Designing Windows 8 or: How to Redesign a Religion
There are lot of hard jobs at Microsoft. Like the guy who collects Steve Ballmer's dry cleaning. But Sam Moreau just might have the hardest gig in Redmond. Or at least the most harrowing. Over the past five years, he's taken on the tiny task of redesigning the operating system used by like a billion people all over the world. You know, NBD.
Sam's the Director of User Experience for Windows, Windows Live and Internet Explorer, and along with Julie Larson-Green, the Vice President of Windows, they've been re-imagining Windows from the ground up with Windows 8, in ways that might surprise you.
I've broken this gigantic interview—which I've edited and condensed—into three parts, roughly descending in order of importance: The big ideas behind Windows 8 and what it's like to design it, Metro across Microsoft, and the nuts and bolts of the Windows 8 redesign.
Make no mistake, kids, you're gonna want to hold onto your pants when Windows 8 drops into your lap. And you'll even be able to use it while you do.
Giz: Metro is going to be the way hundreds of millions of people use their computers, their phones, their Xbox. What does it mean to bring good design to that many people? You have the status quo, and now you're bringing people something that is not radically different, but in some ways it is.
It's one of the coolest design challenges I can ever think of. And definitely the coolest one I've had in my entire career is to design something that is not necessarily broken. It has this whole past and greatness about it. What Windows does is pretty remarkable. It runs a lot of the world. To design something that is really not broken and works really well, and also to design something for a future that's kind of unknown—we don't know a bunch of things that are going on, like that [Intel convertible tablet stuff]. When we started designing this, we didn't know about that. We weren't imagining the hardware this is going to go on. You never know. It's one thing to design with known parameters and to fix whatever fits in the box, but we had this big open-ended thing, to design it for the future. And we play a role in deciding what the future is.
It's super exciting to do that, to design something that you know, by definition, gets to affect the future of computing. Being at Microsoft, being at Windows, that's the seat you get to sit in. Really scary, but really awesome. It's no problem getting up every morning and going to work because you can think, "Oh, I can change the Start menu today." Then I drive home with an ulcer because I can change the Start menu today. And that's how computers work for the whole world. That's harrowing.
When we designed this, none of this stuff existed, so it's pretty hard. We were pretending, carrying pieces of cardboard around, sort of a 16:9 cut out. We noticed when we were doing that with these pieces of cardboard, and we had different sizes of them, that my thumbs are really important. And that led us to, "What if we base the whole interaction model on thumbs? What if the whole thing is just based on thumbs?" Nobody else does that. Notice where my thumb is and where Start is, when I just naturally hold the machine. But we didn't have a machine, it was like cardboard and paper.
[No joke, Moreau starts flying through Windows 8 with just his thumbs right here.]
Giz: Microsoft's been doing concepts for a long time, these cool wiz-bang UIs, and with Metro, it's like, "Oh hey, these are are in real products now." What finally allowed that to happen?
SM: As long as I've been at Microsoft, they've been sort of tired of being an afterthought in the design world. When I went to Microsoft, a lot of my friends in the design community were like, "Seriously?" And my answer was, "Of course. Where else can you create religion?"
So, opportunities are huge, but a company like Microsoft does not want to have low self-esteem about themselves, and design was a place where they didn't necessarily have a high esteem. The design leaders went back to the company and were like, "Let's change this." We wanted to be the place where designers wanted to go.
And also, there were conditions that were created that made it sort of work. There was a strong competitor and there was strong leadership. Those are the conditions that allowed us to do what we do, and we started to have the makings of something that was an idea that was true to us. It authentically came from us; we weren't copying anybody, and we were developing something that we were proud of. We have something. It's beautiful, it works, and it came from us.
Giz: Right, when I talk to people about Windows Phone, some people don't like the cutoff text or whatever. But it has a sense of vision—a really strong sense of vision—and I think that's a good thing, one way or another.
SM: Yeah, because that was the thing that we get blamed for a lot. Or, I would say the thing that annoys me about Android is I don't think it has a point of view. I think it's trying to be this weird sci-fi version of an Apple design language, poorly executed. Some of it is starting to get there, you know, they got Matias there, and some things are starting to get a little better. They're starting to get that—some sense of soul, but I do think that a lot of their soul is derived from some other place.
Metro, Metro, Metro
Giz: When you're talking about Metro, how do you take that from a phone, and put it on Xbox and Windows 8 and apply it in ways that make sense?
SM: It didn't start with the phone, it actually goes older than that. It's something we've been working on for a long time—like at least as long as I've been at the company, which is over five years. One of the things that we did—we don't talk about this as much—but we have these internal design virtual teams that go across the company, and do these things called like, Design Align. We would send groups—a designer from each of the different business divisions—to a project around common commanding or something like that, and they would come off and work on some solutions, and then we would try to bring it back into product cycles.
At the same time, there's this group called the User Experience Leadership team, which is like the heads of state from each of the divisions. It's myself, from Windows, a couple of people like Albert Shung from Windows Phone, and some people from Office and Xbox, etc. We meet regularly to talk about commonizing some things. So for example, when we got to the point of Phone, Albert gave me his motion designers to commonize some of our transitions, and things like that. I gave them back some of our designers that were working on the grid to make sure we get these grid segmentations right, and then we loan it to Xbox to think about how the commanding works in a console space, which is different than how it works in an interaction space.
The main thing that we do is in terms of alignment—the place where we're most aligned—is in principles. We all share the five Metro principles that we talk about, and then we give each other a little bit of room to be specific to the console so that if you're doing Kinect it's different than what you do with touch, so there's some very important places that we don't map one to one.
Giz: What are you most proud of that you've done in terms of interface and design in this build of Windows 8?
SM: We designed a future model that didn't have to leave the past at the same time. If it was like a blank piece of paper, "go design something," that's an easy challenge. But if somebody gives you, "Here's 25 years of code and built-in patterns and interaction models and everything like that now, and here's this whole stack of future needs and desires and wants for interaction or whatever," and then they said, "Now design the whole thing," that's one of the hardest design challenges there is, right? To pass down the future at the same time and make it cohesive. And I think we did. I think we've made it elegantly, gracefully using the entirety of the PC's legacy and potential at the same time in this design.
I remember I said to Steven [Sinofsky] at one point, "You know it's really hard to do that. It's hard to have both of everything and have an elegant solution." His response was, "Well, what if you just design it? Then you just have to make it good." That's sort of a Zen answer.
Giz: How do you think about fonts here, versus the phone or Xbox?
SM: The font we use across the whole company is Segoe. We tune it, though, with little tiny nuances and then we pass them back and forth across the groups. The television has different display characteristics—they need to do some slightly different things and the weights don't hold up as well and resolutions are different—but we just have this common font that we keep passing around, adding to it and augmenting it, trying to use it in a familial way that isn't exactly identical because there's a slightly different personality between a 14-year-old gamer on an Xbox and a 40-year-old executive at Exxon.
It's a font that Microsoft developed. It has roots similar to the Frutiger family of fonts, which are very utilitarian, Germanish type font—clean, very straightforward—but rendered to work well on screens as opposed to paper. Its personality is meant to be relatively benign. It tries to be straightforward and honest, so that it can hold up really well in a messy context. It's like the classic blue suit. It works at funerals and parties.
Giz: What has the negative feedback been like? What do people not like?
SM: To be honest, you know what I think it is? When you change something—this is my own personal observation—a lot of us know how the PC works, become the help desk for all of our friends and family. Inherent in that is a sense that I know. I've got this expertise now, I've got this power. We've changed something now, and leveled the playing field for all those personal help desks, so they're no longer the guy. It's human nature—I had invested in this, I knew this, and some degree of my self was aligned to the fact that I know how this stuff works. I do think that's an aspect of what's going on.
Giz: Do your friends come to you and tell you, this is what I want Windows to look like? What's the craziest request?
SM: Last night at dinner. A friend of mine said, "Look it's all great and everything. But I need you to make the fonts a little bit bigger. My eyes are getting older, so just make them a little bit bigger for me please."
Giz: But the fonts are huge.
SM: You're not the only one who uses it, it turns out. That type of stuff. I'll get that from uncle who'll call me. "It's pretty hard to read. Could you make the button a little bigger? " All the time.
JLG: Steven does this talk, mostly about Office, about how it's like ordering pizza for a billion people. Some people are lactose intolerant. Some people don't like mushrooms. Now make everyone happy.
SM: There's a billion people, and pizza's your only option. That's what it's like designing Windows.
The Nuts and Bolts
Giz: How prioritized is UI in Windows 8, on a technical level?
JLG: We've reduced down the number of processes that are running—reworking how things are layered, so not as many processes are running. Then there's the culture thing of the team around fit and finish—our principle of fast and fluid, reducing steps, make it easier. The whole team has to do that work, even though it's the design team that has to push it and keep us honest on it. That's why we always say it's Windows reimagined, from the base kernel all the way up to the UI and all the way through.
Like when you reboot the machine and it starts up, that was one of those cases of really thinking that through end to end. Remember, you used to see the text screens that would come up or if your battery died when the lid was shut, and when you open it up it'd give you all these old Courier text screens? To get rid of all that stuff and just make it on/off, was 5 teams' work to relayer and rethink all that.
Giz: So this zoom out thing in Windows 8, showing all your apps and stuff, why is it called semantic zoom?
SM: With "optical" zoom basically you'd just make it further away or closer up, but the pixels are all the same. That's not that useful. So the semantic idea was to put it at a different semantic data layer so it's actually more useful. Imagine you have a large collection of stuff. It's a store, or something like that. Or it's a Netflix catalogue. Zooming it out just makes all the movies smaller, like they're smaller tiles. Instead, semantic zoom would make the categories or the metadata come to the foreground as opposed to just the thumbnails of the pictures.
Giz: What else do you have going on in terms of new gestures?
SM: One of the biggest things for us is the selection gesture, the idea that you don't press, hold and wait to select something, but that you can be fast and fluid and just select it. I'm sort of unlocking it from its place, and as I do that, I see that behind it is a self-revealing gesture—there's a little check mark—so when I let go, it selects it. Now you can move it around. I can do a bunch of other stuff too, because when I did that, it invokes the command bar that says, "Oh maybe you want it smaller, maybe you want to turn off the live tile, uninstall it, unpin it, whatever."
Giz: Windows 8 doesn't really seemed to be designed for portrait—this is gonna be a landscape experience?
SM: It's designed to be predominantly in landscape, but also transition gracefully into portrait with the situation, if the app decides. Apps can make themselves portrait, and the system has no problem with that.
Giz: Do you see someone using this interface at all with keyboard and mouse or do you think this is, most people are gonna be using touch?
JLG: If you have a laptop, just a regular old Windows 7 laptop, it all works, mouse and keyboard. We've also created new ways for mouse and keyboard that make it quicker for them.
Giz: So what's the deal with the task manager though?
JLG: You don't have to close stuff—in fact we don't think you need to close anything—but people seem to want to close. It is definitely one where we felt we had to add it, because people want it, not because they need it. None of those things are running. None of them are running. And you could go into task manager and see that nothing is running. You will be able to go in and see that they are all suspended and not taking any battery or doing anything. In RAM. That's all. They're taking up a little bit of RAM, but they're not degrading the system or your battery power, and that's what people think. It's a little more touchable this time. The targets are bigger. The task manager, you shouldn't ever have to use it.
Giz: How much design work are you putting into the desktop? It fundamentally looks like Windows 7 there's a few tweaks, like the ribbon..
JLG: Visually we haven't done a ton, but we've added a lot of features. We have the reset feature, we have all these complicated storage spaces where you can set up your different drives and undock them and dock them around and all these you know more power-user type features, and we did the command center for the file system with the ribbon.
Giz: Can I make a critique? The connection icon for Internet. The computer with the cable. It looks, I don't know, it's a weird looking icon. It's hard to tell what that is. It looks like a pitchfork. Or a petunia.
SM: Poseidon's PC, that's what it is. It's funny actually because we've changed that icon twice now in the last two releases and I'm still not happy with it. Sometimes it's a little lower or a little higher. And I'm one of the only ones who actually knows that it's been moved like two or three pixels.
Giz: And I guess blue is the default color, because that's Microsoft's…
SM: Each release I pick a different color, there's some little story behind it. The first one was purple, at the D Conference. It was complimenting what [Larson-Green] was wearing that day. She was wearing purple.
JLG: You think he's kidding, but he's not kidding.
SM: So yeah, each of the different sort of milestones—the D Conference, which was the preview, and then the developer conference that we made green, for cash. So that the developers knew that this was about making money.
Giz: I find this whole Share thing really interesting too. It's one of the better quote ‘post-PC' things I think that you've brought into Windows 8—how do you get this thing to where I want it to go?
SM: These functions, we thought, should just be native to computing. Searching, sharing that should be just native to how the computer works. So any apps that have registered for the Share contract,—basically you build an app, then you say, "I'm gonna use this API, which means I'm doing the Share contract," Then they get to show up in this list whenever app A has a Share moment. And so FriendSend or Mail have said, "We're registered for the Share contract," so they get to show up in here. So sending that article was that fast, I never left contacts, I never had to do anything.
And it also means that if you make an app, you're now interoperable with every other app there is without having to write any code specific to whatever. You don't know what app is coming in the future, but you're now interoperable with them no matter what. Netflix could decide that they want you to do the star rating when you share it…we try to work to work so let the apps be the best of the apps, put their best foot forward, and at the same time give the user enough of a hand holding that they're not just in the wild west of experiences, like they're not just left on their own.
JLG: And then as you get apps your system becomes more powerful, and you can share more things, you can search through different content, you can do different things. Another way to think of it is like a crazy new version of clipboard.
SM: A two-way modern clipboard...for all apps.
Giz: Search looks a lot different here too.
SM: Imagine a different world where I had typed "hangover." The results for hangover are very different in Internet Explorer on Bing versus Netflix. Let's say I've got a bunch of apps: Netflix, Hulu, Wikipedia, whatever. Each time I click one of those they get to render what they have, their best foot forward for "hangover." And it may be very different things. Google gives you this homogenized version of the internet for a query. Apps, in this one box that drives the query across all the other apps, lets each app give you the best version of what they can do for "hangover." Maybe you can get WebMD and hangover cures.
Giz: When you do get errors now is it going to tell you what happened in English?
JLG. Yes and no. For Metro-style applications, it's a lot different. For older Windows desktop apps, you'll still get some of those. We tried to clean up as many as we can, but there's a lot of error conditions, and we don't know how they all get fired after 25 years. But the newer ones look completely different.
Giz: Are you guys worried at all about performance with the Metro-style HTML5 apps?
JLG: First, you can write them in C++ or whatever you want to do, HTML5. The HTML5 rendering engine is built on the primitives of Windows so you're using the the GPU, it's graphic accelerated, it's calling through—it's not just raw HTML rendered through the browser. It's Windows rendering it. IE9 was the start of us doing that, and that same thing is powering the apps. It's a huge part [of the changes under the hood]. You can't duplicate that on any other operating system.
Giz: Any difference in UI or animations with ARM-based Windows 8?
SM: We had to work on animations and different framerates based on the behaviors of ARM We're working to make them the same. Our goal is always to not have any 'forks' in anything because it's just more efficient. Where we're at now, it's still a goal—we still haven't finished deciding whether that works or not because we have to get a bunch of ARM stuff in and test it.
Giz: Has adding ARM changed a lot of the development process in general?
JLG: It's generally the same code base, it just gets compiled differently. There's one code base. It's not a radical change in the process, but the radical part of the change is that there aren't a lot of machines already available up and running that we can start running stuff on. So we have a lot of work with the silicon vendors to design the chips to work well.