TITLE: what use multiple inheritance?

(Newsgroups: comp.lang.c++.moderated, 27 Apr 98)


RUDOLPH: "Chris Rudolph" <chris_rudolph@claris.com>

> > Q: When should you use multiple inheritance?
> > A: There are three acceptable answers: "Never," "Rarely," and "When the
> > problem domain cannot be accurately modeled any other way."


MCNAMEE: Paul McNamee

> My favorite quote on this obviously religious issue, is by Grady
> Booch, "Object Oriented Design with Applications", Benjamin Cummings,
> 1991, page 111.
> 
>   "In our experience, we find multiple inheritance to be like a
>    parachute: you don't always need it, but when you do, you're really
>    glad to have it on hand."


MARTIN: "Robert C. Martin" <rmartin@oma.com>

I would have thought, by now, the controversy over MI would have
settled down.

MI is most useful in statically typed languages like C++ and
Java for the purpose of separating the interface of a class into 
distinct client specific interfaces.  This decouples the clients
from each other.

If class A and B both use class C, but use different methods of C,
then when A needs a change made in C, B will be affected.  If, however,
A uses an interface named AI and B uses an interface named BI,
and C multiply inherits from AI and BI, then when A needs
a change, AI will change and C will change but BI and B will remain
unaffected.

See "The Interface Segregation Principle" in the 'publications'
section of my website.