This week marks the 10th anniversary of the original iPhone introduction, and with it, a flood of stories about the design and development of the device. One video from Sonny Dickson, showing what looks like an iPod interface with a virtual click wheel, has attracted a ton of attention, since the longstanding story has been that two teams inside Apple competed to make the iPhone — one to turn the iPod into a phone, and the other to shrink OS X to work on a mobile device.
Not so, says Tony Fadell, who led the iPod team at Apple and then the iPhone team. He called me from the Detroit Auto Show, where he’s presenting on panels related to autonomous cars, to clarify what we’re seeing here.
This interview has been edited for clarity and length.
So there’s this video floating around, and a bunch of stories about the early days of the iPhone project, and the two competing approaches to the phone at the beginning.
There was tons and tons of different UI development, between both software and hardware development. It was a competing set of ideas, not teams, and we were all working on it.
“There were probably 16, 17 different concepts.”
Sometimes stupid things only seem stupid at first, but if you break through, it actually becomes smart. Like, "Oh, everyone would want to type on a hardware keyboard, no one wants to type on glass.” Those kinds of things. You had to keep pushing people so that their first knee-jerk reaction isn't the answer. There were probably 16, 17 different concepts.
So tell me, this thing that we're seeing right now looks completely bonkers. There's a touchscreen clickwheel, there's an iPod interface... what is that thing?
So there were two different types of prototypes. There's one, a prototype for the UI team, and typically, because UI teams are using Director — back in the day — and quickly mocking things up on a screen. One team is doing it like it's an iPod, and another team is doing it like it was a touchscreen. The teams were working together. So it wasn't like there were two different people trying different things. And then there was the development board prototypes where we’d rewrite the UI on the hardware to try things like touchscreen and hardware buttons. So there were two tracks in hardware and software UI development running at all times. And so the thing that you're seeing [in that video] was just what the UI guys were doing, devoid of any hardware, doing it on a Mac.
That thing's running on hardware in the latest videos. They're not running Director on that prototype.
Somebody ported something to actual iPhones. So, somewhere along the way, somebody decided to port it just for fun. There were no hardware demos. It first started on the Mac. And then we did the clickwheel, because we had a clickwheel iPod, so we were trying to build on the clickwheel iPod. But that’s too hard to edit the UI on, so you'd do it on the computer and move it over to the iPod.
That explains why there's that Aqua interface element — what we're looking at here is a Mac app for iPod emulation.
It's a Mac app, exactly. It could have been Director, it could have been something else. But yeah.
So how did it end up on an iPhone prototype?
Somebody ported it later.
Someone in Apple, or somewhere else?
I have no idea. Someone, somewhere ported that thing, and that's why it's shown side by side. It was never, we never had anything like that at Apple when we were making decisions.
So clearly there had been work done. The top half of that prototype we're seeing is a bunch of iPod menu trees for a phone. The bottom half is some weird emulation thing. So you clearly got pretty far down the line of "Let's make the iPod a phone." How did the decision to go the other way come about?
Let's back up a little bit. The first thing was, we wanted to make an iPod Video product work better. So let's put a big screen on an iPod, remove the wheel, and make the wheel virtual, so you can look at widescreen videos and pictures. Because the clickwheel was getting in the way, and we wanted to not make the device bigger, but we wanted to just add a bigger screen, we wanted to try to figure out. So that's the virtual wheel, that was another offshoot that we tried.
"Let's look at a multitouch thing."
Then there was the iPod phone, which would keep the screen small — a lot like that Nokia small screen design. Then we'd just use the wheel as the interface because that's so iconic. Let’s not lose the wheel — all the marketing people's minds will blow up. So let's try that one. But what really failed at the end of the day with the iPod Phone was that you couldn't dial a number. Like, 1, 2, 3, like a rotary phone. Everything else was working but the one main thing that didn't work was dialing a regular number — it was so cumbersome. So we said, "This isn't working either."
And then there was a third thing, and this was what led us down the road to the virtual clickwheel, of "let's look at a multitouch thing." And then let's change the interface entirely, to where it came today, which was the tiles. So we started with the wheel and then we went to the tiles, both virtual and physical.
So when you're in the room with Steve and whoever else, what are you advocating for, and what are they advocating for?
A lot of us saw very early on that you couldn't make the clickwheel work for dialing a phone really quickly. So we’d say, “Steve, we don't want to spend time on this.” And Steve would go, "No, I want you to do it. Go and do it. You have to try it. Try."
We tried everything. We tried having little buttons on the clickwheel so you could click. There was a Nokia phone where they had a circular pattern for the numbers, in hard buttons, and Steve was like "Go make that work." So we tried that.
And we went, "Steve, give it up, it's going to be too hard. It's not going to work." So we were halfway through, like four weeks or five weeks into it, and we said "This is not working." We pushed this for like another four, five weeks to keep trying, and we're saying, "This is a waste of time." But we had to be ready, because that's what he wanted. It was the same thing with the hardware keyboard. Steve said, "No, we're going to make this keyboard work, on the glass."
So tell me what ended up happening with the iPod OS version of the iPhone.
Before I took over the division, Jon [Rubinstein] was running the division. There was a Linux-based OS that Jon and a couple of other people had started, and then there was the reduced OS X that Avie [Tevanian] had started with Scott [Forstall]. They were competing to see which one was better. He and Jon had that battle. I took over and we had a few conditions to make sure it could work on the hardware, since macOS was so big at the time. And then I killed the embedded Linux project, because I knew that was the much better way to go. And Steve was happy and all that stuff. It was a couple of weeks from the time I took over the division to the time I was convinced that we would be able to move forward with the Purple OS.
Purple was the codename of what we now know as iOS? It was never the code name of the other thing?
Sure. In terms of choosing the hardware, how did you land on the ARM processor?
It came from the iPod. We took all the stuff from the iPod onto that. Ultimately, later after the iPhone was successful, then Steve was trying to do the switch to Intel, and that's where the Isaacson book and that whole scene about Intel vs. ARM [came from]. But it was always ARM from the beginning, it was not even a decision, because we were building up from the iPod.
“It was a brutal slog to execute it all and ship it, but the design was so good.”
What was the moment that you knew that you had the winner product?
Well, I don't know if we had the winner product, I think we had the winner design. There's a distinction. It became immediately clear when we had enough of the touch working on the real hardware and the software working on the real hardware, compared to the other one, it was so galvanizing, it was so immediate, when we passed enough of those hurdles that we knew we had the winning design. It was a brutal slog to execute it all and ship it, but the design was so good. The design and the whole idea and vision of what we were trying to do, that no one had any questions. It was just, let's burn down all the rest to get this thing to ship.