Google's Chrome O/S: A Fundamentally Flawed Idea That Gets Way Too Much Positive Press

Microsoft’s probably feeling pretty annoyed right about now. Back in July, and just a few days before Microsoft provided business details on the cloud-based Windows Azure operating system along with rolling out the cloud-augmented Office 2010 application suite at the Worldwide Partner Conference, Google pre-empted its competitor by unveiling the cloud-centric and Linux- and Chrome browser-based Chrome O/S. And last week, shortly after Microsoft formally launched Windows Azure at the Professional Developers Conference (including providing a January 1, 2010 production date), Google gave a follow-on Chrome O/S briefing, demonstration and source code release, where company officials opined that ‘Google Chrome OS will be ready for consumers this time next year‘ (emphasis is mine). Didn’t know about last week’s Azure news? I’m not surprised, therefore the title of this post.

I bet Oracle’s Larry Ellison isn’t too happy, either. In September of 1995, he introduced the concept of the Network Computer, which became not only an Oracle trademark but also a division of the company in partnership with IBM, Sun and others. As the Wikipedia entry states:

The NC brand was mainly intended to denote and forecast a range of desktop computers from various suppliers that, by virtue of their diskless design and use of inexpensive components and software, were cheaper and easier to manage than standard fat client desktops.

Sound familiar? However, by the end of that decade, and after substantial time, effort and money spent by alliance members, the concept was mothballed. Why? A few thoughts:

• The ‘open standards’ promoted by NC backers were neither sufficiently ’standard’ nor sufficiently feature-robust to compete against the dominant Microsoft alternative (i.e. there was not yet a HTML v5, for example)
• To the above point, the World Wide Web was in its infancy
• Connectivity was similarly immature. Many corporate environments didn’t have LANs at all, far from ones standardized on Ethernet. There was no such thing as Wi-Fi. And in the home, LANs were virtually nonexistent, as was broadband, and 28.8 kbps analog modems were the latest-and-greatest technology available to access CompuServe, The WELL, and other BBSs.

In all of these respects, admittedly, the situation is much more solid than it was nearly 15 years ago. But it’s not perfect. Witness:

• The continued longevity of proprietary standards, such as Adobe Acrobat and Flash, along with Microsoft Silverlight, and
• Coverage-incomplete, performance-strapped and expensive wireless connectivity

And users’ assumptions of what their computers should do (and how fast they should do it, not to mention form factor, battery life and other requirements) have also expanded over the past 15 years. As such, and just as the Network Computer largely failed in the mid-to-late 1990s, I feel that Chrome O/S-based hardware is also destined to substantially undershoot the expectations currently being hoisted on it both by its backers and by members of the technical press who would love nothing better than to see a credible competitor to Microsoft Windows emerge.

Google’s Chrome O/S is nothing more than the combination of a hardware- and operating system-tuned firmware image mated to a custom-tailored Linux distribution, with the Chrome browser sitting on top of the stack. Fine, so it boots in 10 seconds or so. How often do you cold-boot your computer nowadays? If you’re anything like me, you instead mostly just bring it out of standby or (worst case) hibernation, a process which takes about as long as Google Chrome O/S’s boot latency. Translation: insufficient motivation to make a switch.

More generally, I grant that Google’s Chrome O/S is open-source and its mated hardware is CPU-agnostic. ARM and its licensees love that. So do the anti-Microsoft zealots of the world. But most consumers frankly don’t care about either point. They just want a cost-effective computing and communications platform that meets their functional and other needs. To wit, I have a few questions:

• Have any of the Chrome O/S advocates comparatively priced an Atom- and Windows 7-based netbook or larger-screen notebook PC lately?
• Do they really think that consumers will be satisfied with the restricted experience of being able to only run software that operates within a Chrome browser window, even assuming those consumers are already Google service users?
• What will consumers think the first time their Internet connection goes down and their Chrome O/S-based gadget transforms into a paperweight? Yes, I know all about Google Gears (and its competitors such as Adobe AIR), but the limited offline functionality offered is a further hamper on an already limited-function Google Apps suite. As such, I think Microsoft’s cloud-augmented approach to software is the superior plan, not Google’s cloud- (and Google-) centric strategy.
• And will third-party developers be willing to support multiple CPU architectures, especially if ARM’s power consumption advantage evaporates in the Pine Trail-and-beyond era, and if ARM-based hardware is performance-strapped compared to x86-based alternatives?

Loathing of market leaders Intel and Microsoft is fundamentally fueling Chrome O/S. These are the same fundamental factors that fueled the development of the Network Computer. And in both cases, what’s missing is a credible, compelling potential customer pull for what Google’s pitching. Yeah, I think it’d be interesting to have a Google-centric widget lying around the house. But I’m a techie. It’d be incremental to the fuller-featured computer hardware I’d already own. And it’d therefore have to be very inexpensive to motivate me to pull out my wallet.

Google’s got billions of dollars in the bank. Heck, the company could probably give away Chrome O/S-based netbooks to consumers, operating on the assumption that subsequent advertising revenues would subsidize the initial investment. But it’d be a flawed assumption, because I think that even a free version of such a platform would garner little use after an initial ‘honeymoon’ period. Agree or disagree, readers?

p.s…Happy Thanksgiving to the U.S.-based segment of my audience!