The problem with D/Nimrod/Go/Rust is the kind of people who would use them have already invested heavily in coping with the pain of C++
As someone who has written C++ for eight years now, I don't agree that life works like that. Two jobs back, I worked on a system in Prolog plus C/C++. There, I couldn't have switched to Go or D if I wanted to, binding Prolog with those language runtimes would be a big hassle (for everyone). However, in my current job, I couldn't write C++ if I wanted to. All the work is done on the top of the JVM, so a JVM language is what one has to live with :). In other words: the set of languages that can be used depend highly on the employer and the project.
The larger barrier against the success of these new languages is inertia of your co-workers, employer, etc. And that's for a good reason you touch upon: C++, Java, the JVM, .NET, etc. are all reliable languages tools used by a large community.
As someone who has written C++ for eight years now, I don't agree that life works like that. Two jobs back, I worked on a system in Prolog plus C/C++. There, I couldn't have switched to Go or D if I wanted to, binding Prolog with those language runtimes would be a big hassle (for everyone). However, in my current job, I couldn't write C++ if I wanted to. All the work is done on the top of the JVM, so a JVM language is what one has to live with :). In other words: the set of languages that can be used depend highly on the employer and the project.
The larger barrier against the success of these new languages is inertia of your co-workers, employer, etc. And that's for a good reason you touch upon: C++, Java, the JVM, .NET, etc. are all reliable languages tools used by a large community.