Google Chrome: The interface is beside the point

There's been a lot chatter around the office and internet about Chrome, the recently launched (or leaked) Google Web browser. I've got to say that much of it misses the aspect of the application that I find most inspirational. Google Chrome exists for one reason and one reason only: To provide a framework for web-based applications to look, feel, and act like desktop applications.

It doesn't seem that Google has real interest in replacing IE or Firefox as the dominant web browser. (The tech business press can't see beyond this point). Instead, Google wants to see the technological underpinnings of Chrome adopted by mainstream browsers. The Chrome team is explicitly inviting other browsers to use their code base — they've open sourced everything — and they have explicitly acknowledged adopting best-of-breed UI features from others.

So how does Chrome elevate web apps to desktop app status? Six ways: Separate processes for each tab; Google Gears for local storage, offline functionality and "native app" behaviors; application shortcuts; a modern JavaScript virtual machine; and minimal-to-absent browser interface, aka "chrome". Let's look at each one in more detail, with snippets from Scott McCloud's fantastic graphic novella product tour. chrome-process.jpgChrome provides a separate processes for each window and tab, just like the OS provides a separate process for each application. As a result, hitting a page that slows down or crashes your browser doesn't take down your mail, spreadsheet, and photo editor.

Gears lets you work offline. You can edit your presentation in the plane on the way to the client or read your mail at the cafe with spotty WiFi, and developers can create web apps that function more like desktop apps. chome-gears2.jpg

chrome-shortcuts.jpg Application shortcuts tear web apps from the browser and place them along side your other desktop applications. Want to work on a document? Fire up Google Docs. Check your schedule? Launch your MobileMe calendar. Each app is represented in your taskbar or dock just like any other.

chrome-js.jpg AJAX web applications have elevated javascript into a fully functional programming language, but browsers today just aren't optimized to run it. The V8 JavaScript virtual machine is dramatically more proficient, robust, and stable.

Chrome? What chrome? Web applications don't need back buttons, or favorites, or search boxes, or bookmark bars, or Google Toolbars, or... They are self-contained, and any browser interface is a distraction. As a browser, Chrome strips away everything it can. When apps run from applications shortcuts, there is no browser interface at all. chrome-nochrome.jpg

Google is practically screaming "Take these technologies and run." They are doing a great job of promoting the availability and advantages of their approach. In addition to the various overviews at the main Chrome site there's extensive documentation at, the open source project behind Chrome. There, the introduction to the User Experience section sums up everything I've been talking about:

In the long term, we think of Chromium as a tabbed window manager or shell for the web rather than a browser application. We avoid putting things into our UI in the same way you would hope that Apple and Microsoft would avoid putting things into the standard window frames of applications on their operating systems ... The tab is our equivalent of a desktop application's title bar; the frame containing the tabs is a convenient mechanism for managing groups of those applications. In future, there may be other tab types that do not host the normal browser toolbar.

With that, Google welcomes us all to the age of cloud computing, where your applications and data are services seamlessly available locally and over the network. For Chrome, the browser just isn't the point.


Michael Jackson
What struck me as I watched the webcast introduction to Chrome was that these are some major geek guys here and the whole basis for the interface design was “Well I want it to work this way... When I use a browser I don’t like….” It is clear the inmates are running that asylum. Being a software engineer for > 20 yrs now I have a great deal of experience where that line of thinking has taken me. I can just see it now trying to explain to my neighbor that you can enter searches and or addresses in the “omnibox”
Tim McCoy
Yeah, we've had a thread internally about the engineer-driven design process. I think what makes it different from an inmates situation is that (no offense please) these are *really bright* software developers who are extremely talented across a range of disciplines. I mean, Alan Cooper is an engineer. And there is a bunch of self-referential design in there, but it seems well-informed and well-intentioned. The beauty of the omnibox is that you don't have to explain it to your neighbor, for two reasons. First, he's not the target user for Chrome, so he probably won't encounter it here (though the idea is already on its way into more mainstream browsers). Second, there's nothing to explain--got a URL? type it in! Looking for something? type it in! Start typing the address for your local paper for the 80th time? Hey, it filled it in for you! That rounds back to the first point--the self-referential design succeeds partly because the engineers on the team are good proxies for their target users. By solving for "how would I like it to behave" they are also addressing how the early-adopter crowd would like it to behave. And as Chrome isn't setting out to become the next generation web-browser but to help foster its development, that's OK.
Doug LeMoine
What Tim says reminds me of an interesting debate took place on the 37signals blog a couple of years ago. Someone asked them for their thoughts on personas, and Jason Fried responded by saying: "We don’t use personas. We use ourselves. I believe personas lead to a false sense of understanding at the deepest, most critical levels." And of course their products are made for people like themselves, so why bother with the abstraction of a persona? (I ask that rhetorically, wondering if anyone has thoughts in this vein). Also, there's some interesting chatter on this IxDA thread about personas and self-referential design.
Dave Cronin
@Tim, while I agree that the engineers who worked on this are probably smarter than the average bear, let's not go overboard here. I'm not sure they wouldn't have benefited from just a bit more attention to interaction and visual interface design. For one thing, that color blue isn't exactly consistent with "if you can just ignore the browser, we've done a good job." (Just because MSFT did it with XP, doesn't mean it's ok.) But more to the point, actual interaction design might have resulted in slightly more UI innovation than just moving the tabs over the URL bar to a more logically expressive position. (Which, I nevertheless like, both conceptually, and because if you're using a tabbed interface like yahoo mail, there's a bit of separation between browser tabs and application/site tabs.) Yes, they did a nice job keeping things simple (first, do no harm!), and to your well-made points, this product is more about innovation at the application layer than at the UI layer, but I'm not sure they've really taken advantage of their unique and privileged position to bring a new browsing experience to the world. (And for all of the promise of open source, it's rare to see real interaction innovation in an open source project. Though peeps, please pipe up with counter-examples.)
Nick Coster
Google has created for themselves a fairly unique position in the online technology market place. They make money from their Adwords network. Lots of it. This provides them with the luxury of being able to be led by their engineering teams, investing in a wide variety of product inventions and technology innovations without having to be burdened by their short term profitability. This makes them exciting to use as an innovation model, but it is not a very practical one. Most new product ideas that are developed are built with a clear intention to return a value to the business in one form or another. The engineers at Google are simply not required to do this. While this frees up their creativity and the range of experimentation it does not guarantee a commercial success. In fact any commercial success becomes a significant gamble. In my view the purpose of identifying a clear target market and the subsequent use of personas is to minimise the gamble for organisations that cannot afford to gamble their product development resources away. By understanding the customer and the problems that they would exchange “value” to have solved, the risk of product failure is significantly reduced. Note: I am a Google Chrome user. I am glad they had the time to dedicate to the project.
Skot Nelson
I wish that Google had, instead, focused resources to help the Mozilla project. The world needs stable, reliable browsing: not more browsers. Having said that, I'm happy to see the use of WebKit, which is at the heart of Safari. I basically agree with the premise of your article, and the /business/ press doesn't get it although some tech writers do. I personally suspect that we'll see an Asus EEE model dedicated to running Google Chrome fairly soon. With WiFi prevalent, and Gears designed to fill the gap, it could be an interesting machine.
Paul Neervoort
It seems to me that it is fairly irrelevant whether the inmates are desinging and building or not. I believe the core is Fit for Purpose and Natural Language in terms of what is the thing the user wants to achieve. The point I see in the (promise of) Chrome concept is that they are really thinking very hard about using the most natural way to do things and designing a UI solution which is fit for purpose. They are also trying to break the artificial barriers between applications. I say artificial since most of the time a user does not really consider the tool to be the purpose of his activity or a goal he wants to achieve. Whether that means using tabs or taskbars or other means is irrelevant as neither one is fitting the profile in all cases. If you think about the natural language of a task neither a canvas and toolbox approach as in the Adobe interfaces, or the toolbars and 'ribbons' from MS are intuitive natural interactions. For that matter over using the Omnibox would cut it either. I think the apporach of Enso is getting closer to the point. Although typing in the command is also not really intuitive. But their point is at least addressing the issue of What Does the User Really Want to Achieve. What does fit for purpose mean if you are writing a report. If the promise of Chrome is true at least some other inmates may be able to get one more step closer to that goal.

Post a comment

We’re trying to advance the conversation, and we trust that you will, too. We’d rather not moderate, but we will remove any comments that are blatantly inflammatory or inappropriate. Let it fly, but keep it clean. Thanks.

Post this comment