I rather like c++ closures in that they are pretty explicit about how capture works. Given the other concepts in the language it is absolutely necessary to cons

I rather like c++ closures in that they are pretty explicit about how capture works. Given the other concepts in the language it is absolutely necessary to consider how closure capture works with the copy constructor, for example. I would contrast it with objc blocks, where capture is pretty implicit and sometimes feel ambiguous. But I think that copy constructor is the root of a lot of evil performance penalty in c++. In the old days it was especially easy to over-allocate with vector or string. Move semantics fixed some of this. But passing around closures without a lot of care is also a consequence. An "easy" solution, to capture a shared_ptr, is not without its costs.