the phone, complete with versions of Mac apps. They knew the operating system inside and out—it was based on code they’d worked with for over a decade. “We knew for sure that there was enough horsepower to run a modern operating system,” Williamson says, and they believed they could use a compact ARM processor—Sophie Wilson’s low-power chip architecture—to create a stripped-down computer on a phone.
The iPod team thought that was too ambitious and that the phone should run a version of Linux, the open-source system popular with developers and open-source advocates, which already ran on low-power ARM chips. “Now we’ve built this phone,” says Andy Grignon, “but we have this big argument about what was the operating system it should be built on. ’Cause we were initially making it iPod-based, right? And nobody cares what the operating system in an iPod is. It’s an appliance, an accessory. We were viewing the phone in that same camp.”
Remember, even after the iPhone’s launch, Steve Jobs would describe it as “more like an iPod” than a computer. But those who’d been in the trenches experimenting with the touch interface were excited about the possibilities it presented for personal computing and for evolving the human-machine interface.
“There was definitely discussion: This is just an iPod with a phone. And we said, no, it’s OS X with a phone,” Henri Lamiraux says. “That’s what created a lot of conflict with the iPod team, because they thought they were the team that knew about all the software on small devices. And we were like, no, okay, it’s just a computer.”
“At this point we didn’t care about the phone at all,” Williamson says. “The phone’s largely irrelevant. It’s basically a modem. But it was ‘What is the operating system going to be like, what is the interaction paradigm going to be like?’” In that comment, you can read the roots of the philosophical clash: The software engineers saw P2 not as a chance to build a phone, but as an opportunity to use a phone-shaped device as a Trojan horse for a much more complex kind of mobile computer.
The Incredible Shrinking Operating System
When the two systems squared off early on, the mobile-computing approach didn’t fare so well.
“Uh, just the load time was laughable,” Andy Grignon says. Grignon’s Linux option was fast and simple. “It’s just kind of prrrrrt and it’s up.” When the Mac team first got their system compiling, “it was like six rows of hashtags, dink-dink-dink-dink-dink, and then it just sat there and it would shit the bed for a little bit, and then it would finally come back up and you’d be like, Are you even kidding me? And this is supposed to be for a device that just turns on? Like, for real?”
“At that point it was up to us to prove” that a variant of OS X could work on the device, Williamson says. The mafia got to work, and the competition heightened. “We wanted our vision for this phone that Apple was going to release to become a reality,” Nitin Ganatra says. “We didn’t want to let the iPod team have an iPod-ish version of the phone come out before.”
One of the first orders of business was to demonstrate that the scrolling that had wowed Jobs would work with the stripped-down operating system. Williamson linked up with Ording and hashed it out. “It worked and looked amazingly real. When you touched the screen, it would track your finger perfectly, you would pull down, it would pull down.”
That, Williamson says, put the nail in the Linux pod’s coffin. “Once we had OS X ported and these basic scro
lling interactions nailed, the decision was made: We’re not going to go with the iPod stack, we’re going to go with OS X.”
The software for the iPhone would be built by Scott Forstall’s NeXT mafia; the hardware would go to Fadell’s group. The iPhone would boast a touchscreen and pack the power of a mobile computer. That is, if they could get the thing to work.
Fadell looked at the multitouch contraption again. “I didn’t say, ‘Sure,’ and I didn’t say, ‘No.’ I said, ‘Okay, we’ve got a lot of things to work on,’” he says. “We had to create a whole basically separate company just to build the prototype.”
Long after its launch, the iPhone would not only require the creation of such “separate companies” inside Apple, it would lead to the absorption of entirely new ones outside it. It would lead to new breakthroughs, new ideas, new hurdles. These next chapters engage with the world reorganized by the iPhone—from the ascent of Siri and the Secure Enclave, to the making, marketing, and trashing of the one device.
CHAPTER 10
Hey, Siri
Where was the first artificially intelligent assistant born?
Of all the places I might have expected to have a deep conversation with one of the founders of Siri about the evolving state of artificial intelligence, a cruise ship circling Papua New Guinea was not a frontrunner. But here we are, in my cabin on the National Geographic Orion, the buzz of its engine providing a suitably artificial thrum to the backdrop of our talk, a tropical green-blue expanse out the window.
The One Device Page 23