Like any human language, C++ provides a way to express concepts. If successful, this medium of expression will be significantly easier and more flexible than the alternatives as problems grow larger and more complex.

You can’t just look at C++ as a collection of features; some of the features make no sense in isolation. You can only use the sum of the parts if you are thinking about design, not simply coding. And to understand C++ this way, you must understand the problems with C and with programming in general. This book discusses programming problems, why they are problems, and the approach C++ has taken to solve such problems. Thus, the set of features I explain in each chapter will be based on the way that I see a particular type of problem being solved with the language. In this way I hope to move you, a little at a time, from understanding C to the point where the C++ mindset becomes your native tongue.

Throughout, I’ll be taking the attitude that you want to build a model in your head that allows you to understand the language all the way down to the bare metal; if you encounter a puzzle, you’ll be able to feed it to your model and deduce the answer. I will try to convey to you the insights that have rearranged my brain to make me start “thinking in C++.”

This book is a thorough rewrite of the first edition to reflect all of the changes introduced in C++ by the finalization of the C++ Standard, and also to reflect what I’ve learned since writing the first edition. The entire text present in the first edition has been examined and rewritten, sometimes removing old examples, often changing existing examples and adding new ones, and adding many new exercises. Significant rearrangement and re-ordering of the material took place to reflect the availability of better tools and my improved understanding of how people learn C++. A new chapter was added which is a rapid introduction to the C concepts and basic C++ features for those who don’t have the C background to tackle the rest of the book. The CD ROM bound into the back of the book contains a seminar that is an even gentler introduction to the C concepts necessary to understand C++ (or Java). It was created by Chuck Allison for my company (MindView, Inc.), and it’s called “Thinking in C: Foundations for Java and C++.” It introduces you to the aspects of C that are necessary for you to move on to C++ or Java, leaving out the nasty bits that C programmers must deal with on a day-to-day basis but that the C++ and Java languages steer you away from (or even eliminate, in the case of Java).

So the short answer to the question “what’s different in the 2^nd edition?” is: what isn’t brand new has been rewritten, sometimes to the point where you wouldn’t recognize the original examples and material.

The completion of the C++ Standard also added a number of important new libraries, such as string and the containers and algorithms in the Standard C++ Library, as well as new complexity in templates. These and other more advanced topics have been relegated to Volume 2 of this book, including issues such as multiple inheritance, exception handling, design patterns, and topics about building and debugging stable systems.

Just like the book you currently hold, Thinking in C++, Volume 2 is downloadable in its entirety from my Web site at www.BruceEckel.com. You can find information on the Web site about the expected print date of Volume 2.

The Web site also contains the source code for both of the books, along with updates and information about other seminars-on-CD ROM that MindView, Inc. offers, public seminars, and in-house training, consulting, mentoring, and walkthroughs.

In the first edition of this book, I decided to assume that someone else had taught you C and that you have at least a reading level of comfort with it. My primary focus was on simplifying what I found difficult: the C++ language. In this edition I have added a chapter that is a rapid introduction to C, along with the Thinking in C seminar-on-CD, but I am still assuming that you already have some kind of programming experience. In addition, just as you learn many new words intuitively by seeing them in context in a novel, it’s possible to learn a great deal about C from the context in which it is used in the rest of the book.

I clawed my way into C++ from exactly the same position I expect many of the readers of this book are in: as a programmer with a very no-nonsense, nuts-and-bolts attitude about programming. Worse, my background and experience was in hardware-level embedded programming, in which C has often been considered a high-level language and an inefficient overkill for pushing bits around. I discovered later that I wasn’t even a very good C programmer, hiding my ignorance of structures, malloc( ) and free( ), setjmp( ) and longjmp( ), and other “sophisticated” concepts, scuttling away in shame when the subjects came up in conversation instead of reaching out for new knowledge.

When I began my struggle to understand C++, the only decent book was Bjarne Stroustrup’s self-professed “expert’s guide,[1]” so I was left to simplify the basic concepts on my own. This resulted in my first C++ book,[2] which was essentially a brain dump of my experience. That was designed as a reader’s guide to bring programmers into C and C++ at the same time. Both editions[3] of the book garnered enthusiastic response.

At about the same time that Using C++ came out, I began teaching the language in seminars and presentations. Teaching C++ (and later, Java) became my profession; I’ve seen nodding heads, blank faces, and puzzled expressions in audiences all over the world since 1989. As I began giving in-house training to smaller groups of people, I discovered something during the exercises. Even those people who were smiling and nodding were confused about many issues. I found out, by creating and chairing the C++ and Java tracks at the Software Development Conference for many years, that I and other speakers tended to give the typical audience too many topics, too fast. So eventually, through both variety in the audience level and the way that I presented the material, I would end up losing some portion of the audience. Maybe it’s asking too much, but because I am one of those people resistant to traditional lecturing (and for most people, I believe, such resistance results from boredom), I wanted to try to keep everyone up to speed.

For a time, I was creating a number of different presentations in fairly short order. Thus, I ended up learning by experiment and iteration (a technique that also works well in C++ program design). Eventually I developed a course using everything I had learned from my teaching experience. It tackles the learning problem in discrete, easy-to-digest steps and for a hands-on seminar (the ideal learning situation) there are exercises following each of the presentations. You can find out about my public seminars at www.BruceEckel.com, and you can also learn about the seminars that I’ve turned into CD ROMs.

The first edition of this book developed over the course of two years, and the material in this book has been road-tested in many forms in many different seminars. The feedback that I’ve gotten from each seminar has helped me change and refocus the material until I feel it works well as a teaching medium. But it isn’t just a seminar handout; I tried to pack as much information as I could within these pages, and structure it to draw you through onto the next subject. More than anything, the book is designed to serve the solitary reader who is struggling with a new programming language.

My goals in this book are to:

C++ is a language in which new and different features are built on top of an existing syntax. (Because of this, it is referred to as a hybrid object-oriented programming language.) As more people pass through the learning curve, we’ve begun to get a feel for the way programmers move through the stages of the C++ language features. Because it appears to be the natural progression of the procedurally-trained mind, I decided to understand and follow this same path and accelerate the process by posing and answering the questions that came to me as I learned the language and those questions that came from audiences as I taught the language.

This course was designed with one thing in mind: to streamline the process of learning C++. Audience feedback helped me understand which parts were difficult and needed extra illumination. In the areas in which I got ambitious and included too many features all at once, I came to know – through the process of presenting the material – that if you include a lot of new features, you have to explain them all, and the student’s confusion is easily compounded. As a result, I’ve taken a great deal of trouble to introduce the features as few at a time as possible; ideally, only one major concept at a time per chapter.

The goal, then, is for each chapter to teach a single concept, or a small group of associated concepts, in such a way that no additional features are relied upon. That way you can digest each piece in the context of your current knowledge before moving on. To accomplish this, I leave some C features in place for longer than I would prefer. The benefit is that you will not be confused by seeing all the C++ features used before they are explained, so your introduction to the language will be gentle and will mirror the way you will assimilate the features if left to your own devices.

Here is a brief description of the chapters contained in this book:

Chapter 1: Introduction to Objects. When projects became too big and complicated to easily maintain, the “software crisis” was born, with programmers saying, “We can’t get projects done, and if we can, they’re too expensive!” This precipitated a number of responses, which are discussed in this chapter along with the ideas of object-oriented programming (OOP) and how it attempts to solve the software crisis. The chapter walks you through the basic concepts and features of OOP and also introduces the analysis and design process. In addition, you’ll learn about the benefits and concerns of adopting the language and suggestions for moving into the world of C++.

Chapter 2: Making and Using Objects. This chapter explains the process of building programs using compilers and libraries. It introduces the first C++ program in the book and shows how programs are constructed and compiled. Then some of the basic libraries of objects available in Standard C++ are introduced. By the time you finish this chapter you’ll have a good grasp of what it means to write a C++ program using off-the-shelf object libraries.

Chapter 3: The C in C++. This chapter is a dense overview of the features in C that are used in C++, as well as a number of basic features that are available only in C++. It also introduces the “make” utility that’s common in the software development world and that is used to build all the examples in this book (the source code for the book, which is available at www.BruceEckel.com, contains makefiles for each chapter). Chapter 3 assumes that you have a solid grounding in some procedural programming language like Pascal, C, or even some flavors of Basic (as long as you’ve written plenty of code in that language, especially functions). If you find this chapter a bit too much, you should first go through the Thinking in C seminar on the CD that’s bound with this book (and also available at www.BruceEckel.com).

Chapter 4: Data Abstraction. Most features in C++ revolve around the ability to create new data types. Not only does this provide superior code organization, but it lays the groundwork for more powerful OOP abilities. You’ll see how this idea is facilitated by the simple act of putting functions inside structures, the details of how to do it, and what kind of code it creates. You’ll also learn the best way to organize your code into header files and implementation files.

Chapter 5: Hiding the Implementation. You can decide that some of the data and functions in your structure are unavailable to the user of the new type by making them private. This means that you can separate the underlying implementation from the interface that the client programmer sees, and thus allow that implementation to be easily changed without affecting client code. The keyword class is also introduced as a fancier way to describe a new data type, and the meaning of the word “object” is demystified (it’s a fancy variable).

Chapter 6: Initialization and Cleanup. One of the most common C errors results from uninitialized variables. The constructor in C++ allows you to guarantee that variables of your new data type (“objects of your class”) will always be initialized properly. If your objects also require some sort of cleanup, you can guarantee that this cleanup will always happen with the C++ destructor.

Chapter 7: Function Overloading and Default Arguments. C++ is intended to help you build big, complex projects. While doing this, you may bring in multiple libraries that use the same function name, and you may also choose to use the same name with different meanings within a single library. C++ makes this easy with function overloading, which allows you to reuse the same function name as long as the argument lists are different. Default arguments allow you to call the same function in different ways by automatically providing default values for some of your arguments.

Chapter 8: Constants. This chapter covers the const and volatile keywords, which have additional meaning in C++, especially inside classes. You’ll learn what it means to apply const to a pointer definition. The chapter also shows how the meaning of const varies when used inside and outside of classes and how to create compile-time constants inside classes.

Chapter 9: Inline Functions. Preprocessor macros eliminate function call overhead, but the preprocessor also eliminates valuable C++ type checking. The inline function gives you all the benefits of a preprocessor macro plus all of the benefits of a real function call. This chapter thoroughly explores the implementation and use of inline functions.

Chapter 10: Name Control. Creating names is a fundamental activity in programming, and when a project gets large, the number of names can be overwhelming. C++ allows you a great deal of control over names in terms of their creation, visibility, placement of storage, and linkage. This chapter shows how names are controlled in C++ using two techniques. First, the static keyword is used to control visibility and linkage, and its special meaning with classes is explored. A far more useful technique for controlling names at the global scope is C++’s namespace feature, which allows you to break up the global name space into distinct regions.

Chapter 11: References and the Copy-Constructor. C++ pointers work like C pointers with the additional benefit of stronger C++ type checking. C++ also provides an additional way to handle addresses: from Algol and Pascal, C++ lifts the reference, which lets the compiler handle the address manipulation while you use ordinary notation. You’ll also meet the copy-constructor, which controls the way objects are passed into and out of functions by value. Finally, the C++ pointer-to-member is illuminated.

Chapter 12: Operator Overloading. This feature is sometimes called “syntactic sugar;” it lets you sweeten the syntax for using your type by allowing operators as well as function calls. In this chapter you’ll learn that operator overloading is just a different type of function call and you’ll learn how to write your own, dealing with the sometimes-confusing uses of arguments, return types, and the decision of whether to make an operator a member or friend.

Chapter 13: Dynamic Object Creation. How many planes will an air-traffic system need to manage? How many shapes will a CAD system require? In the general programming problem, you can’t know the quantity, lifetime, or type of objects needed by your running program. In this chapter, you’ll learn how C++’s new and delete elegantly solve this problem by safely creating objects on the heap. You’ll also see how new and delete can be overloaded in a variety of ways so you can control how storage is allocated and released.

Chapter 14: Inheritance and Composition. Data abstraction allows you to create new types from scratch, but with composition and inheritance, you can create new types from existing types. With composition, you assemble a new type using other types as pieces, and with inheritance, you create a more specific version of an existing type. In this chapter you’ll learn the syntax, how to redefine functions, and the importance of construction and destruction for inheritance and composition.

Chapter 15: Polymorphism and virtual Functions. On your own, you might take nine months to discover and understand this cornerstone of OOP. Through small, simple examples, you’ll see how to create a family of types with inheritance and manipulate objects in that family through their common base class. The virtual keyword allows you to treat all objects in this family generically, which means that the bulk of your code doesn’t rely on specific type information. This makes your programs extensible, so building programs and code maintenance is easier and cheaper.

Chapter 16: Introduction to Templates. Inheritance and composition allow you to reuse object code, but that doesn’t solve all of your reuse needs. Templates allow you to reuse source code by providing the compiler with a way to substitute type names in the body of a class or function. This supports the use of container class libraries, which are important tools for the rapid, robust development of object-oriented programs (The Standard C++ Library includes a significant library of container classes). This chapter gives you a thorough grounding in this essential subject.

Additional topics (and more advanced subjects) are available in Volume 2 of this book, which can be downloaded from the Web site www.BruceEckel.com.

I’ve discovered that exercises are exceptionally useful during a seminar to complete a student’s understanding, so you’ll find a set at the end of each chapter. The number of exercises has been greatly increased over the number in the first edition.

Many of the exercises are fairly simple so that they can be finished in a reasonable amount of time in a classroom situation or lab section while the instructor observes, making sure all students are absorbing the material. Some exercises are a bit more challenging to keep advanced students entertained. The bulk of the exercises are designed to be solved in a short time and are intended only to test and polish your knowledge rather than present major challenges (presumably, you’ll find those on your own – or more likely, they’ll find you).

Solutions to selected exercises can be found in the electronic document The Thinking in C++ Annotated Solution Guide, available for a small fee from www.BruceEckel.com.

The source code for this book is copyrighted freeware, distributed via the Web site www.BruceEckel.com. The copyright prevents you from republishing the code in print media without permission, but you are granted the right to use it in many other situations (see below).

The code is available in a zipped file, designed to be extracted for any platform that has a “zip” utility (Most do; you can search the Internet to find a version for your platform if you don’t already have one installed). In the starting directory where you unpacked the code you will find the following copyright notice:

You may use the code in your projects and in the classroom as long as the copyright notice is retained.

Throughout this book, when referring to conformance to the ISO C standard, I will generally just say ‘C.’ Only if it is necessary to distinguish between Standard C and older, pre-Standard versions of C will I make a distinction.

At this writing the C++ Standards Committee was finished working on the language. Thus, I will use the term Standard C++ to refer to the standardized language. If I simply refer to C++ you should assume I mean “Standard C++.”

There is some confusion over the actual name of the C++ Standards Committee and the name of the standard itself. Steve Clamage, the committee chair, clarified this:

There are two C++ standardization committees: The NCITS (formerly X3) J16 committee and the ISO JTC1/SC22/WG14 committee. ANSI charters NCITS to create technical committees for developing American national standards.

J16 was chartered in 1989 to create an American standard for C++. In about 1991 WG14 was chartered to create an international standard. The J16 project was converted to a "Type I" (International) project and subordinated to the ISO standardization effort.

The two committees meet at the same time at the same location, and the J16 vote constitutes the American vote on WG14. WG14 delegates technical work to J16. WG14 votes on the technical work of J16.

The C++ standard was originally created as an ISO standard. ANSI later voted (as recommended by J16) to adopt the ISO C++ standard as the American standard for C++.

Thus, ‘ISO’ is the correct way to refer to the C++ Standard.

Your compiler may not support all of the features discussed in this book, especially if you don’t have the newest version of the compiler. Implementing a language like C++ is a Herculean task, and you can expect that the features will appear in pieces rather than all at once. But if you attempt one of the examples in the book and get a lot of errors from the compiler, it’s not necessarily a bug in the code or the compiler; it may simply not be implemented in your particular compiler yet.

The primary content of the CD ROM packaged in the back of this book is a “seminar on CD ROM” titled Thinking in C: Foundations for Java & C++ by Chuck Allison (published by MindView, Inc., and also available in quantities at www.BruceEckel.com). This contains many hours of audio lectures and slides, and can be viewed on most computers if you have a CD ROM player and a sound system.

The goal of Thinking in C is to take you carefully through the fundamentals of the C language. It focuses on the knowledge necessary for you to be able to move on to the C++ or Java languages instead of trying to make you an expert in all the dark corners of C. (One of the reasons for using a higher-level language like C++ or Java is precisely so we can avoid many of these dark corners.) It also contains exercises and guided solutions. Keep in mind that because Chapter 3 of this book goes beyond the Thinking in C CD, the CD is not a replacement for that chapter, but should be used instead as a preparation for this book.

Please note that the CD ROM is browser-based, so you should have a Web browser installed on your machine before using it.

There are seminars-on-CD-ROM planned to cover Volume 1 and Volume 2 of this book. These comprise many hours of audio lectures by me that accompany slides that cover selected material from each chapter in the book. They can be viewed on most computers if you have a CD ROM player and a sound system. These CDs may be purchased at www.BruceEckel.com, where you will find more information and sample lectures.

My company, MindView, Inc., provides public hands-on training seminars based on the material in this book and also on advanced topics. Selected material from each chapter represents a lesson, which is followed by a monitored exercise period so each student receives personal attention. We also provide on-site training, consulting, mentoring, and design and code walkthroughs. Information and sign-up forms for upcoming seminars and other contact information can be found at www.BruceEckel.com.

I am sometimes available for design consulting, project evaluation and code walkthroughs. When I first began writing about computers, my primary motivation was to increase my consulting activities, because I find consulting to be challenging, educational, and one of my most enjoyable experiences, professionally. Thus I will try my best to fit you into my schedule, or to provide you with one of my associates (who are people that I know well and trust, and often people who co-develop and teach seminars with me).

No matter how many tricks a writer uses to detect errors, some always creep in and these often leap off the page to a fresh reader. If you discover anything you believe to be an error, please use the correction form you will find at www.BruceEckel.com. Your help is appreciated.

The first edition of this book had my face on the cover, but I originally wanted a cover for the second edition that was more of a work of art like the Thinking in Java cover. For some reason, C++ seems to me to suggest Art Deco with its simple curves and brushed chrome. I had in mind something like those posters of ships and airplanes with the long sweeping bodies.

My friend Daniel Will-Harris, (www.Will-Harris.com) whom I first met in junior high school choir class, went on to become a world-class designer and writer. He has done virtually all of my designs, including the cover for the first edition of this book. During the cover design process, Daniel, unsatisfied with the progress we were making, kept asking “How does this relate people to computers?” We were stuck.

On a whim, with no particular outcome in mind, he asked me to put my face on the scanner. Daniel had one of his graphics programs (Corel Xara, his favorite) “autotrace” the scan of my face. As he describes it, “Autotracing is the computer's way to turn a picture into the kinds of lines and curves it really likes.” Then he played with it until he had something that looked like a topographic map of my face, an image that might be the way a computer could see people.

I took this image and photocopied it onto watercolor paper (some color copiers can handle thick stock), and then started creating lots of experiments by adding watercolor to the image. We selected the ones we liked best, then Daniel scanned them back in and arranged them into the cover, adding the text and other design elements. The whole process happened over several months, mostly because of the time it took me to do the watercolors. But I’ve especially enjoyed it because I got to participate in the art on the cover, and because it gave me incentive to do more watercolors (what they say about practice really is true).

The book’s interior design was created by Daniel Will-Harris, who used to play with rub-on letters in junior high school while he awaited the invention of computers and desktop publishing. However, I produced the camera-ready pages myself, so the typesetting errors are mine. Microsoft^® Word for Windows Versions 8 and 9 were used to write the book and to create camera-ready pages, including generating the table of contents and index. (I created a COM automation server in Python, called from Word VBA macros, to aid me in index marking.) Python (see www.Python.org) was used to create some of the tools for checking the code, and would have been use for the code extraction tool had I discovered it earlier.

I created the diagrams using Visio^® – thanks to Visio Corporation for creating a useful tool.

The body typeface is Georgia and the headlines are in Verdana. The final camera-ready version was produced in Adobe^® Acrobat 4 and taken directly to press from that file – thanks very much to Adobe for creating a tool that allows e-mailing camera-ready documents, as it enables multiple revisions to be made in a single day rather than relying on my laser printer and overnight express services. (We first tried the Acrobat process with Thinking in Java, and I was able to upload the final version of that book to the printer in the U.S. from South Africa.)

The HTML version was created by exporting the Word document to RTF, then using RTF2HTML (see http://www.sunpack.com/RTF/) to do most of the work of the HTML conversion. (Thanks to Chris Hector for making such a useful, and especially reliable, tool.) The resulting files were cleaned up using a custom Python program that I hacked together, and the WMFs were converted to GIFs using JASC^® PaintShop Pro 6 and its batch conversion tool (thanks to JASC for solving so many problems for me with their excellent product). The color syntax highlighting was added via a Perl script kindly contributed by Zafir Anjum.

First, thanks to everyone on the Internet who submitted corrections and suggestions; you’ve been tremendously helpful in improving the quality of this book, and I couldn’t have done it without you. Special thanks to John Cook.

The ideas and understanding in this book have come from many sources: friends like Chuck Allison, Andrea Provaglio, Dan Saks, Scott Meyers, Charles Petzold, and Michael Wilk; pioneers of the language like Bjarne Stroustrup, Andrew Koenig, and Rob Murray; members of the C++ Standards Committee like Nathan Myers (who was particularly helpful and generous with his insights), Bill Plauger, Reg Charney, Tom Penello, Tom Plum, Sam Druker, and Uwe Steinmueller; people who have spoken in my C++ track at the Software Development Conference; and often students in my seminars, who ask the questions I need to hear in order to make the material more clear.

A huge thank-you to my friend Gen Kiyooka, whose company Digigami has provided me with a web server.

My friend Richard Hale Shaw and I have taught C++ together; Richard’s insights and support have been very helpful (and Kim’s, too). Thanks also to KoAnn Vikoren, Eric Faurot, Jennifer Jessup, Tara Arrowood, Marco Pardi, Nicole Freeman, Barbara Hanscome, Regina Ridley, Alex Dunne, and the rest of the cast and crew at MFI.

A special thanks to all my teachers and all my students (who are my teachers as well).

And for favorite writers, my deep appreciation and sympathy for your efforts: John Irving, Neal Stephenson, Robertson Davies (we shall miss you), Tom Robbins, William Gibson, Richard Bach, Carlos Castaneda, and Gene Wolfe.

To Guido van Rossum, for inventing Python and giving it selflessly to the world. You have enriched my life with your contribution.

Thanks to the people at Prentice Hall: Alan Apt, Ana Terry, Scott Disanno, Toni Holm, and my electronic copy-editor Stephanie English. In marketing, Bryan Gambrel and Jennie Burger.

Sonda Donovan helped with the production of the CD Rom. Daniel Will-Harris (of course) created the silkscreen design that’s on the Disc itself.

To all the great folks in Crested Butte, thanks for making it a magical place, especially Al Smith (creator of the wonderful Camp4 Coffee Garden), my neighbors Dave & Erika, Marsha at Heg’s Place bookstore, Pat & John at the Teocalli Tamale, Sam at the Bakery Café, and Tiller for his help with audio research. And to all the terrific people that hang out at Camp4 in and make my mornings interesting.

The supporting cast of friends includes, but is not limited to, Zack Urlocker, Andrew Binstock, Neil Rubenking, Kraig Brockschmidt, Steve Sinofsky, JD Hildebrandt, Brian McElhinney, Brinkley Barr, Larry O’Brien, Bill Gates at Midnight Engineering Magazine, Larry Constantine, Lucy Lockwood, Tom Keffer, Dan Putterman, Gene Wang, Dave Mayer, David Intersimone, Claire Sawyers, the Italians (Andrea Provaglio, Rossella Gioia, Laura Fallai, Marco & Lella Cantu, Corrado, Ilsa and Christina Giustozzi), Chris and Laura Strand (and Parker), the Almquists, Brad Jerbic, Marilyn Cvitanic, the Mabrys, the Haflingers, the Pollocks, Peter Vinci, the Robbins, the Moelters, Dave Stoner, Laurie Adams, the Cranstons, Larry Fogg, Mike and Karen Sequeira, Gary Entsminger and Allison Brody, Kevin, Sonda, & Ella Donovan, Chester and Shannon Andersen, Joe Lordi, Dave and Brenda Bartlett, the Rentschlers, Lynn and Todd, and their families. And of course, Mom and Dad.

[1] Bjarne Stroustrup, The C++ Programming Language, Addison-Wesley, 1986 (first edition).

[2] Using C++, Osborne/McGraw-Hill 1989.

[3] Using C++ and C++ Inside & Out, Osborne/McGraw-Hill 1993.