I get the sense that many of you just don't want ANY answers that aren't "Yes, you're right, we should have implemented support for WinRT's object model and runtime enviroment in an ISO C++ solution and never seriously considered the benefits of a small, platform-specific (WinRT) language extension for simplying the expression of developer intent when programming shared, ref-counted COM objects in a new hetrogeneous programming model for Metro style application development on Windows 8".
No, Charles. We want either "Yes, you are right" or "No, you are wrong, here is why". What we get from Herb is "No, you are wrong, but I won't show you why, this is all terribly difficult, you just don't understand, you don't have a complete technical spec so I will just point out to various things which you haven't yet mentioned in your posts and this will supposedly show that you, well, don't have a complete technical spec and thus I win". This is pathetic.
C++/CX is real, not imaginary or speculative like the reverse of #import with wrapper classes solution.
Right, this is your entire argument. Since we seem to be at our final words, here is the final word from me:
Me and my team are going to stay away from C++/CX. This means staying away from WinRT for some time, too.