VS2005: Uh, wow! Late change fixes System.Nullable type confusion....

David sent me this gem recently and it frankly surprised me a bit to see (what I consider to be) such a high-impact change to C# and the CLR so late in the game.

On the other hand, I think the solution is quite elegant and makes nullablility behave exactly as you would expect.  I think the whole issue with losing nullability for boxed types was a brain-dead obvious problem, but after it surviving this late in the development schedule I had assumed it was a dead issue.

Note: If I recall correctly, the new C# behavior is how the possibly-null-values syntax behaved with the C-Omega (Cω) language from MS Research.

Regardless, its nice to see Microsoft's responsiveness to the developer community and eagerness to choose to make it Right instead of just making it work.   This, in my opinion, is what seperates the .NET world from the Java world.   Both are quality tools, but its the individual design decisions in areas like delegates, generics, and other mainstream feature improvements that give the .NET Framework the advantage.