There are several benefits to OOP.
First, there is the obvious idea that functions and data are closely associated within a particular topic area. OOPLs make this easy to do by using namespaces (classes) within which the data and functions can be defined and the functions have easy access to the data.
This has been called encapsulation; and many non-oo languages have facilities for achieving it. It's actually very each to create encapsulations in C that are even better than most OOPLS can achieve. But overall OOPLs offer adequate support for it.
The second is inheritance. The relationship has been much maligned in recent years, and with reason. Much of the recent resistance is due to the fact that statically typed languages have required some kind of inheritance in order to support polymorphism. This has led to an unfortunate menagerie of different kinds of inheritance. (public, private, protected, virtual, implements, extends...)
In reality inheritance should not be necessary for polymorphism. Duck typing works as well; but is more prevalent in dynamically typed languages.
The real benefit of inheritance is the ability to inherit implementation. It is handy to have a suite of encapsulated data and functions that can be inherited into other encapsulations.
The third benefit of OO is the strongest. Dynamic Polymorphism. This facility allows the programmer to safely invert source code dependencies against the flow of control. This could be done in C by using pointers to functions, but was risky and awkward. In OOPLs it is safe and natural.
The ability to invert dependencies makes it much easier to create flexible and robust architectures. Without it architectures are dependent upon the function call hierarchy. With dependency inversion the architecture of the system is _independent_ of the function all hierarchy. This allows plugins, and high level modules that are independent of the low level details they control.
There is much more that could be said about these benefits. But perhaps the most important is that these benefits are completely orthogonal to the benefits of procedural and functional programming. A functional programmer can ALSO do OO and gain the benefits of both. The best programmers use procedural, OO, and Functional together and good programming languages seamlessly support procedural code, OO code, and Functional code without bias.
First, there is the obvious idea that functions and data are closely associated within a particular topic area. OOPLs make this easy to do by using namespaces (classes) within which the data and functions can be defined and the functions have easy access to the data.
This has been called encapsulation; and many non-oo languages have facilities for achieving it. It's actually very each to create encapsulations in C that are even better than most OOPLS can achieve. But overall OOPLs offer adequate support for it.
The second is inheritance. The relationship has been much maligned in recent years, and with reason. Much of the recent resistance is due to the fact that statically typed languages have required some kind of inheritance in order to support polymorphism. This has led to an unfortunate menagerie of different kinds of inheritance. (public, private, protected, virtual, implements, extends...)
In reality inheritance should not be necessary for polymorphism. Duck typing works as well; but is more prevalent in dynamically typed languages.
The real benefit of inheritance is the ability to inherit implementation. It is handy to have a suite of encapsulated data and functions that can be inherited into other encapsulations.
The third benefit of OO is the strongest. Dynamic Polymorphism. This facility allows the programmer to safely invert source code dependencies against the flow of control. This could be done in C by using pointers to functions, but was risky and awkward. In OOPLs it is safe and natural.
The ability to invert dependencies makes it much easier to create flexible and robust architectures. Without it architectures are dependent upon the function call hierarchy. With dependency inversion the architecture of the system is _independent_ of the function all hierarchy. This allows plugins, and high level modules that are independent of the low level details they control.
There is much more that could be said about these benefits. But perhaps the most important is that these benefits are completely orthogonal to the benefits of procedural and functional programming. A functional programmer can ALSO do OO and gain the benefits of both. The best programmers use procedural, OO, and Functional together and good programming languages seamlessly support procedural code, OO code, and Functional code without bias.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
