It seems the network computer is an idea that, like a zombie army, resolutely refuses to lie down and die. A discussion thread within the Enterprise Irregulars group this week started off with an observation about where Google expects to find market demand for its ChromeOS:
"Businesses are accelerating in their adoption of cloud-based applications like SFDC and Google Apps. As the center of gravity of enterprise computing moves here, the need for large swathes of workers at enterprises (especially more 'task' oriented workers) to have full blown local computing systems that can run executables not originating from the web itself diminishes significantly."
For the past fifteen years I have been listening to people talking about foisting restricted-function web terminals on 'large swathes' of the workforce (beginning with Larry Ellison's NC and Scott McNealy's Javastation). The thing I've always noticed is that no one ever says, 'I want one of those.' The proposition is always promoted as an ideal solution for someone else's computing needs — and almost without exception it's a use case the speaker has no direct familiarity with. Thus the Gen-Y software engineers that Google employs straight out of college apparently imagine that ChromeOS is a perfect platform for workers in call centers and manufacturing plants.
It's another of those irregular verbs, isn't it? I use Mac OS X, you run Windows, they'll be fine with ChromeOS.
In reality, call center workers are more likely to remain on full-function desktop platforms than the rest of us. Speed of response is paramount for these workers who are often interacting with multiple applications. A powerful desktop machine that turbocharges their productivity — even with a SaaS application — makes a big difference to their ability to resolve questions fast and deliver customer satisfaction.
Of course if ChromeOS taps the underlying power of the hardware platform, it may give a better performance than traditional network computer designs. Microsoft is pursuing a similar path, though from a different direction, with its use of native client capabilities in its forthcoming Windows-based browser. But Google's designers shouldn't imagine they can get away with palming off substandard performance on powerless workers across its enterprise customers. As Google well knows, speed matters. Any browser-only computer has to be able to deliver performance and user experience that's at least as good as, and preferably better than, the PCs it seeks to replace. Designing it with any lesser target in mind will condemn it to market failure.