Just a thought:
1. D needs capable compiler programmers.
2. Converting C++ programmers is a good recruitment strategy.
3. C++ programmers are very concerned about memory management
features.
4. When C++17 is released, current C++ programmers will go "I was
ready to leave C++, but I will give it a second chance now that
it has all these new interesting features". So 2017-2018 is a bad
period for converting C++ programmers.
It seems to me that a competitive production-ready solution that
satisfies the current crop of C++ programmers has to be
implemented within 6 months. When C++17 is perceived as being
near release the psychology will be "I will wait for C++17 before
looking at other languages".
Would it be possible to make a fully compatible
unique_ptr/shared_ptr solution that acts as the default memory
management scheme in D within 6 months?
If not, what is the strategy behind focusing on interfacing with
C++? (Evaluated in relation to the C++17 release.)
