0.4 — Introduction to development

Before we can write and execute our first program, we need to understand in more detail how C++ programs get developed. Here is a graphic outlining a simplistic approach:

The software development process

Step 1: Define the problem that you would like to solve.

This is the “what” step, where you figure out what you are going to solve. Coming up with the initial idea for what you would like to program can be the easiest step, or the hardest. But conceptually, it is the simplest. All you need is an idea that can be well defined, and you’re ready for the next step.

Here are a few examples:

  • “I want to write a program that will allow me to enter numbers, then calculates the average.”
  • “I want to write a program that generates a 2d maze and lets the user navigate through it.”
  • “I want to write a program that reads in a file of stock prices and predicts whether the stock will go up or down.”

Step 2: Determine how you are going to solve the problem.

This is the “how” step, where you determine how you are going to solve the problem you came up with in step 1. It is also the step that is most neglected in software development. The crux of the issue is that there are many ways to solve a problem -- however, some of these solutions are good and some of them are bad. Too often, a programmer will get an idea, sit down, and immediately start coding a solution. This often generates a solution that falls into the bad category.

Typically, good solutions have the following characteristics:

  • They are straightforward.
  • They are well documented (especially any assumptions being made).
  • They are built modularly, so parts can be reused or changed later without impacting other parts of the program.
  • They are robust, and can recover or give useful error messages when something unexpected happens.

When you sit down and start coding right away, you’re typically thinking “I want to do _this_”, so you implement the solution that gets you there the fastest. This can lead to programs that are fragile, hard to change or extend, or have lots of bugs.

Studies have shown that only 20% of a programmer’s time is actually spent writing the initial program. The other 80% is spent debugging (fixing errors) or maintaining (adding features to) a program. Consequently, it’s worth your time to spend a little extra time up front before you start coding thinking about the best way to tackle a problem, what assumptions you are making, and how you might plan for the future, in order to save yourself a lot of time and trouble down the road.

We’ll talk more about how to effectively design solutions to problems in a future lesson.

Step 3: Write the program

In order to write the program, we need two things: First we need knowledge of a programming language -- that’s what these tutorials are for! Second, we need an editor. It’s possible to write a program using any editor you want, even something as simple as Window’s notepad or Unix’s vi or pico. However, we strongly urge you to use an editor that is designed for coding. Don’t worry if you don’t have one yet. We’ll cover how to install one soon.

A typical editor designed for coding has a few features that make programming much easier, including:

1) Line numbering. Line numbering is useful when the compiler gives us an error. A typical compiler error will state “error, line 64”. Without an editor that shows line numbers, finding line 64 can be a real hassle.

2) Syntax highlighting and coloring. Syntax highlighting and coloring changes the color of various parts of your program to make it easier to see the overall structure of your program.

3) An unambiguous font. Non-programming fonts often make it hard to distinguish between the number 0 and the letter O, or between the number 1, the letter l (lower case L), and the letter I (upper case i). A good programming font will differentiate these symbols in order to ensure one isn’t accidentally used in place of the other.

Your C++ programs should be called name.cpp, where name is replaced with the name of your program. The .cpp extension tells the compiler (and you) that this is a C++ source code file that contains C++ instructions. Note that some people use the extension .cc instead of .cpp, but we recommend you use .cpp.

Also note that many complex C++ programs have multiple .cpp files. Although most of the programs you will be creating initially will only have a single .cpp file, it is possible to write single programs that have tens if not hundreds of individual .cpp files.

Step 4: Compiling

In order to compile a program, we need a program called a compiler. The job of the compiler is twofold:

1) To check your program and make sure it follows the rules of the C++ language. If it does not, the compiler will give you an error to help pinpoint what needs fixing.

2) To convert each file of source code into a machine language file called an object file. Object files are typically named name.o or name.obj, where name is the same name as the .cpp file it was produced from. If your program had 5 .cpp files, the compiler would generate 5 object files.

The compilation process

For illustrative purposes only, most Linux and MacOS systems come with a C++ compiler called g++. To use g++ to compile a file from the command line, we would do this:

g++ -c file1.cpp file2.cpp file3.cpp

This would create file1.o, file2.o, and file3.o. The -c means “compile only”, which tells g++ to just produce .o (object) files.

Other compilers are available for Linux, Windows, and just about every other system. We will discuss installing a compiler in the next section, so there is no need to do so now.

For complex projects, some development environments use a makefile, which is a file that tells the compiler which files to compile. Makefiles are an advanced topic, and entire books have been written about them. Fortunately, you don’t need to worry about them, so we will not discuss them here.

Step 5: Linking

Linking is the process of taking all the object files generated by the compiler and combining them into a single executable program that you can run. This is done by a program called the linker.

The linking process

In addition to the object files for a program, the linker also includes files from the C++ standard library (or any other precompiled libraries you’re using, such as graphics or sound libraries). The C++ language itself is fairly small and simple. However, it comes with a large library of optional components that may be utilized by your programs, and these components live in the C++ standard library. For example, if you wanted to output something to the screen, your program would include a special command to tell the compiler that you wanted to use the I/O (input/output) routines from the C++ standard library.

Once the linker is finished linking all the object files (assuming all goes well), you will have an executable file.

Again, for illustrative purposes, to link the .o files we created above on a Linux or MacOS machine, we can again use g++:

g++ -o prog file1.o file2.o file3.o

The -o tells g++ that we want an executable file named “prog” that is built from file1.o, file2.o, and file3.o

The compile and link steps can be combined together if desired:

g++ -o prog file1.cpp file2.cpp file3.cpp

Which will combine the compile and link steps together and directly produce an executable file named “prog”.

Step 6: Testing and Debugging

This is the fun part (hopefully)! You are able to run your executable and see whether it produces the output you were expecting. If not, then it’s time for some debugging. We will discuss debugging in more detail soon.

Note that steps 3, 4, 5, and 6 all involve software. While you can use separate programs for each of these functions, a software package known as an integrated development environment (IDE) bundles and integrates all of these features together. With a typical IDE, you get a code editor that does line numbering and syntax highlighting. You get a compiler and a linker. The IDE will automatically generate the parameters necessary to compile and link your program into an executable, even if it includes multiple files. And when you need to debug your program, you can use the integrated debugger. Furthermore, IDE’s typically bundle a number of other helpful editing features, such as integrated help, name completion, hierarchy browsers, and sometimes a version control system.

We will talk more about installing and using IDEs in the next section.

0.5 -- Installing an Integrated Development Environment (IDE)
0.3 -- Introduction to C/C++

121 comments to 0.4 — Introduction to development

  • Can you make "Dive inn" Topics that we can read from? I need a quick tutorial on how to make a declaratoin.

    • Alex

      You can use the site index to find specific topics.

      • Kevin Rommens

        Well. I'm trying to get how to make a declaration from scratch.
        And I don't get it, even if I try my hardest.
        So maybe give people some tips on how to get topics when I read something and I don't get it?

        • Alex

          Find the article on declarations (using the site index), and leave a comment there with what you're trying to do, and what you've tried that didn't work. I'm sure some nice soul will help.

  • vinayak ruhela

    sir i have some doubt in "what and how" section
    if my what section says that " I want to create a program which remove a particular type of files from my pc so what "How section" consist
    please give me some more examples

    • knelse

      Some of the "how" examples can be:
      1. (naive, slow) Find all the files on your PC, then check for those you want to remove and remove them.
      2. (pro-active) Make a system service that would look for file creation events and trigger removal program when specific file type is created.
      3. (somewhat strange) Modify file system code to make that file type work as if it was free space.
      4. (resembling actual solution development) Limit possible folders/locations to reasonable ones (for example, you're probably not looking for .pdf in Windows DLL folders), find out the best time to run such program (for example, if you need to delete temp files that some software creates once a month, you don't need to run removal every second), think about speed and concurrency issues (what happens if you delete the file that your operating system is currently trying to read? Is it a problem for everything else on your PC if you try to delete 1M files simultaneously?), then check for theoretically possible bugs, then implement solution 1 with all the data you collected.

  • I work in a unix environment with Mac OS X.
    How can i execute step 5 if the terminals gives back this message:
    << MacBook-Air-di-Iman:calcoloipotenusa iman.rosignoli$ g++ -o prog 2numeri.o file.o
    duplicate symbol _main in:
    ld: 1 duplicate symbol for architecture x86_64
    clang: error: linker command failed with exit code 1 (use -v to see invocation) >>
    In syntesis, there's this message of "duplicate symbol _main"
    Can you explain it to me in order to fix the files to execute step 5?
    Good day.

  • Oneslayed

    Is it possible to create your own coding engine? (e.g. Java, Python, C++, etc. etc.)

    • Alex

      What's a "coding engine"?

      If you mean programming language, sure. You can write your own language, develop a parser that converts it into C++ (or some other language), and then use an existing compiler for that language to compile it.

      If you mean compiler, yes, but this is a lot more complicated. I wouldn't even try to do this.

  • TheMightyChuy!

    Hi there Mr. Alex, first of all, many thanks for creating these lessons. I have faith that they will turn out very useful (i.e. They will actually work). However, I have a yearning question. Why shouldn't we use .cc as a file extension? What is the difference between .cc and .cpp. Please tell us, we are curious minds.

    *raises right eyebrow and gives sinister smile

    • Alex

      Because the website is named, not

      I kid, I kid.

      There's really no difference. If you want to use .cc, you can. Just be consistent.

  • Matthew

    I started learning, or trying to learn c++ like a week ago and I started with this one site, that I won't name but it really sucked. Anyways so far I'm enjoying this one but before I start learning to code in c++ do I need to learn HTML first? Also is using ideone just as good as an IDE, or would you prefer the IDE?

    • nascardriver

      Hi Metthew!
      HTML has nothing at all to do with c++, you can start with c++.
      I have no experiences with ideone, judging by a quick glance it doesn't seem to let you choose the compiler or compiler options for your program and I don't see any file management. I'd go with a regular IDE.

      Linux: eclipse or JetBrains CLion, I prefer eclipse
      Windows: Visual Studio

      There are many other IDEs around but these are the major ones which you'll find the most help for should you encounter any problems.

      • Matthew

        Yea I actually went with the free visual studio 2017 and so far I love it. Thanks to whoever started this site and took their time to actually teach beginners on how to step by step. It means a lot and I'll learn way more here with it being hands-on than I ever would've in some classroom. I just hope I'll be able to learn enough to get a job with it...This is something I've been wanting to do for a long time and I now have an opportunity to do it...

    • Alex

      You definitely don't need to know HTML.
      Ideone is an online compiler, and looks pretty limited. It will work for simple programs, may not be able to support multiple files or robust debugging. I'd install an IDE application if that's an option for you.

  • Sir!

    I wanted to know if I could program in C++ without using any IDE.

    Best regards.

  • Dear Sirs!

    Sub: Tutorial on C++

    Re: Lesson 0.4 -Step: 3

    I have learnt HTML and CSS online from where they have said(and I also do opine same):

    What is needed?

    Most likely you already have everything you need.

    You have a "browser." A browser is the program that makes it possible to browse and open websites. Right now you are looking at this page in your browser.

    It is not important which browser you use. The most common are Google Chrome, Firefox, and Internet Explorer. But there are others such as Opera and Safari and they can all be used and they are all free.

    You might have heard about, or even used, programs such as Microsoft FrontPage, Macromedia Dreamweaver or even Microsoft Word, which can — or claim that they can — create websites for you. Forget these programs for now! They are not of any help to you when learning how to code your own website.

    Instead, you need a simple text editor. If you are using Windows you can use Notepad.

    The problem with many of the programs that claim they can create websites is that they have a lot of standard functions that you can choose from. The downside is that everything needs to fit into these standard functions. Thus, this type of programs often cannot create a website exactly as you want it. Or — even more annoying — they make changes to your hand-written code. With Notepad or other simple text editors, you only have yourself to thank for your successes and errors.

    I started learning from your tutorial today only with great enthusiasm and liking it and enjoying it; because I myself is the proof of the system(way of learning anything).

    Still I would appreciate your comment in the context.

    Thank you.

    Best regards.
    [N.B.: I have given my website (hosted for FREE) which i have done myself based on their tutorial. I have seen that you also offer HTML Tutorial which I will learn to enrich my knowledge further.]

    • Alex

      I'm not sure what you would like me to comment on.

      With websites, there are many programs that exist to help people layout website visually without having to write HTML and CSS. These programs generate the code for you. However, if you want the best level of control, you need to write the HTML and CSS yourself.

      The C++ core language doesn't include visual elements. So in these tutorials, you'll write all of your own code.

  • Would it be advisable to try designing the solution by writing in a different language (e.g. Python) than what you intend to use for the ACTUAL solution, so you can see how it completely fails?

    • Alex

      That sounds like a lot of work for a questionable benefit. So no, I don't think I'd advise that.

      If you're interested in making sure your solutions are robust and error-free, I'd focus your time on learning how to _really_ test your code well. Lesson 5.11 has some tips in this regard.

Leave a Comment

Put all code inside code tags: [code]your code here[/code]