Given I've been using this Pimpl idiom quite extensively lately I've noticed writing the same scaffolding over and over again. Being the lazy bugger as I am I looked around if there was anything available. Turns out there were proposals floated around (like Asger Mangaard's dated around May 2006). However, I did not see anything that would preserve idiom's original purpose/value - the separation of interface and implementation. So, I jogged something with that in mind. That seemed to do the job for me as something tedious like class Test { public: Test (int); Test (int, int); int get() const; bool operator==(Test const& p) const { return impl_ == p.impl_; } bool operator!=(Test const& p) const { return impl_ != p.impl_; } operator bool() const { return implementation_; } void swap(pimpl& that) { impl_.swap(that.impl_); } private: struct Internal; boost::shared_ptr<Internal> impl_; }; shrunk down do pure interface struct Test : public boost::pimpl<Test> { Test (int); Test (int, int); int get() const; }; with implementation details safely tacked away in an implementation file. However, it looked suspiciously easy. So, I am throwing it out for all the respected community to see (uploaded as vb_pimpl_0.1.zip) and to set me straight. Is there any value in that basic idea? Don't hold back your criticism. Regards, Vladimir.