On top of that, this "solution" doesn't really solve anything -- it still takes time away from development, requires developers to learn the new abstraction layer, and won't make up for functional incompatibilities/missing features between different toolkits or versions. ?It sounds like a great idea in theory -- and IIRC, the Dillo renderer (but not the rest of the interface) is already built on an abstraction layer -- but in practice it would mostly amount to needlessly re-inventing the wheel.
Disclaimer: I'm not one of the core developers, and this is only my personal view.
~Benjamin
It sounds like your issues would be more effectively resolved if you posted them to the fltk mailing list. Or, the news & blog threads concerning fltk versioning. Although from what they have stated, almost all development has occurred within the fltk-1.3 branch for the past year, still somebody posted concerning pushing further development on fltk-2.0 recently. <sigh> I'm just going with the flow here.