I say dump IE core, adopt the gecko engine (or khtml, whichever...) and develop on top of that, contributing to fixes, it's not like if that was not possible. What do you have to lose?

The fact remains that IE7 lacks support of basic CSS2 features that we're hoping to implement, and it's the one stopping us.

It's just an opinion. You could either support the features, or adopt an engine that already supports them. As always, Internet Explorer still sucks, apart from the UI stupidness that's been done.