Question

In the following code I have placed anonymous structures inside of my class declaration to hopefully improve the readability of it.

class example {
    private:
        struct barrier {
            boost::barrier playlist_avaliable;
            boost::barrier display_sync;
            barrier( ) : playlist_avaliable( 2 ), display_sync( 3 ) { }
        } barrier;
    public:
        example( ) { }

        void playlist_avaliable( ) {
            barrier.playlist_avaliable.wait();
        }
};

Is this any better than the alternative of having variables with names like ‘barrier_playlist_avaliable’?

Answer 1

Ok, some random thoughts:

1. Leave out the ‘barrier’ after
   struct. It is redundant.
2. It’s
   barely more than syntactic sugar. It
   is a nice way to group variables.
3. Whether you should use the struct may be
   influenced by the question how much
   transparent the inner struct should
   be to the class example. For example, do you want the members of the struct to initialized in the constructor of example?
4. A similar question goes for the destructor.
5. It is more writing. This is bad.
6. It helps wrapping a lot of annoying stuff into a single object. Interesting for copying these members or for wrapping a set of rapidly rewritten internals into a single name. This might help.
7. you should keep it consistent throughout your code, in whatever way you define consistent.
8. Opposite of (1): Keep the inner struct namend, but change the name slightly. Maybe it is not too far to a situation where you want to use this struct in a more extensive way.