If you look at the pros and cons of both architectures, there is one glaring problem with Chrome OS: if the network goes down, so does the app. Sure, the network may go down less and less but do we really want to lose all use of an app when the network isn’t available? What if we’re in an airplane? Or on the South Pole? Or in an underground bunker? Should computing devices become useless without the network?

On the other hand, the advantages of the thin client are clear. Data stored in a device-independent fashion. Seamless backup and sync. Immediate and simple software updates that flow out to devices. I’m sure I’m forgetting more.

The next generation of apps are bringing these advantages together. SugarSync is a good example, with seamless clients running on Windows, Mac, BlackBerry, iPhone, Web and Mobile Web. If I go offline, all the changes I make are queued up and transmitted when I go back online. Installed clients are auto-updated over the network when needed. (I would also add SMS as a potential client although SugarSync doesn’t use it.)

The challenge is: does every app have to build six or eight or ten clients? Or can development tools improve to the point that this gets much less expensive and time consuming to do right?