MeeGo 1.2 on the N900

Definitely rewriting the toolkit stack is what kept Nokia from releasing a second Maemo phone, but your timing of events is mixed up. That started with the Harmattan cycle, in 2009, and well before MeeGo. For whatever reason, after the Trolltech purchase, someone decided that they had to dogfood Qt and Qt alone in all of their products, immediately, regardless of the costs.

The Deb-to-RPM transition came after the Maemo/Moblin->MeeGo merger, and isn't really related. Little more than a LSB-compliance move. So while Harmattan->MeeGo might could have been "a bump in the road" as you put it, Fremantle->Harmattan before it is what actually kept Nokia from expanding its offerings.

MeeGo is not simply "Maemo 6" or even "Maemo 7." In spite of Nokia's product-release problems, MeeGo has some significantly orthogonal goals, such as cross-UX-compatibility, and that's the really significant part. It may even be healthier for MeeGo that Nokia isn't pursuing it as hard, since that "Qt-only" emphasis (which was strategic- but not engineering-based) could prove to be thorny again (see Moblin's Clutter layer, for example).

Nate