Hi Matt,
In a lot of circles the choice of 'engine/codebase/framework/API/tools' is an extremely contentious topic and I don't want to start another argument now.
The one thing that reaches out and grabs me from your introduction is that you seem to be painting yourself into a windows-only corner. If you're going to take a year to develop your toolset/api/marketing strategy/first game, then why not do something that's more likely to be cross platform - revenue from Mac sales should not be ignored. I'm aware of things like Wine and Mono but they're still a bolt-on, extra-hoop for your users to jump through, with no guarantee of compatibility or future support.
My codebase is entirely windows/directx and I'm unhappy about it (recent updates mean I should be able to switch 3d api's and host platform pretty easily, but I've still to do it and test it).
If I was starting again today, in your situation, I would seriously consider licensing and learning a 3rd party api or even a complete solution like BlitzMax - there are plenty of cheap and proven paths to choose from. Demote your 2d engine to a sideline and improve it with what you learn in the future.
This way you'll get a game out of the door (no worries of 'engine feature creep') which ultimately (I imagine) is what you want to do. You might want to license your successful 'engine', so did I 5/6 years ago when I started mine, but look around at all the successful engine developers that license their stuff - they always tend to be primarily engine developers, not game developers. The obvious exception is Torque/Garage games, but they're not a one-man shop.
Just my thoughts
Cheers
Matt