However, this meant adding entire new input/graphics/widget backends to the Gecko platform using a combination of WinRT and Win32 APIs, and a long slog to get those stable and responsive enough to ship to millions of users. And designing and build a new UI (even a minimal one) means we don't get existing desktop Firefox features "for free" -- stuff like private browsing, add-ons, bookmarking, password management, sync...
Even minimal browsers have a surprisingly large surface area if you want them to be usable for daily browsing. We cut out a lot of that stuff for the first release, but that also means we had a long road ahead toward a feature set that was competitive with other browsers (including desktop Firefox, or Firefox for Android). And that's not even getting into technical issues with the Metro environment, like the inability to run the NPAPI version of Flash Player...
Just as one example: When writing a browser, you need to do your own text rendering and layout. That means you can't rely on the OS to handle things like select/copy/paste. So you end up implementing your own touch-friendly selection UI, and making it match the OS behavior as much as possible. This turns out to be a minefield of subtleties and edge cases.