Inner Classes and Shtuff

Inner Classes




By: Troy D. Travis
CMSC335






























Table of Contents



I. Introduction
II. History and Basic Facts of Inner Classes in Java
III. 4 Tyes of Inner Classes
a. Nested Inner Classes
b. Inner Classes
c. Local Inner Classes
d. Anonymous Inner Classes
IV. Advantages of using Inner Classes
V. Inner Class Uses and Strategies
a. Callbacks
b. Inner Class Strategy, Data Access Objects
VI. References































Introduction

Inner classes are a very useful concept in the java language. They can help to keep code cleaner, they can help to control access to objects and they provide a certain since of robustness in code that aids in object oriented design.
This paper will address the history of the inner class as it figures into the Java language. We will also discuss the various kinds of inner classes, their names, how many and what types of files are generated by the compiler, access restrictions and why the restrictions are what they are. We will then discuss some examples and purposes for inner classes in the development world.

History and Basic Facts of Inner Classes in Java

Up until Jdk 1.1, the java language only supported top-level classes. At the introduction of jdk1.1 Sun introduced the concept of nested/inner classes in java. This allowed a developer to use inner classes as some sort of swiss army knife—they can be used to do many different things, and can be used to reproduce some C++ features that are otherwise unavailable in Java. This java feature of inner classes is limited to the compiler. To compile code that contains inner classes you will need jdk1.1 or later, BUT because of the way the class files are generated during compilation, the compiled code can be executed on any jvm. The naming of anonymous inner classes varies from one compiler implementation to another. Sun uses numbers to provide anonymous class names. For instance, if a top-level class was named Car and one of the methods contained an anonymous inner class, that class might be named Car$1.class by the compiler.

As far as the java interpreter is concerned, there is no such thing as an inner class: all classes are normal top-level classes. There are three types of inner classes and will be discussed shortly but they are: member classes – static and non-static or ‘nested inner’ and just plain ‘inner’ classes. To make this ‘magic’ work the compile performs various ‘tricks’ that can be viewed by using the javap disassembler on the generated class files. Once disassembled it can be seen that the compiler actually inserts hidden fields, methods, and constructor arguments into the classes it generates. This information will be used to help explain the various behaviors by each type of inner class. It should also be known that a file can only contain one outer public class. Any attempt to create a file with more than one public class will cause the compiler to yell. A file can contain multiple non-public classes (inner classes) though.


Four Types of Inner Classes

When referring to inner classes we will make a reference to the containing class. This containing class is usually referred to as a/the top-level class. A top-level class is a class that is defined as a member of a package. In general, inner classes make it easier for a developer to connect objects together, since classes can be defined closer to other objects. They are classes defined from within the same java file. They are ‘contained’ inside a ‘container’ class or a ‘top-level’ class.

There are three types of inner classes. But, if we consider a static inner class then there are four. The first type of inner class we will discuss is the nested inner class or static member class (because it is a member of the top-level class). We will then go on to discuss the inner classes or plain old member classes of which these classes are instance classes of the top-level class in which it is contained. We will then discuss local inner classes, which are defined inside of a block of code or more generally a method perhaps. Finally, we will briefly discuss the anonymous inner class.


Nested Inner Classes (static member classes)
Nested inner classes are classes within classes. A nested inner class has the same behavior as any static member of a class. You can access it without initializing the parent class, and they can access the parent class’ static methods and variables, even if the variable is private. These types of inner classes are always defined with the keyword ‘static’. An access tag (public, protected, private) can be defined, but by default a nested class takes the default package access. This produces an interesting question: if the effect of marking it as static means there is only one instance of any variables, no matter how many instances of the outer class are created, how could the static inner class know which variables to access of its non-static outer class? The answer is obvious and it is that it could not know, which is why a static inner class cannot access instance variables of its enclosing class. Sun considers nested classes to be top-level classes. Since the static member class is compiled into an ordinary top-level class, however, there is no way it can directly access the private members of its container. As mentioned above in the history section, the compiler will generate extra code. This makes since, because if a static member class uses a private member of its containing class (or vice versa), the compiler automatically generates non-private access methods and converts the expressions that access the private members into expressions that access these specially generated methods. These generated methods are given the default package access, which is sufficient, as the member class and its containing class are guaranteed to be in the same package. Compiled nested inner classes are named as ‘OuterClassName$InnerClassName.class’. Where OuterClassName is the container class (or package) and InnerClassName is the name of the static nested inner class defined inside of the containing class. The $ in this name is automatically inserted by the compiler. So if you were to create a class OuterClass.java and within that class a class called InnerClass the compiler will generate two class files. This is why inner classes can be run from any jvm (as mentioned above). Since, by definition, a nested inner class is static it can be accessed from anywhere in your code by using OuterClass.InnerClass. The same applies to writing import statements.


Inner Classes (member classes)
These ‘member classes’ are implemented much like a static member class. They get compiled into separate top-level class files and the compiler performs various code ‘tweeks’ to get it to work. Inner classes differ from a nested class because they do not have a static tag to them. Therefore, every instance of an inner class requires an instance of the container class and the container class can have multiple instances of the contained class/inner class. The compiler enforces this association by defining a synthetic field named this$0 in each member class. This field is used to hold a reference to the enclosing instance. Every inner class (non-static) constructor is given an extra parameter that initializes this field.
A non-static inner class can be declared public, protected, or private or given the default package visibility, but as mentioned previously, member classes are compiled to top-level class files but top-level classes can only have public or package access, which means member classes can only have public or package visibility (to the Java interpreter). So if a member class is declared as protected it is actually treated as a public class and if it is declared as private it actually has package visibility. According too anonymous, “Although the interpreter cannot enforce these access control modifiers, the modifiers are noted in the class file. This allows any conforming Java compiler to enforce the access modifiers and prevent the member classes from being accessed in unintended ways.” Just a thought but can you deduce how this relates in a security centric way?


Local Inner Classes
These types of inner classes are defined inside a bloc of code, such as a method and are not available outside of the method in which it is defined, even inside class outer. However, like member classes, a local class is able to refer to fields and methods in its containing class for exactly the same reason and mechanism available to a member class; it is passed a hidden reference to the containing class in its constructor and saves that reference away in a private field added by the compiler. Also, the private fields and methods of the container class are also available to local inner classes because the compiler will insert any required accessor methods. What separates a local class from a member class is the ability to refer to local variables in the scope that defines them. There is, hence, one restriction to this: local classes can only reference local variables and parameters that are declared final. According to the Java spec, this is because the compiler automatically gives the local class a private instance field to hold a copy of each local variable the class uses and as previously mentioned the compiler will add special hidden parameters to each local class constructor to initialize the compiler created private fields. Therefore a local class isn’t actually accessing local variables but merely copies, albeit private copies, of them. And to make this work a guarantee must be in place that says these values won’t change and that is by using the keyword final. Almost magic!


Anonymous Inner Classes
An anonymous inner class usually does not have a name. They are often used in gui development to implement listeners. An anonymous class must either extend a class or implement an interface but it can’t do both like a member class. An access modifier is never specified and cannot be specified. Since the class doesn’t have a name it doesn’t get a constructor.

By now the reader should have a pretty good understanding of inner classes and how they are created via the compiler and how they are interpreted by the interpreter and how and why their access behavior exist. Next we will touch on some basic advantages to using inner classes.


Advantages of Using Inner Classes

There are quite a few advantages to using inner classes. These advantages include, but not limited to, blocs of code, integration of objects in objects and enabling a more object-oriented structure.

Blocs of Code
Nested inner classes give developers the ability to implement blocs of code that can be used to realize special functions inside a given object. Using nested inner classes a developer can integrate blocs of code inside an object to perform special tasks pretty much like method pointers in C.

Integration of Objects in Objects
By giving developers the ability to integrate blocs of code inside other objects via an inner class, a developer now has the ability to integrate objects in objects which can lead to a more object oriented design. A developer can create objects that are specific to the enclosing class. This makes the code easier to read and to maintain and leads to a more object oriented design.

More Object Oriented Structure
By allowing developers to define inner objects top-level nested classes can be defined inside another object. Multiple class files do not need to be created for every distinct object in the program. Objects can be integrated into one another, which makes the code easier to understand and makes since, from an object oriented perspective, to have classes only used by a certain class inside of that class.


Inner Class Uses and Strategies

There are various uses and many strategies implemented that take advantage of the powerful nature of inner class design. We will present two of these approaches. The first approach is in the implementation of ‘callbacks’. Java doesn’t necessarily use ‘callbacks’ like C, but we can certainly imitate them and is clearly seen in the AWT framework. The other approach is merely a design strategy used among various persistence frameworks.


Callbacks
A common technique in the C language is to pass a pointer to a function in the argument list of another function. The receiving function can invoke the passed in function using the pointer. This approach is referred to as a callback. Callbacks are very useful in situations where you want to invoke different functions without needing to know what particular function is being invoked. For example, a plotting function could receive a pointer to a function that takes a value on the horizontal axis as an argument and returns a value for the vertical axis. The plotting function could then plot any such function that is passed to it. Inner classes are very useful when you want to implement callbacks. Java, however, does not provide for pointers (direct memory addresses) to methods. Instead, interfaces are used for callbacks. In this case, a method holds an interface reference in its argument list and then invokes a method in the interface. A callback method is a method intended to be passed as a parameter. The callee can then call the passed method asynchronously and with whatever parameters it deems appropriate. One of the basic techniques used in java to fake a call back is to use inner classes to define an anonymous callback class, instantiate an anonymous callback delegate object, and pass it as a parameter all in one line. This allows a developer to efficiently have a separate bloc of code to handle the actionPeformed function of the Actionlistener interface for every graphical component object. This is one of the many features found in other languages that the java language has been able to implement.


Inner Class Strategy, Data Access Objects (DAO)’s
Let’s say a developer wants to isolate data access and manipulation in a separate layer, while also keeping member variables private and hence retaining full encapsulation and object integrity. How do we handle this?
A Data Access Object (DAO) is found when dealing with object-relational persistence. The DAO is the primary object of the ‘Data Access Object’, one of the core j2EE patterns. The Data Access Object abstracts the underlying data access implementation for the BusinessObject to enable transparent access to the data source. There is often a need for the DAO implementation to modify data on the domain object (Business Object) that is being persisted. For instance, when a new object is being inserted into a relational database table, it can sometimes be auto-assigned a surrogate primary key as opposed to a natural primary key. In a properly designed entity class, the field representing the primary key should not be modifiable to the public, as this would compromise the integrity of the object’s state. A solution to this is to create the DAO interface as an abstract static inner class of the domain object it persists. This is a viable solution because inner classes have access to the private members of the outer class. This allows the DAO implementation to have access to select private members, while still isolating the actual persistence behavior in its own class. For instance, the Customer class has three private members: name, customerNumber(natural primary key) and customerId( auto generated surrogate primary key). We want to make sure the customerId can not be changed outside of the Customer class (the business object). So to satisfy this requirement we create the CustomerDAO as an abstract inner class of the Customer class. In this example, the customerId is available only as a readable property to the public while retaining encapsulation and guaranteeing no change while the DAO implementation is allowed to modify the attribute.

In this paper we have discussed inner classes as they are applied in java. We have discussed the history and basic ideas behind inner classes. We have also discussed the various types of inner classes, the nested-static inner class the non-static inner class, the local inner class and the anonymous inner class. We discussed how inner classes are invoked, their scope and how they relate to the java compiler and interpretor. We also discussed some concepts and design practices that currently use the inner class design, such as the implementation of ‘callbacks’ and it’s use in persistence frameworks.












































References

http://mindprod.com/jgloss/callback.html
http://www.javaworld.com/javaworld/javatips/jw-javatip10.htmll
http://java.about.com/cs/javadocumentation/g/callback.html
http://www.particle.kth.se/~lindsey/JavaCourse/Book/Part1/Java/Chapter04/interface.html
http://nils.kilden-pedersen.net/DAO_pattern/InnerClassStrategy.html
http://www.unix.org.ua/orelly/java-ent/jnuy/ch03_13.htn