The thing is, even if all of the currently "common" hardware features become part of some standard, there will be additional features in the future that will not, and browser-based applications will not be able to take advantage of this hardware as quickly as native applications can.

So, as long as you're comfortable waiting for WC3 to "bless" support for a device, and then browser vendors to turn out apps (which are these so-called irrelevant native apps themselves), then in that case I agree.

The way this works is simple: The Browser becomes the OS. Where you had an abstraction protected by API's in a previous OS, you now just have an interface provided to you by the local client (browser). All of this can be abstracted easily enough, and will be .. its already happening on Android, where you can access GPS and Accelerometer from the browser ..

It depends, also, on the hardware integration. Hey, good thing there is a VM around to handle this issue.

It won't be the VM or App developers slowing down progress in this area. It will be the "standards" issue so that developers don't have to write three different modules for three different accelerometers and properly load the right one depending on which iPhone or Palm Pre you have. That will definitely slow things down as vendors fight over which of their API is "the bestest" for all phones/netbooks.