Remember the Browser Wars of the Aughts? Internet Exploder gripped the web. Firefox 1.0 challenged the king. Six years later, IE is waning. (But still strong.) WebKit rules smartphones. Where does Mozilla, and the web, go from here?
Christopher Blizzard is the Director of Developer Relations at and Open Source Evangelist at Mozilla. We sunk into the plush furniture at the Ace Hotel to talk about the future of Mozilla and the web over Stumptown coffee—decaf for Christopher.
Giz: So you've got an iPhone app coming up, right?
Christopher: It isn't just about the browser as a piece of technology, but it's also sort of who you are, where you go, what you know, and trying to make everything really easy on mobile devices. 'Cause it's still a huge pain in the butt to type on devices and get something. So connecting data, connecting who you are—which your browser knows you better than anyone really. It's kinda creepy, actually. And bringing those cross-mobile and desktop. So I use that technology [Firefox Sync] on multiple operating systems and a machine at home and for me it's been revolutionary. Like everything syncs. I start up a new browser and it knows exactly who I am.
Giz: So I know on the desktop side, the big thing has kinda been a Javascript speed race. Where do you see Firefox competing on the desktop at this point? Chrome and Safari now have extensions, which was a really big thing that Firefox had first. So what's sort of the next step for Firefox? If you look at growth rates, Chrome is sort of speeding up faster than anybody in part because Google's pushing it kinda hard. Like, if you go to YouTube on a non-Chrome browser it says, 'Hey, try Chrome'.
Christopher: Yeah, so, there's a bunch of stuff. Javascript and the C front, that battle's not finished. They have what we call—well, Chrome-inspired—a baseline JIT [compiler] in their browsers. And it basically, when you download the code, it compiles it and it runs it into the native code. It's pretty fast and it's consistent.
In Firefox, since 3.5, we showed this thing called tracing. So tracing takes code as it's running and optimizes it. It's a very different approach.
We're going to do a baseline JIT for Firefox 4. It's not done yet, and it hasn't landed in any tree yet, so nobody's tested it. It's gonna give similar performance characteristics to Chrome. But we're also gonna do tracing on top of that. What we discovered, is that for a lot of applications, especially when you want to do anything that's actually CPU intensive, we smoke everybody. It's not even fair.
What we're gonna have, I think, with Firefox 4 is that we're gonna have a Javascript engine that's a generation ahead of everybody else, which will be pretty interesting. We're kind of excited about that. Our larger mission is drive apps on the web...so I think that's going to be really important to see.
You see Google pushing native client now—it's part of a solution for building performance on the web. And what we've seen is actually with tracing is that we can get a lot of those performance benefits or get really close. And we know we're only gonna get faster to what you can do with native client or you can even do with native code. So, this is gonna go on for a couple more years.
Giz: It's interesting that you have the three biggest players in tech with each one pushing their own browser: Microsoft has theirs; Apple has theirs; now Google's pushing theirs; and Mozilla's like the true independent at this point.
Christopher: We're the only one that makes browsers! We don't sell ads, we don't make operating systems. We don't sell hardware. We just make browsers.
Giz: Before, when it was Mozilla vs. Microsoft, it's easier for people to get behind that in a way. People don't feel so bad about using Google's browser. They're not like, ‘Fuck Google.' That was kind of like a big part of [Mozilla vs. Microsoft]. So where does Mozilla go from there, in this kind of environment? How do you compete?
Christopher: We ask that question a lot internally too, and it always comes back to the fact that we care about people. It actually is the most important thing. And the web is a democratizing force. So as long as we continue to make things that people love, people are gonna continue to use it. As long as we stand behind the technology that we have, and continue to improve and invest in it as a mechanism for getting the larger goals done, we're gonna be fine. I'm not that worried about competing with them.
We don't have as many resources in terms of cash, but in terms of team coherency, we're much better in terms of team size. We're about the same as everybody else. Sometimes larger, sometimes smaller.
We have a lot of assets too. We have strong brand identity. It's really interesting that people continue to choose Firefox over and over and over again. We've seen a lot of pickup from Chrome, but we know we're still adding users at twice their rate. We know that people love us. When we go out, they go ‘oh I love Firefox, it's changed my experience' and that is huge. In a consumer space, not even in tech space, so we have those assets. But I also think that if we just stay focused on what we care about, we'll be fine. We continue to reinvent ourselves and we continue to build things people like.
Giz: How do you continue to balance keeping things lean and adding features? Obviously, sync is a big new feature. It's one makes a lot of sense when you're talking about a cohesive browsing experience.
Christopher: We spend a lot of time on that. I would say that reinventing yourself every once in a while is a pretty good idea. And we continue to do that. Most people don't know this, but extensions were originally created so that we could create that lean experience. It gave us the ability to say no to a feature and say, we don't want them in a core browser experience, it's gonna have to be an extension. And that I think is one of the reasons why we see thousands and thousands of extensions.
You know Chrome has extensions, and Safari will have extensions, but it is not gonna be the same. The platform that they've built their extensions on, and most people don't realize this, but the difference is a very constrained API. Firefox is a full-on platform. The UI is written in the same language as the web, you can modify it, you can change it. It's a completely flexible platform, like, in a hacking way. And that gives a very large creative outlet for anybody to try anything. So it doesn't help with our stability, but it helps with the kind of creativity people want to be able to do.
You can't do that in other browsers. They've taken all the language we've used to describe extensions and applied it, clearly, but they're not the same. The ability to be creative, the ability to change, and the ability to discover what people really want is something extensions have given us and will allow us to continue to grow.
We're bringing that to mobile. We care a lot about data. We're moving from beyond just browsers into identity and sync. So we're doing a lot of really interesting stuff.
Giz: Opera's interesting in that they've pioneered a lot of features, like the speed-dial thing that Safari and Chrome use. Do you guys worry about winding up where they are? They have so many features in their browsers and they're kind of like a weird little experimenting ground for other people, 'Oh that works pretty well, we're going to take that.' But even on the phone, they seem to be slipping.
Christopher: They're in a weird place. I don't know what they're going to do. Their mobile products have been fantastic, especially their feature phones, because of the server side piece. As far as I know, nobody's taken that space.
Giz: Skyfire does server-side compression as well, on a handful of smartphones....
Christopher: There's an interesting story there that most people won't realize. The reason they chose Mozilla is compatibility. As soon as you step outside the tech web sites, stuff doesn't work in WebKit. The founder is Indian, and as soon as you go into India, there are a lot of sites that still don't work in Firefox, but they are more likely to work in Firefox than they are in other browsers. So, when you get to the edges of web, Firefox's compatibility is still really great, but not so much with the WebKit-based browsers. That's what we've heard.
So, I don't think we're gonna end up in the same place as Opera, because our architecture's different; our community's different. We have much more marketshare than they ever did, so I'm not that worried about turning into Opera. I also say that we're a wildly different organization than they are. I mean, they're huge. They're a public company, right? So you can look at their statements. One of the guys in the office said that they're—you might wanna check—750 people or something. Like, they're pretty big. They spend all their time porting to phones. Their desktop browser has been a testbed for their mobile platforms, so I'm not that worried about ending up in the same place as them.
Giz: Do you guys hope to be on most phones, or some phones? Is Android good enough?
Christopher: We need to pick. We want to bring a full browsing experience, which means that we're basically tied to smarter phones at this point. We need to be on open platforms to allow us to innovate how and where we need to. The iPhone is terrible for that. When I talk about all the creativity we can do around extensions, there's no way that stuff would be allowed on an iPhone. While other people innovate, it's pretty scary, right?
So we're doing Android. We were going to do Windows Mobile because Windows Mobile really needed a good browser, but they shut down their platform.
[On Windows Phone 7] they said, 'You can use stuff in Silverlight if you want, but you have to come through our app store, and we get to veto.' We're not going to bother. They're gonna ship some version of IE, which is gonna be terrible...
Giz: It's like a mutant version, based on IE7...
Christopher: The wonderful thing about the mobile web is the people like being able to use modern browser platforms and Microsoft's gonna single-handedly gonna drag that back.
Giz: So where do you see web apps in the future? Cause Google's very big on the web app thing on this point. Android seems like a stopgap for them in some ways. That seems to be where things are going. Do you guys feel pretty much the same way, like it's gonna be a web app world?
Christopher: I come from a slightly different standpoint. I mean, there's a bunch of new issues. There's monetization; there's agility; there's engines. What's interesting is that browsers have gotten now to the point where it's possible to build really useful, intensive applications in and of themselves, that are useful in and of themselves, not just viewing data or making it convenient to get to.
I think that's the big sell with the Chrome Store with what they're trying to do. Part of it is monetization, but part of it is basically making sure their content is easy to get to and easy to navigate to and easy to find. So, they're gonna do a bunch of stuff for that.
I actually think that we need to have a bigger picture view than just that. I think that how you actually build applications needs to change in a different way with the web so that something can cross desktop and mobile, and I think we're in a unique position to be able to do that. Some of that's gonna be standards based; some of it's gonna be experimentation, and some of it's gonna be services. We're doing this thing as a service for doing identity as part of that. Bring your identity with you.
In terms of the big picture for applications, it's something we're gonna be involved with, and something we're helping drive. So, we're gonna be there. I also think that it is time to think about how to monetize and it's time to think about how other people monetize to allow them to build great apps. I mean I've heard from like the people on the iPad they basically say 'Oh we take this content from the web and we wrap it up and we call it an eBook and we sell it for 10 bucks on the iPad and people will buy it.'
Giz: What do you think of VP8? Do you see the HTML5 video thing being resolved anytime soon? [Be sure to check out Christopher's huge rip on Apple's HTML5 marketing.]
Christopher: There's a bunch of issues there. The first is as an implementation and as a standard, VP8 is pretty good. The standard piece of it is basically a lot of stuff hasn't been written down, a lot of of stuff is like, ‘Well, look at the code, this is how it works'. So we need to get to the point where we have a spec. It is just a spec. I think adoption is gonna drive a lot of that, so VP8 is interesting because it is gonna be available on YouTube. That's 20 percent right there, of all web content. They're an order magnitude larger than anyone else put together. So that means it's gonna be relatively accessible. I would also say that on the what's interesting on the encoding side is that VPA isn't just useful as something that you can watch a video on, but it's something you're gonna be able to use to encode videos as well, so you'll be able to use it for realtime chat. You're gonna be able to use it royalty free for all of these use cases that H.264 doesn't look bad, but it's not great. It's extremely CPU intensive to encode and extremely complicated.
H.264 gets the compression it does because they trade off a huge amount of CPU for encoding and decoding. If you're decoding h.264 video, depending on the profile you're using, it can be extremely CPU intensive. Some graphics cards are starting to support the actual decoding. Although some graphics cards it turns out are not that good at decoding. They're not good at decoding video codecs. One of the guys who does video codecs explained this to me, because of the way that video codecs work, you have to have a lot of knowledge about data in multiple directions, so if you're encoding and decoding you have to look a lot of places. Video cards are good at a piece of data and a computation all at the same time, and you can't do that with video. You need dedicated hardware if you're gonna do that.
H.264 has these three profiles, right? Baseline, main and high. I think those are the names. The interesting thing is that basically everybody sets their videos at baseline, essentially cause of the iPod. The iPod only decodes baseline. You can't use it to get higher profiles, so you sort of end up with a fair-to-mid level format. If you decode high, it is a tremendously intensive process to both encode and decode, so you need a lot of CPU.
So the interesting thing about VP8 is it always has the same profile. It's relatively computationally cheap to encode and decode, it comes with an immediate code encoder, so you can basically do real-time stuff (though there are other tradeoffs there). There's a lot of space left here for innovation and there's a lot of space here left to try new things in new markets. So it's not just gonna be about the web, it's gonna be about these other markets as well. But on the web, we're not gonna see enough adoption of HTML5 browsers for another couple years, so this is gonna be a game that's based on patience.
Thanks to Christopher Blizzard from Mozilla for talking to us! Check out the Firefox 4 Beta here, and Fennec for Android here.
Via: (Gizmodo)
No comments:
Post a Comment