Professor Baclawski has the following feedback to the Law of Demeter for C++:

Karl, I am a little puzzled by your statement of the Law of Demeter in terms
of C++.  You refer to member functions as being "attached to" a class.  
While I can understand what you might mean by this, it is not a C++ term.
member functions are declared as members and then defined.  They are not
defined and then attached to a class.  You also speak about a member
which "sends a message".  I don't know what this could mean.  Neither
"sending" nor "messages" are C++ terms.  

Thanks, -- Ken

Ken is right; we have to formulate it in C++ terminology:

OLD FORMULATION: ----- LAW OF DEMETER (C++, object version) -------------
In any member function M defined for class C only send
messages to the following objects:
  - M's argument objects, including the object pointed to by "this" 
  - a data member object of class C.
(Objects created by the member function, or by functions which it
calls and objects in global 
variables are viewed as being passed by arguments.)
----------------------------------------------------------
NEW FORMULATION: ----- LAW OF DEMETER (C++, type version) -------------
In any member function M defined for class C only call member functions
of the following classes:
  - M's argument classes, 
  - data member classes of class C.
(Objects created by the member function, or by functions which it
calls and objects in global 
variables are viewed as being passed by arguments.)
----------------------------------------------------------

It is interesting to notice that for C++ it is easier to give the
type version, since C++ is strongly typed.

Ken, how do you like this formulation?

-- Karl