No, it's definitely not my idea (I believe igor introduced me to the term, actually), but it's definitely the best way I've heard the concept described.

I wish I knew. Is USB's current poor power saving a technical limitation or simply due to the applications most USB dongles end up in (definitely not attached to power-sipping mobile devices)?

Let's look at it another way. How many people with very modular desktops ever upgrade them (install a new graphics card, CPU, hard drive, whatever)? How many people with modular laptops ever upgrade them (RAM, hard drives, wireless connectivity, whatever)?

I don't have any hard numbers to back this up (other than observing friends, family, and coworkers), but I'd say very few people ever upgrade their computers. They're much more likely to simply purchase new machines.

Accepting that, how much sense do you think it would make for Nokia to make feature(s) like cellular data modular? Not much, I'd say. It's an awful large compromise to ask of the vast majority of users who will never, ever take advantage that modularity.

Let's propose something else. What if, instead of a costly (to the customer) modular option Nokia offered several different devices for the Maemo platform. Each with a different set of options, some lower-end, some higher (much like Nokia's current cellular phone lineup).

That way, you could pick the device that suits you (say, the lower-end $250 tablet without 3G, accelerometers, FM RX/TX, GPS, or a hardware keyboard etc.) and somebody else could pick the device that suits them ($600, all the fixin's, hardware keyboard, etc.) and you're both happy without having to deal with the modularity compromise (which is a big one to swallow on mobile devices).