First look: Chrome OS beta's Achilles' heel is its reliance on the Web
- — 13 December, 2010 22:10
Computers and their software today are too complicated, and users are increasingly looking at iPads and cloud-based services such as Google Docs to handle the basics that most of us stick to: document editing, photo management, emailing, Web browsing, and the like. Running Office on a PC or Mac is beyond overkill for most people. Google proposes we do away with the PC altogether, at least part of the time, and replace it with Google's cloud-based laptop -- an appliance in which the Chrome browser serves as operating system. With the Chrome OS, all actions occur in the browser and the cloud.
Google announced Chrome OS in July 2009, formally introduced it 13 months ago, and then went silent. Last week, it re-introduced Chrome OS and this time gave an ETA for the real thing: mid-2011. It also distributed prototype "Chromebook" laptops to people like me for ongoing testing of what Google CEO Eric Schmidt said would be an alternative to both Windows and Mac OS X. I've now had some quality time with that laptop, its Chrome OS, and the early apps available for the platform.
Chrome OS could be an option for grandmas and office drones: people who do very basic tasks, have a single email account, don't often share documents and data with other applications, don't use professional features such as revisions tracking when editing documents, and work with just forms. At this stage, the Web apps available for Web browsers such as Chrome OS -- including Microsoft Office Web Apps, Google Docs, and the first apps available at Google's new Chrome Web Store -- are rudimentary at best. And they don't play well with each other.
Unless Google and others succeed in developing really compelling Web apps that also operate with the rest of the Web, I believe the concept of a cloud laptop will fail. Yes, Office and many other apps are way too complicated, but Chrome apps are today way too simplistic and limited. I would point Google and Chrome developers to the apps you see on the iPad as examples of a better balance between simplicity and capability, where sophistication is not sacrificed.
It's clear that Google still has a long way to go to make Chrome OS viable; six months out from formal rollout, the operating system and the apps are in no way ready for prime time. I'm frankly surprised how primitive it all is half a year away from launch, but Google did warn that it has work to do. Maybe there's a lot going on behind the scenes that will quckly come together to make Chrome OS worth considering when it is formally released. Anyhow, keep Chrome OS' early nature in mind -- the prototype laptops, for example, aren't even running a beta version of the Chrome 9 browser that they'll ship with, but instead use a version of today's Chrome 8 browser.
The Chrome OS experienceThe Chrome browser is very spare, following the general Google strategy of eliminating clutter. What you get is simply the Chrome browser as your operating system. Your "desktop" is merely a browser tab with the icons of the apps you've installed, and any app you open is just a Web page in a browser tab. You won't see many floating windows or dialog boxes. The available on-screen controls are very simple: Back, Forward, Refresh, Add to Favorites, and Settings. You can also add a Home button to the on-screen controls, an option that is turned off by default. Additionally, there's a Search key on the laptop -- that's it.
At first, it's disconcerting to do work entirely in a browser -- I kept wanting to go to the equivalent of the Mac OS X Dock or Windows Start menu to access my apps and Web page favorites. Instead, you open a new tab to get your "home screen" equivalent each and every time you want to get to your apps. You also need to open a new tab to get to the menu to see your bookmarks. It's more work to switch tasks in Chrome OS than in Windows or Mac OS X because there's no quick-access mechanism yet.
You'll find basic interface controls in the Chrome OS, such as for the default font size for Web pages. On any Web page you're viewing, you can zoom in the usual way, by pressing Ctrl-+. But forget about a customized user interface or the ability to add fonts, desktop backgrounds, and the like -- at least in this early version. It's a spare, generic experience. Aesthetically, Google's vision of cloud computing is not personal computing. (You can apply Chrome browser themes from the Web Store, but they're ugly.)
There is local storage on the Chromebook, but there's no equivalent of a file system, so don't expect drive icons or folders. If you try to download a file, it's placed in a floating window called Downloads that acts like a folder for Web apps; you can upload files from that "folder" into those apps when you click their Browse for File buttons. This upload/download via a scratch space also is time-consuming compared to the drag-and-drop and open-in-app functionality we're used to on desktops and smartphones.
If you want to print from a Chromebook, you have to use Google's forthcoming CloudPrint service, which will work with printers designed to use it and with printers connected to a Windows, Mac OS X, or Linux PC (using the PC as a waystation) -- similar in approach to Apple's disappointingly limited printing capability in iOS 4.2. Right now, CloudPrint works only via Windows PCs, after an amazingly confusing and complicated install process. I couldn't get it to work from a Windows XP virtual machine on my Mac, but InfoWorld Editor in Chief Eric Knorr did get it to work from his XP laptop. But we can't test what the direct printing is like yet.
Google has bragged that Chrome OS is really fast and that Chromebooks would turn on or awaken in just a few seconds. Both statements are true, but the operating system is overall much slower to use because it lacks the interoperability that desktop and mobile operating systems provide. I turn on my computer once a day, so saving a couple minutes of boot time is meaningless. What is meaningful is all the extra time that the Chrome approach takes to switch among resources and work with files -- I do those all day long. Plus, my MacBook Pro and iPad both awaken in just seconds from sleep mode, so a Chromebook has no advantage there.
I'd gladly lose the faster bootup for faster operations. Operational speed matters much more than boot speed.
Using Chrome appsAs I mentioned previously, the first Chrome apps are primitive, like most Web apps. They also tend to assume you work with only a single app. I believe that apps are the Achilles' heel of the Chrome OS. They will need to be significantly more capable for most users to accept them. I can tell you I gained a newfound appreciation for desktop and mobile apps after using the Web apps available for Chrome OS.
For example, there's currently no unified email client. In Chrome OS, I keep separate windows open for my Gmail account, my Exchange account, and my personal IMAP email account. I have a Gmail account only because one is required to use Google tools; otherwise, it doesn't need to be open. But in Chrome OS, I have to switch back and forth between the Webmail pages for my two regular accounts. That's a pain I don't experience on my desktop or on my mobile devices, all of which support a unified mail account. In Chrome OS, my calendars and contacts are also separated -- unlike with my desktop and mobile devices. Frequent switching among these accounts costs me time, context, and flexibility.
Worse, Webmail sucks. Working with folders is very difficult, for example, and you don't typically get message previews to help you prioritize what you read in depth. And Exchange's Webmail warned me that Chrome is not a supported browser (of course not!), so I might have some functional limits, the details of which are currently unknown to me. The net result is that one of the key daily tasks -- communicating -- is harder in Chrome OS. If I used only Gmail and didn't care about folders and the like, it'd be fine -- but that's grandma-and-office-drone usage. The rest of us need more.
I had a better experience using Microsoft Office Web Apps, Google Docs, and InfoWorld's Drupal-based content management system. These Web apps work pretty much the same in Chrome OS as they do in a Windows or Mac browser, which is to say good enough for basic editing and formatting. (Mobile browsers don't support them, unfortunately.) You can't track revisions in either Office Web Apps or Google Docs, so they may not be sufficient for many business workflows. And Chrome OS doesn't support their drag-and-drop text editing, but they are fine for basic document work.
I witnessed some odd behavior in Drupal: In Chrome OS, Drupal would sometimes apply unwanted span tags to my text when I deleted paragraph breaks, which never happens in my Mac or PC browsers. The cursor also tended to jump around the screen randomly, whether I used an external mouse or the prototype Chromebook's quirky trackpad. I didn't experience these issues in Office Web Apps or Google Docs, though.
I also lodged a productivity issue with Office Web Apps and Google Docs: Getting files to and from them is a pain. To open a file attachment for editing requires first downloading it to that Downloads area, then uploading it into the editing app. Emailing it to someone requires a similar chain of steps. That's an issue in both Chrome OS and in these particular services. Chrome OS doesn't facilitate interapplication communication, and these apps assume you use only them.
Plus, copying and pasting text across apps results in formatting being partially or fully removed; in some cases, lots of unnecessary HTML formatting is added (such as span and o tags, a frequent problem with HTML exports in Microsoft Office). This issue is common even in the desktop Web, where the open source MCE and other rich text facilities assume that text comes from Microsoft Office only and can't properly handle other sources. Without a local version of Office to use as a waystation, Chrome OS suffers severely from this transportability flaw.
Right now, all these apps require ongoing Internet access. Google says it is revamping Google Docs to support HTML5's offline storage mechanism, so you can work on documents when not connected via Wi-Fi or Verizon 3G service (the only connectivity options). Support for offline storage will be critical for Chrome OS' success. Without it, a bad Internet connection can cause you to lose your data, as changes you make go poof when you try to save them or upload them to the cloud server the app uses. This dependency on live Internet connectivity is one aspect of Chrome OS that makes me very nervous.
Also, the cloud dependency means, at least right now, you can't launch an "installed" application when disconnected. The app on your home screen is just a link to the app on a Web server somewhere. If you're not at home or at an office with a good Wi-Fi network, that means you'll never be sure if you can launch an app. After all, 3G service is spotty and inconsistent: Even within a fixed location such as a home, you may get a signal in one location but not another, and certainly if you commute by train or bus you know that the 3G signal comes and goes. To address this issue, maybe Google will use HTML5's offline storage capacity to keep "installed" applications locally cached, not just their data, so you can launch them at any time. We'll see.
In using Chrome OS over both Wi-Fi and Verizon 3G service (which is slow to reconnect after you've been idle or put the Chromebook asleep), I frequently noted the slowdowns as applications and documents loaded. That didn't bother me for the InfoWorld CMS, which is a Web app whose pauses I'm already used to. But it was disconcerting in Office Web Apps and Google Docs compared to using a local app such as Microsoft Office or Apple iWork. I'll probably get used to the pauses, but for now it triggers fear that I'll lose my data because of a connectivity snafu -- not a worry about when using my MacBook Pro or iPad, as they can always save locally.
At this point, there aren't any serious Chrome apps to explore how far you can go with Web apps. I'll do a follow-up report when there are.
Using the ChromebookThe all-black, unlabeled Cr-48 laptop that Google has provided for beta testers is definitely not something you'd want to buy. It's a cheap system whose keyboard is a bit imprecise and thus typo-prone, and whose trackpad is intermittently unresponsive and subject to random selection and cursor movements (known bugs at the Chrome OS help site). But Google was clear the Cr-48 is not a production-quality device. Acer, Samsung, and perhaps others will make the real thing, and they'll likely have better quality.
However, the prototype Chromebook does demonstrate the assumptions Google has for real models. Basically, they're small laptops, with 12-inch screens and no hard drves or optical drives, so they weigh only about 4 pounds. The 12-inch screen felt tight to me, even after I upped the default font size by 25 percent -- a 13-inch screen would be better for my middle-aged eyes.
You get a full-size keyboard that lacks two critical keys: Page Up and Page Down. That makes it hard to scroll through text-oriented applications, and I found no shortcuts to simulate these essential keys. The keyboard is very much modeled after an Apple keyboard, and Apple dropped Page Up and Page Down years ago from its laptops -- an Apple interface decision that should not be copied. You can scroll using the trackpad with the two-finger gesture common on other operating systems, but it's not the same as paging up and down.
There's also no Caps Lock key (a Search key takes its place); Google said the goal was to eliminate "yelling" in website comments. That's a charmingly naive rationale for people who must only spend time on the Web. Getting rid of it to control silly Web commenters is dumb social engneering on Google's part. (What's next: getting rid of the question-mark symbol to prevent dumb questions?) I can live without a Caps Lock key, but it is actually useful in writing; fortunately, Chrome OS has a setting to make the Search key act as a Caps Lock key instead.
The Chromebook also gets rid of the function keys that PCs borrowed from mainframes 30 years ago. I won't miss them, although I'll be curious how remote control applications such as the promised Citrix Receiver for Chrome OS will deal with Windows Server apps that use them. Instead, there's a row of keys that are heavily inspired by Apple MacBook keyboards, with keys for brightness, audio volume, full-screen mode, and tab switching (which doesn't work yet), plus Web-oriented keys for Back, Forward, and Refresh. No Home key, though. I also can't find a way to enter accented letters or special symbols such as the euro symbol.
There's not much else on a Chromebook. The protototype has one USB port to support mice, keyboards, and (not yet implemented) storage. There's also an SD card slot, an audio jack, and a VGA video port. Nothing happened when I inserted an SD card or USB thumb drive, and attaching an external USB drive just caused it to click repeatedly. At this point, the only files accessible are those stored in the Downloads "folder" (a browser tab like everything else) after being downloaded from the Web. (Google's help page suggests the storage device capabilities aren't yet implemented.)
In any event, it's clear that Chromebooks will be simple devices. I'd add a second USB port, but there's no reason to lard up Chromebooks with all the ports that most laptops have these days. The support for Bluetooth devices such as keyboards also helps reduce the need for dedicated ports.
Looking forward to the next beta roundHardware, of course, is not what the Chrome OS is all about. Frankly, the Chromebook device is nothing more than a package to contain Google Docs and other Web services from Google and others. It's a Google Docs-plus-Web appliance.
Conceptually, that makes sense -- if the apps are rich enough. If the final Chrome OS' app universe six months from now is similar to what's available today, the Chrome OS will be at best a niche product: a simple PC on which Grandma only reads email and IMs, peruses the Web, and shares photos of the family as well as a simple PC that requires no IT administration for office drones who work only in Google Apps. It would also have niche uses as a survey tool by marketers at shows and malls, and (if there's a non-3G version) for bedside data entry in hospitals and clinics.
Done right, the Chrome OS could power the typical family computer and the common business computer for use within the comfort of local, reliable Wi-Fi networks. (It'll need parental controls and remote-wipe capabilities, respectively, though.) That would create an interesting new world, with Chromebooks becoming the first commonly used thin client (the netbook reinvented); iPads and future Android tablets staking out the role of the highly personal, use-it-anywhere "compu-entertainer"; and traditional laptops and PCs finding themselves confined to a role as the new workstation for specialized tasks such as graphics creation. Every home would have a PC or Mac, a Chromebook or two, and a bunch of iPads or Android slates. Most businesses would have a Chromebook on every desk, lots of slates and smartphones for those not confined to a desk, and a just a few workstation PCs for the CFO, in-house developers, and the creative teams.
But in its current incarnation, the cloud-only Chrome OS is no threat to the local-plus-cloud PC nor to the new generation of local-plus-cloud slates such as the iPad.
I'm hoping the folks at Google, despite the apparent slow progress, have been working on a compelling, rich experience that doesn't confine itself to Google's universe of services or to today's primitive style of Web apps. I'll be looking for tangible signs of this during the next several months of Chrome OS' beta development. Stay tuned!
This article, "First look: Chrome OS beta's Achilles' heel is its reliance on the Web," originally appeared at InfoWorld.com. Follow the latest developments in cloud computing and mobile technology at InfoWorld.com.
Read more about cloud computing in InfoWorld's Cloud Computing Channel.