A Critique of the Inheritance Principles of C++

Although multiple inheritance (MI) is al- ready a feature of the C# language, there is a debate going on about its good and bad sides. In this paper, I am defending MI. At the same time, many of the rules and principles of inheritance in current C# seem to me to need improvements. The suggested modifications are relatively simple, do not introduce many new re- served words, and should not affect other parts of the language. I do not find the current rules totally adequate even for single inheritance (SI). The problems lie in the mean- ing of access levels and in the redefinability of virtual functions. Additional inconsistencies in C# virtual functions appear in so-called independent multiple in- heritance (IMI), which is in principle the easy case of MI. The most difficult problems are caused by so- called fork-join inheritance (FJI), which is the most complicated kind of inheritance. One essential cause of complexity is private inheritance because of its intransitive nature. The main idea sug- gested here is that, simply put, private inheritance should be implicitly "non-virtual" and public inheri- tancs "virtual." This is actually a simplification of the language, at least from the programmer's if not from the implementor's point of view. The main rule has