Previous | Beginning | Next
Notes
Substitutability is the property that different elements have if they are grouped -- and then treated -- in terms of their commonality; variations that are not relevant to their use are not a part of this abstraction, or rather they do not dominate. In the context of OO, such transparency normally equates to the use of inheritance and polymorphism, and bound by the Liskov Substitution Principle [Liskov1987, Coplien1992]. C++ is a rich language, supporting many other mechanisms for communicating common usage to both the human and machine reader, in addition to its many forms and uses of inheritance. A fuller exploration of the commonality and variability mechanisms in C++ supports the view that the mechanisms of the language can be used to greater effect than just in support of a subset-OO style [Coplien1999]. The behavioural concepts behind LSP can be generalised to cover all forms of substitutability. Classification is the basis of substitutability, and this notion is enforced to differing degrees by the compiler and developer. Where one element can be classified as another, or in terms of another, then that element may be used in place of the other element transparently. The degree of substitutability varies across specific mechanisms and problems. Transparency is often not completely crystal clear, and transparent should not be taken to mean "not there". If something is not immediately visible, it may reveal itself under a change of circumstance, e.g. the effect of failure in distributed systems cannot be made totally transparent, and thus distributed programming models cannot seamlessly and completely be made to look like local, sequential programming models.
|
|
