Firefox 3.5 vs. Chrome 3 Showdown – Running Web apps

While some are still sounding the trumpets over Google’s proclamation that its Chrome Web browser technology will be elevated to the role of operating system sometime next year, there are some Web applications engineers who wonder why it isn’t one already. Once Chrome 3 is proclaimed fully functional on Linux, it will essentially make the same Web applications accessible from the desktop that Chrome OS plans to do. And if you don’t mind the fact that the Chrome 3 Web browser is in the development stage, whether you’re running Linux or Windows XP on your netbook, it does that right now.

Though I’ve made the point before that most businesses and a majority of consumers today still prefer Microsoft Office, there will be a small, though potentially beneficial, market of consumers who appreciate the flexibility and versatility that a Web application may offer. If they can suffer through the bugs, users will have the opportunity to produce some respectable, if not altogether spectacular, documents. But what is there for users to gain by installing these Web apps on their desktops, and use them like software installed on their computers, as opposed to simply running them in their browsers?

If “the Web is the platform,” as Google likes to say, then Chrome 3 provides that platform with a front door. For Mozilla Firefox 3.5, it’s more of a side door, but a door nonetheless: The organization offers a plug-in called Prism that enables Firefox users to convert any active Web page into a browser-less application.

With the understanding that Google Apps, one of the most prominent Web application suites, might have an automatically easy time running in Chrome, and that Microsoft’s Web applications are geared around Internet Explorer (and require separate downloads), we decided to test both Chrome and Prism with Web apps produced by groups other than Mozilla and Google, and other than Microsoft. Specifically, we chose: Zoho, a complete suite that uses Google Gears but was not built with Google’s direct assistance; Buzzword, part of Adobe’s Acrobat.com online suite; and Pixlr, an online photo editing tool that uses filters and layers.

Getting the browser out of the way

One refreshing option offered by both Prism and Chrome 3, to both manufacturers’ credit, is the option to frame the Web app with nothing other than a plain window. At this most basic level, you’re given no clue as to whose platform is hosting the app — nothing on the window, no “About” box, nothing to right-click to see a trademark. The application appears to own the window without any obstruction from foreign manufacturers, and that’s as it should be. For testers, though, it merely poses the minor problem of forcing them to remember which platform they launched each app in.

It’s worth noting here that Chrome doesn’t display converted Web apps in a “Chrome window” — in other words, it’s not sensitive to the new “snap-to” events that premiered in Chrome 3, and that are part of the basis for Google’s claim to a wholly new windowing environment. So you can’t bump a Web app window alongside a Chrome window, for instance, and have it hug the side of it, as you can with two Chrome windows. This is to Chrome’s credit; in fact, Web app developers would probably prefer their hosting platform to be as neutral as possible.

The Web app “installation” process

The concept of “installing” a Web application is not self-evident, and for the market that most Web apps serve (novice and everyday computer users), it should be. It’s a bit of a problem, because Web app producers such as Adobe and Zoho will be looked to by users to provide the necessary step-by-step instructions — no one would expect Microsoft to provide instructions for installing, say, Photoshop Elements. But Adobe and Zoho cannot provide those instructions because they’re not the ones making the browser; and even if they tried by providing a “Firefox,” “Chrome,” or “Safari” page, the browsers themselves change rapidly and without notice.

Mozilla knows that folks who intentionally install the Prism add-on for Firefox will probably have the good sense to read its online instructions, and from there learn what they need to do to “install” Web apps on the desktop. But Chrome’s Web app installation ability is built-in, which means it’s up to Chrome’s online documentation not only to inform the user as to how to do this, but to alert the new user as to the feature’s existence. To Google’s credit, its online docs do a good job of this, making “Web Applications” a top category on Chrome’s Help page.

Chrome’s method of making a Web application into a stand-alone app is, like much of Google, characteristically understated — so much so that finding out the proper metaphor Google chose to use is a lot of needless guesswork. Chrome portrays the job as an “application shortcut to the desktop,” which many users won’t immediately recognize as any different from making a Web page shortcut to the desktop. If an app runs via shortcut, then one might conclude that dragging the app’s URL to the desktop — which does create a shortcut — would make the right kind of shortcut. No, this creates a URL that launches the browser, just as with any Web page in the past.

But the act itself is not difficult. First, you run the app in the regular browser. Next, click on the Page icon near the address bar, and from the menu, select Create Application Shortcuts. This brings up a dialog box marked Gears, from which you select whether shortcuts are to be placed on the desktop, the Start menu, or a Quick Launch toolbar. It would be nice here if Chrome could recognize whether or not it was running in Windows 7 (if the developers can’t Google it themselves, they might try looking it up on Microsoft’s developers’ site, so that it might give users there the option of pinning the Web app to the taskbar instead of the Quick Launch toolbar (which goes away in Win7). But you can still pin it there eventually, after creating an icon for it first on the desktop.

The moment you create the shortcut, Chrome transfers the Web app to its stand-alone window. And just as Chrome doesn’t give the person installing the browser many options, like where to stick the Start Menu entry or whether to even make one, Chrome doesn’t stop to ask what the desktop shortcut should be called. So you often have to rename it the old-fashioned way, which isn’t difficult, but it does leave you feeling sometimes that the Google way of doing things is to force the user to fall into the default mode, and correct it later if that’s wrong.

Firefox/Prism’s installation method is precisely what and where a long-time Firefox user would expect it to be: Once the Prism add-on is installed, and you’re running the Web app you want to install in the browser, in the Tools menu, you’ll find Convert Website to Application. From there, you’ll get a dialog box letting you choose how much or how little of Firefox’s normal browser components you want displayed, as well as where you want shortcuts to appear (same problem here with not detecting Windows 7). You’ll also get an opportunity to change the icon, and you might want to do that. By default, the icon given to a Web app is the “favicon.ico” graphic that normally appears in the browser’s address bar, which doesn’t look all that swift magnified on the desktop — often like a Commodore 64 sprite.

Once that job is done, Prism leaves open the Web app you launched in Firefox, so the “conversion” process isn’t actually complete, even after the shortcut appears on the desktop. I’ve had a bit of a debate with myself over whether this is the behavior an everyday user would expect, or rather the complete transfer of the app to a separate window that Chrome provides…and here, I think Chrome’s final behavior makes the most sense. In this particular instance, I think Chrome exhibits the behavior most users would expect.

On the other hand, Chrome’s commands are a little less intuitive than Firefox’s, and Prism gives the Firefox user more self-explanatory choices up front. So thus far, I would give the edge overall to Firefox/Prism, although it would be nice if Prism gave the user a check box enabling her to relocate the running Web app to that separate window instantly.

Next: Managing the behaviors a browser can’t control…