10 min read

Modular programming enables one to organize code into independent, cohesive modules, which can be combined to achieve the desired functionality.

This article is an excerpt from a book written by Nick Samoylov and Mohamed Sanaulla titled Java 11 Cookbook – Second Edition. In this book, you will learn how to implement object-oriented designs using classes and interfaces in Java 11.

The complete code for the examples shown in this tutorial can be found on GitHub.

You should be wondering what this modularity is all about, and how to create a modular application in Java. In this article, we will try to clear up the confusion around creating modular applications in Java by walking you through a simple example. Our goal is to show you how to create a modular application; hence, we picked a simple example so as to focus on our goal.

Our example is a simple advanced calculator, which checks whether a number is prime, calculates the sum of prime numbers, checks whether a number is even, and calculates the sum of even and odd numbers.

Getting ready

We will divide our application into two modules:

  • The math.util module, which contains the APIs for performing the mathematical calculations
  • The calculator module, which launches an advanced calculator

How to do it

  1. Let’s implement the APIs in the com.packt.math.MathUtil class, starting with the isPrime(Integer number) API:
        public static Boolean isPrime(Integer number){
          if ( number == 1 ) { return false; }
          return IntStream.range(2,num).noneMatch(i -> num % i == 0 );
        }
  1. Implement the sumOfFirstNPrimes(Integer count) API:
        public static Integer sumOfFirstNPrimes(Integer count){
          return IntStream.iterate(1,i -> i+1)
                          .filter(j -> isPrime(j))
                          .limit(count).sum();
        }
  1. Let’s write a function to check whether the number is even:
        public static Boolean isEven(Integer number){
          return number % 2 == 0;
        }
  1. The negation of isEven tells us whether the number is odd. We can have functions to find the sum of the first N even numbers and the first N odd numbers, as shown here:
        public static Integer sumOfFirstNEvens(Integer count){
          return IntStream.iterate(1,i -> i+1)
                          .filter(j -> isEven(j))
                          .limit(count).sum();
        }
public static Integer sumOfFirstNOdds(Integer count){ return IntStream.iterate(1,i -> i+1) .filter(j -> !isEven(j)) .limit(count).sum(); }

We can see in the preceding APIs that the following operations are repeated:

  • An infinite sequence of numbers starting from 1
  • Filtering the numbers based on some condition
  • Limiting the stream of numbers to a given count
  • Finding the sum of numbers thus obtained

Based on our observation, we can refactor the preceding APIs and extract these operations into a method, as follows:

Integer computeFirstNSum(Integer count,
                                 IntPredicate filter){
  return IntStream.iterate(1,i -> i+1)
                  .filter(filter)
                  .limit(count).sum();
 }

Here, count is the limit of numbers we need to find the sum of, and filter is the condition for picking the numbers for summing.

Let’s rewrite the APIs based on the refactoring we just did:

public static Integer sumOfFirstNPrimes(Integer count){
  return computeFirstNSum(count, (i -> isPrime(i)));
}
public static Integer sumOfFirstNEvens(Integer count){ return computeFirstNSum(count, (i -> isEven(i))); }

public static Integer sumOfFirstNOdds(Integer count){ return computeFirstNSum(count, (i -> !isEven(i)));

So far, we have seen a few APIs around mathematical computations. These APIs are part of our com.packt.math.MathUtil class. The complete code for this class can be found at Chapter03/2_simple-modular-math-util/math.util/com/packt/math, in the codebase downloaded for this book.

Let’s make this small utility class part of a module named math.util. The following are some conventions we use to create a module:

  1. Place all the code related to the module under a directory named math.util and treat this as our module root directory.
  2. In the root folder, insert a file named module-info.java.
  3. Place the packages and the code files under the root directory.

What does module-info.java contain? The following:

  • The name of the module
  • The packages it exports, that is, the one it makes available for other modules to use
  • The modules it depends on
  • The services it uses
  • The service for which it provides implementation

Our math.util module doesn’t depend on any other module (except, of course, the java.base module). However, it makes its API available for other modules (if not, then this module’s existence is questionable). Let’s go ahead and put this statement into code:

module math.util{
  exports com.packt.math;
}

We are telling the Java compiler and runtime that our math.util module is exporting the code in the com.packt.math package to any module that depends on math.util.

The code for this module can be found at Chapter03/2_simple-modular-math-util/math.util.

Now, let’s create another module calculator that uses the math.util module. This module has a Calculator class whose work is to accept the user’s choice for which mathematical operation to execute and then the input required to execute the operation. The user can choose from five available mathematical operations:

  • Prime number check
  • Even number check
  • Sum of N primes
  • Sum of N evens
  • Sum of N odds

Let’s see this in code:

private static Integer acceptChoice(Scanner reader){
  System.out.println("************Advanced Calculator************");
  System.out.println("1. Prime Number check");
  System.out.println("2. Even Number check");
  System.out.println("3. Sum of N Primes");
  System.out.println("4. Sum of N Evens");
  System.out.println("5. Sum of N Odds");
  System.out.println("6. Exit");
  System.out.println("Enter the number to choose operation");
  return reader.nextInt();
}

Then, for each of the choices, we accept the required input and invoke the corresponding MathUtil API, as follows:

switch(choice){
  case 1:
    System.out.println("Enter the number");
    Integer number = reader.nextInt();
    if (MathUtil.isPrime(number)){
      System.out.println("The number " + number +" is prime");
    }else{
      System.out.println("The number " + number +" is not prime");
    }
  break;
  case 2:
    System.out.println("Enter the number");
    Integer number = reader.nextInt();
    if (MathUtil.isEven(number)){
      System.out.println("The number " + number +" is even");
    }
  break;
  case 3:
    System.out.println("How many primes?");
    Integer count = reader.nextInt();
    System.out.println(String.format("Sum of %d primes is %d", 
          count, MathUtil.sumOfFirstNPrimes(count)));
  break;
  case 4:
    System.out.println("How many evens?");
    Integer count = reader.nextInt();
    System.out.println(String.format("Sum of %d evens is %d", 
          count, MathUtil.sumOfFirstNEvens(count)));
  break;
  case 5: 
    System.out.println("How many odds?");
    Integer count = reader.nextInt();
    System.out.println(String.format("Sum of %d odds is %d", 
          count, MathUtil.sumOfFirstNOdds(count)));
  break;
}

The complete code for the Calculator class can be found at Chapter03/2_simple-modular-math-util/calculator/com/packt/calculator/Calculator.java.

Let’s create the module definition for our calculator module in the same way we created it for the math.util module:

module calculator{
  requires math.util;
}

In the preceding module definition, we mentioned that the calculator module depends on the math.util module by using the required keyword.

The code for this module can be found at Chapter03/2_simple-modular-math-util/calculator.

Let’s compile the code:

javac -d mods --module-source-path . $(find . -name "*.java")

The preceding command has to be executed from Chapter03/2_simple-modular-math-util.

Also, you should have the compiled code from across both the modules, math.util and calculator, in the mods directory. Just a single command and everything including the dependency between the modules is taken care of by the compiler. We didn’t require build tools such as ant to manage the compilation of modules.

The --module-source-path command is the new command-line option for javac, specifying the location of our module source code.

Let’s execute the preceding code:

java --module-path mods -m calculator/com.packt.calculator.Calculator

The --module-path command, similar to --classpath, is the new command-line option  java, specifying the location of the compiled modules.

After running the preceding command, you will see the calculator in action:

Congratulations! With this, we have a simple modular application up and running.

We have provided scripts to test out the code on both Windows and Linux platforms. Please use run.bat for Windows and run.sh for Linux.

How it works

Now that you have been through the example, we will look at how to generalize it so that we can apply the same pattern in all our modules. We followed a particular convention to create the modules:

|application_root_directory
|--module1_root
|----module-info.java
|----com
|------packt
|--------sample
|----------MyClass.java
|--module2_root
|----module-info.java
|----com
|------packt
|--------test
|----------MyAnotherClass.java

We place the module-specific code within its folders with a corresponding module-info.java file at the root of the folder. This way, the code is organized well.

Let’s look into what module-info.java can contain. From the Java language specification (http://cr.openjdk.java.net/~mr/jigsaw/spec/lang-vm.html), a module declaration is of the following form:

{Annotation} [open] module ModuleName { {ModuleStatement} }

Here’s the syntax, explained:

  • {Annotation}: This is any annotation of the form @Annotation(2).
  • open: This keyword is optional. An open module makes all its components accessible at runtime via reflection. However, at compile-time and runtime, only those components that are explicitly exported are accessible.
  • module: This is the keyword used to declare a module.
  • ModuleName: This is the name of the module that is a valid Java identifier with a permissible dot (.) between the identifier names—similar to math.util.
  • {ModuleStatement}: This is a collection of the permissible statements within a module definition. Let’s expand this next.

A module statement is of the following form:

ModuleStatement:
  requires {RequiresModifier} ModuleName ;
  exports PackageName [to ModuleName {, ModuleName}] ;
  opens PackageName [to ModuleName {, ModuleName}] ;
  uses TypeName ;
  provides TypeName with TypeName {, TypeName} ;

The module statement is decoded here:

  • requires: This is used to declare a dependency on a module. {RequiresModifier} can be transitive, static, or both. Transitive means that any module that depends on the given module also implicitly depends on the module that is required by the given module transitively. Static means that the module dependence is mandatory at compile time, but optional at runtime. Some examples are requires math.util, requires transitive math.util, and requires static math.util.
  • exports: This is used to make the given packages accessible to the dependent modules. Optionally, we can force the package’s accessibility to specific modules by specifying the module name, such as exports com.package.math to claculator.
  • opens: This is used to open a specific package. We saw earlier that we can open a module by specifying the open keyword with the module declaration. But this can be less restrictive. So, to make it more restrictive, we can open a specific package for reflective access at runtime by using the opens keyword—opens com.packt.math.
  • uses: This is used to declare a dependency on a service interface that is accessible via java.util.ServiceLoader. The service interface can be in the current module or in any module that the current module depends on.
  • provides: This is used to declare a service interface and provide it with at least one implementation. The service interface can be declared in the current module or in any other dependent module. However, the service implementation must be provided in the same module; otherwise, a compile-time error will occur.

We will look at the uses and provides clauses in more detail in the Using services to create loose coupling between the consumer and provider modules recipe.

The module source of all modules can be compiled at once using the --module-source-path command-line option. This way, all the modules will be compiled and placed in their corresponding directories under the directory provided by the -d option. For example, javac -d mods --module-source-path . $(find . -name "*.java") compiles the code in the current directory into a mods directory.

Running the code is equally simple. We specify the path where all our modules are compiled into using the command-line option --module-path. Then, we mention the module name along with the fully qualified main class name using the command-line option -m, for example, java --module-path mods -m calculator/com.packt.calculator.Calculator.

In this tutorial, we learned to create a simple modular Java application. To learn more Java 11 recipes, check out the book Java 11 Cookbook – Second Edition.

Read next

Brian Goetz on Java futures at FOSDEM 2019

7 things Java programmers need to watch for in 2019

Clojure 1.10 released with Prepl, improved error reporting and Java compatibility

Data science enthusiast. Cycling, music, food, movies. Likes FPS and strategy games.