If you already "migrated" to ILE and free-format RPG, learning a modern programming language is easier than it was before. Why not learn C#?

This article is a first step in the path to C# mastery, from the perspective of an RPG programmer. You're probably wondering why you should learn C#. Well, this modern and very powerful programming language allows you to supplement—or, in some situations, even replace—part of the development you'd do on an IBM i.

Many online forums provide help, either as simple samples of code or complete components that are ready to use just by plugging them into your project and writing a few lines of code. Instead of reinventing the wheel, there are plenty of "wheels" out there for nearly anything you can imagine—from financial calculation to mobile application templates and everything in between.

You might argue the same could be said about Java. It's true, and the C#-versus-Java debate is as old as C# itself. For some tasks, C# is more adequate, while for others Java is a better fit. It might come as a surprise, but there are a lot of similarities between C# and RPG, as I'll show you in this article. Arguably, C# is easier to learn than Java, because C#'s "native" Integrated Development Environment (IDE), Visual Studio, provides rich editing features that go way beyond Java's IDEs and the language syntax is somewhat friendlier than Java's.

Some C#ontext: The .NET Framework

C# is one of the languages you can use in the .NET framework. This framework is, in a lot of ways, similar to IBM i's ILE. Here are a few common traits:

Programming Language Interoperability

They both provide interoperability between programming languages by "translating" code into an intermediary language. This allows programmers to use code written in another programming language, which they might not even be familiar with, in the program they're working on. While an ILE programmer can use code written in RPG, COBOL, C, C++, and CL, a .NET programmer can use code written in Microsoft's Visual Basic, Visual C#, Visual C++, and many other programming languages from various vendors. These include, among others, versions of COBOL, Perl, Eiffel, Python, Pascal, Salford FTN95 (Fortran), SmallTalk, and Dyalog APL tailored for the .NET Framework. A few years ago, even RPG was made available for the .NET framework, with the launch of ASNA's Visual RPG!

This interoperability is possible, in .NET's case, thanks to the Microsoft Intermediate Language (MSIL). This "managed code," as Microsoft calls it, runs on the Common Language Runtime, or CLR. CLR is the equivalent to ILE's heart or, if you're familiar with Java, a Java Virtual Machine.

Modular and Reusable Components

ILE allows you to create modules, programs, and service programs to organize your code. .NET implements the exact same concept with assemblies. And just like ILE, .NET provides a set of common tools available to all the supported languages, which allow the programmer to simplify tasks and save a lot of time. The difference is that .NET's tools go way beyond ILE's Common Runtime Services in their functionality and ease of use. A significant part of this ease of use is due to the fact that .NET's languages are object-oriented (OO). If you're not familiar with the concept, the next section will help you understand it. Learning what OO is and how it works is fundamental to learning C#.

What's This Object-Oriented (OO) Stuff Everybody Is Talking About?

We RPG programmers are used to procedural languages, in which the operations are isolated from the data they manipulate. In an object-oriented language, things are a bit different. To put it simply, data and the operations to manipulate it are parts of the same things, aptly named "objects." These objects intend to mimic real-world objects, thus facilitating the way code is written, stored, and understood. An object is built based upon a class, which is in some ways similar to a service program.

Let's look at an example to try to make this clearer. If you wanted to write a program to simulate the life of a horse, here's how you could do it:

Life of a Horse, RPG Version:

You'd create a file, where each record would represent a horse. This record format would contain fields like name, breed, date_of_birth, mile_ran, and other relevant information. Then you'd write a set of procedures to bring the horse to life: Change_Horse_Name, Set_Horse_Date_of_Birth, Make_Horse_Neigh, Make_Horse_Run, and so on. These procedures would receive the record, or part of it, as parameters and return something, depending on their purpose. Ideally, these procedures would reside in a service program so they could be easily reused.

Life of a Horse, Object-Oriented Version:

You'd define a Horse class (a kind of blueprint of what a horse is) with the data (here named properties or attributes) and the stuff you'd want for the horse (here called methods). In your program, you'd then create a new horse object (a "building" based on the blueprint). To make this new horse object do stuff, you'd just invoke its methods. For instance, to make it neigh, you'd write horse.neigh(), instead of neigh(horse).

The big difference is that while the data and procedures are isolated in the RPG version of things, the object's methods are part of the object. In a way, you can think of classes, with their data and code, like service programs. The thing is that, while RPG limits you to one "copy" of the service program per program, you can create as many "copies" of the object as you want in your program. This means that you can easily handle multiple similar things simultaneously, while keeping them totally independent from one another. You can write very specific code and "hide" it within an object, thus making sure it won't be used with the wrong parameters or in the wrong situation. In other words, by storing the code and the data inside the object, you make crystal clear what belongs where: the Horse's "neigh" method will never be applied to a Dog object, for instance.

Another big difference is the way you can easily derive and extend functionality in an object-oriented language. For instance, in a procedural language like RPG, if you want to build upon the existing functionality of a service program but you don't want the programs that currently use that service program to be affected, you'd have create a new service program, starting with the same modules and adding new ones, to extend the functionality. In an object-oriented language, you can simply create a class derived from an existing class, thus inheriting all of its data and functionality, while keeping a link to the original one. This inheritance operation is very simple to achieve; you just need to specify a keyword and the name of the parent class in the new class definition. You can still create objects based on both old and new classes without any risk. This is, in a brief and over-simplified way, the concept of inheritance. Object-oriented languages have a few additional fundamental concepts, abstraction, polymorphism, and encapsulation, which I won't be discussing at this time because they're a bit too weird for someone used to procedural languages.

Your First C# Program

There's much, much more to know about object-oriented programming languages, but I'm sure what you really want is to see C# in action. Let's dive into your first C# program! Naturally, it is the obligatory "Hello World!" program:

using System;

namespace MyFirstCSharpProgram

{

   class Program

   {

       static void Main(string[] args)

       {

           Console.WriteLine("A big Hello to the C# World from an RPG programmer!");

       }

   }

}

You'll notice a couple of similarities to free-format RPG: most lines of code end with a semicolon, and the methods take parameters just like RPG's procedures and functions. But other than that, it's…well, different. Let's analyze this code one instruction at a time and try to establish a comparison with RPG code.

The first line of code, using System;, tells the compiler that you're going to use or import something from a class named System. Remember, classes are more or less equivalent to service programs, so this can be compared to a /COPY line in which you import the prototypes for your service program's functions and procedures.

The following line defines the namespace MyFirstCSharpProgram. A namespace is a concept that doesn't have a match in RPG. The namespace keyword declares a scope that contains a set of related objects. It's used to organize code elements and to create globally unique types. You can have classes with the same name coexisting in the same project as long as they belong to different namespaces. The usefulness of this keyword will become more obvious when you dig deeper into C#.

Then comes what makes this program work: the definition of the class, with the line class Program, followed by curly braces ({ and }). These are C#'s equivalent to Begin and End. While in an RPG subroutine you'd use BegSr and EndSr to signal the beginning and end of a subroutine, in C# you use curly braces not only for signaling the beginning and end of a class, but for all the blocks of code: namespaces, classes, if statements, loops, everything. Without the proper indentation, this becomes really confusing really quick. Just like in free-format RPG, indentation is of paramount importance in C#.

After the class definition comes a method definition. In this particular case, it's the Main method, which will be the entry point of the program. When you create an ILE program, you need to tell the compiler which module is the entry point—or in other words, where the code is that makes the program start—by running statements and calling procedures or functions from itself or other modules. In C#, this is achieved via the Main method: your project can have dozens of classes, but you'll need to tell the compiler where to begin. For the moment, let's ignore the static and void keywords and focus on the parameter definition of the method: (string[] args) defines an array of strings named "args." This array is represented here by the string keyword followed by the opening and closing square brackets ([ and ]). This is the same as a Char(<some number to represent the length>) Dim(<number of elements>) in an RPG procedure's prototype definition. While in RPG you need to indicate the number of elements, in C# you're allowed to define arrays and most data types without specifying their length. Just as in RPG, you can create an array of nearly all the existing data types. Let's get back to the method parameter definition: you can have multiple parameters, separated by commas and following the pattern <parameter type> <parameter name>. Similar to RPG, C# also has keywords designed to constraint the usage of the parameters, but let's leave that for later.

The next line of code, Console.WriteLine("A big Hello to the C# World from an RPG programmer!");, executes the WriteLine method of the Console object, passing the sentence as parameter or argument, as it's usually referred to in C#. Note that the Console object is not part of the language itself; it's one of the more than 4,000 objects the .NET framework provides to all the programming languages it supports. In order to use it, you need to tell the compiler where it comes from. That's what the first line of code, using System;, is doing there: it's telling the compiler where to look when it finds something that it doesn't recognize as part of the C# programming language. You can think of the using keyword as a hybrid of the /COPY statement and the binding directory, as it provides both the "prototype definition" and the location of a certain object.

The rest of the lines of code simply serve to close all the blocks of code, with properly indented curly braces. That's it, your first C# program! When you compile and run it, you get this output:

Figure 1: Press any key to continue…

Sorry for the Portuguese: it's simply the standard "Press any key to continue…" message.

Real C# programs are much more complex, but they follow a typical structure: they're composed of several classes and can import many others, designed by the programmer or downloaded from the Internet. One of the classes has a Main method, which is executed when the program is called. This is somewhat similar to an ILE program, which can have several modules and "import" others via the respective service programs. When the ILE program is compiled, the programmer indicates which module is the entry point (the equivalent to the Main method) to get the program running.

This simple program failed to illustrate the object creation, which is a core concept in any object-oriented programming language, but I tried to keep things as simple as possible for a first contact. There were several other important aspects that I had to leave out of this article, such as the tools to program in C# (Microsoft did a great job with Visual Studio, but there are other tools on the market that you might want to try) and the nitty-gritties of the language itself: how the instructions and their syntax work, how to handle data, how to build graphical user interfaces, and many other things.

Let me know in the Comments section below if you found this article useful and if you'd like to know more about C# from an RPG programmer's perspective. I might turn this topic into a running TechTip series, like RPG Academy and SQL 101!