CS102: Introduction to Computer Science II

Inheritance in C++ 

In this section we introduce another important aspect of object-oriented programming. Inheritance is the ability for one class to be related to another class in much the same way that people can be related to one another. Children inherit characteristics from their parents. Similarly, C++ child classes can inherit characteristic data and/or behaviors from a parent class. The child classes are often referred to as subclasses or derived classes, and the parent class is often called the base class or superclass. 

Figure 8 shows the built-in C++ collections and their relationships to one another. We call a relationship structure such as this an inheritance hierarchy. For example, the string is a child of the sequential collection. In this case, we call the string the child and the sequence the parent (or subclass string and superclass sequence). This is often referred to as an IS-A Relationship (the string IS-A sequential collection). This implies that strings inherit important characteristics from sequences, namely the ordering of the underlying data and operations such as concatenation, repetition, and indexing.

Figure 8: An Inheritance Hierarchy for C++ Collections

Vectors, arrays, and strings are all types of sequential collections. They all inherit common data organization and operations. However, each of them is distinct as well. The children all gain from their parents but distinguish themselves by adding additional characteristics.

By organizing classes in this hierarchical fashion, object-oriented programming languages allow previously written code to be extended to meet the needs of a new situation. In addition, by organizing data in this hierarchical manner, we can better understand the relationships that exist. We can be more efficient in building our abstract representations. 

A child class inherits both behaviors and properties from the parent subject to some access restrictions. These variables and functions become members of the derived class. A virtual function is a member function which is declared within a base class and is overwritten by a derived class. In C++, the keyword virtual is used. A simple example of using a virtual function in C++ is shown below. In this example, the two derived subclasses inherit the printType method from the Base class.

ActiveCode: 1 Using a virtual function with inheritence (virtualfunction)

1.13.1. Logic Gates and Circuits To explore this idea further, we will construct a simulation, an application to simulate digital circuits. The basic building block for this simulation will be the logic gate. These electronic switches represent Boolean algebra relationships between their input and their output. In general, gates have a single output line. The value of the output is dependent on the values given on the input lines. 

AND gates have two input lines, each of which can be either 0 or 1, representing false or true , respectively. If both of the input lines have the value 1, the resulting output is 1. However, if either or both of the input lines is 0, the result is 0. OR gates also have two input lines and produce a 1 if one or both of the input values is a 1. In the case where both input lines are 0, the result is 0. 

NOT gates differ from the other two gates in that they only have a single input line. The output value is simply the opposite of the input value. If 0 appears on the input, 1 is produced on the output. Similarly, 1 produces 0. Figure 9 shows how each of these gates is typically represented. Each gate also has a truth table of values showing the input-to-output mapping that is performed by the gate.

Figure 9: Three Types of Logic Gates

In order to implement a circuit, we will first build a representation for logic gates. Logic gates are easily organized into a class inheritance hierarchy as shown in Figure 11. At the top of the hierarchy, the LogicGate class represents the most general characteristics of logic gates: namely, a label for the gate and an output line. The next level of subclasses breaks the logic gates into two families, those that have one input line and those that have two. Below that, the specific logic functions of each appear.

Figure 10: Circuit

In order to implement a circuit, we will first build a representation for logic gates. Logic gates are easily organized into a class inheritance hierarchy as shown in Figure 11. At the top of the hierarchy, the LogicGate class represents the most general characteristics of logic gates: namely, a label for the gate and an output line. The next level of subclasses breaks the logic gates into two families, those that have one input line and those that have two. Below that, the specific logic functions of each appear.

Figure 11: An Inheritance Hierarchy for Logic Gates

We can now start to implement the classes by starting with the most general, LogicGate . As noted earlier, each gate has a label for identification and a single output line. In addition, we need methods to allow a user of a gate to ask the gate for its label. 

The other behavior that every logic gate needs is the ability to know its output value. This will require that the gate perform the appropriate logic based on the current input. In order to produce output, the gate needs to know specifically what that logic is. This means calling a method to perform the logic computation. The complete class is shown in Listing 8.

Listing 8

A protected member variable or function is similar to a private member but it has the additional benefit that they can be accessed by derived classes. The access keyword protected is used for this. 

At this point, we will not implement the performGateLogic function. The reason for this is that we do not know how each gate will perform its own logic operation. Those details will be included by each individual gate that is added to the hierarchy. This is a very powerful idea in object-oriented programming. We are writing a method that will use code that does not exist yet. The parameter virtual is a reference to the actual gate object invoking the method. Any new logic gate that gets added to the hierarchy will simply need to implement the performGateLogic function and it will be used at the appropriate time. Once done, the gate can provide its output value. This ability to extend a hierarchy that currently exists and provide the specific functions that the hierarchy needs to use the new class is extremely important for reusing existing code. 

We categorized the logic gates based on the number of input lines. The AND gate has two input lines. The OR gate also has two input lines. NOT gates have one input line. The BinaryGate class will be a subclass of LogicGate and will add two input lines. The UnaryGate class will also subclass LogicGate but will have only a single input line. In computer circuit design, these lines are sometimes called "pins" so we will use that terminology in our implementation.

Listing 9

Listing 10

Listing 11

The only thing AndGate needs to add is the specific behavior that performs the Boolean operation that was described earlier. This is the place where we can provide the performGateLogic method. For an AND gate, this method first must get the two input values and then only return 1 if both input values are 1. The complete class is shown in Listing 11. 

We can show the AndGate class in action by creating an instance and asking it to compute its output. The following session shows an AndGate object, gand1 , that has an internal label "gand1" . When we invoke the getOutput method, the object must first call its performGateLogic method which in turn queries the two input lines. Once the values are provided, the correct output is shown.

The same development can be done for OR gates and NOT gates. The OrGate class will also be a subclass of BinaryGate and the NotGate class will extend the UnaryGate class. Both of these classes will need to provide their own performGateLogic functions, as this is their specific behavior. 

We can use a single gate by first constructing an instance of one of the gate classes and then asking the gate for its output (which will in turn need inputs to be provided). For example:

1.13.2. Building Circuits Now that we have the basic gates working, we can turn our attention to building circuits. In order to create a circuit, we need to connect gates together, the output of one flowing into the input of another. To do this, we will implement a new class called Connector . 

The Connector class will not reside in the gate hierarchy. It will, however, use the gate hierarchy in that each connector will have two gates, one on either end (see Figure 12). This relationship is very important in object-oriented programming. It is called the HAS-A Relationship. Recall earlier that we used the phrase "IS-A Relationship" to say that a child class is related to a parent class, for example UnaryGate IS-A LogicGate .

Figure 12: A Connector Connects the Output of One Gate to the Input of Another

Now, with the Connector class, we say that a Connector HAS-A LogicGate meaning that connectors will have instances of the LogicGate class within them but are not part of the hierarchy. When designing classes, it is very important to distinguish between those that have the IS-A relationship (which requires inheritance) and those that have HAS-A relationships (with no inheritance). 

Listing 12 shows the Connector class. The two gate instances within each connector object will be referred to as the fromgate and the togate , recognizing that data values will "flow" from the output of one gate into an input line of the next. The call to setNextPin is very important for making connections (see Listing 13). We need to add this method to our gate classes so that each togate can choose the proper input line for the connection.

Figure 13: Circit of NOT(AND(ganda,gnadb)OR AND(gandc,gandd))

ActiveCode: 2 Implementing our desired circuit (desiredcircuit)

Source: Brad Miller, David Ranum, and Jan Pearce, http://cs.berea.edu/cppds/Introduction/ObjectOrientedProgrammingDerivedClasses.html
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License.