robotfindskitten is a very fun adventure game for robots and humans. There are many versions of the game, but this one is mine.

Part of my intent was for it to be a decent learning vehicle for people new to the C programming language.

I based this version on a version by Alexey Toptygin, but simplified it and re-cast it in a modern style. It also sheds all vestiges of the GNU Autotools build system. I noticed that the build system boilerplate and infrastructure was 2 orders of magnitude larger than the code itself. That seemed wrong.

By perusing the Git history, you can see the steps I followed to get from Alexey’s original to this version. Thank you, Alexey! And thank you to Leonard Richardson, who originated robotfindskitten.

This version of robotfindskitten works on POSIX operating systems (tested on Ubuntu and macOS) with reasonably modern C compilers (tested with Clang).

You will need a C compiler. On macOS, get Xcode. On Ubuntu and other Debian-based systems, run sudo apt install build-essential. That should do the trick.

The ncurses library is a C code library for manipulating the display in Terminal windows, and robotfindskitten depends on it. robotfindskitten will not compile properly unless your system has the ncurses header files. They come with Xcode on macOS, and you can install them on Ubuntu or other Debian-based systems with the command sudo apt install libncurses-dev.

To build and run robotfindskitten, open a Terminal window and simply type make. To run, type ./robotfindskitten. Like so:

git clone https://github.com/noncombatant/robotfindskitten.git
cd robotfindskitten
make
./robotfindskitten

To install, do the preceding and then copy the robotfindskitten binary to a place in your $PATH. A good place might be $HOME/bin, /usr/games, or /usr/local/games.

In my experience, a good way to learn how to learn a programming language is to interleave the activities of reading tutorial books, reading code written in the language, and writing code in the language. The best first book on programming in C is probably The C Programming Language by Kernighan and Ritchie. (It is also one of the best technical books of any kind!) This implementation of robotfindskitten is intended to be as easy to read as C can be. (Let me know if I succeeded or failed!) So, that covers the reading part of learning.

For the writing part of learning, try your hand at modifying robotfindskitten to change the game play in ways you find fun. Here are some ideas:

• Add some more non-kitten item descriptions
• Add functionality to read additional non-kitten item descriptions from a filename provided on the command line
• Add command line options to set the size of the game field (currently, the game uses the entire screen)
• Develop a scoring system, and display the score on the screen (e.g., as Robot sees more non-kitten items, its Zen Wisdom increases)
• Enable Kitten and the non-kitten items to move about on the game field, rather than staying stationary
• ...whatever else seems fun to you!

I assume that most people reading this are coming from a background programming in a very high-level language like Python, JavaScript, or Ruby. C is significantly different than those languages in several ways:

• Before you can run a program written in C, you must first compile (translate) the C code into code that your computer can understand.
• Each variable must be declared to refer to a value of a particular type, such as char (character), int (integer), double (floating-point number), and so on.
• C distinguishes between values of particular types and pointers to values. Pointers are declared with the * symbol; int foo; means “foo is an integer”, while int* bar; means “bar is a pointer to an integer”.
• The unary operator & (address-of) gets you a pointer to a value. Given an int foo, the expression &foo evaluates to the address of foo, and the type of this expression is int*.
• The unary operator * (dereference) follows an address to get the value it refers to. Given an int* wump, *wump evaluates to whatever int value wump is currently pointing to.
• You can declare arrays, but the size of arrays is fixed at compile time (!). char noodles[12]; means “noodles is an array of 12 characters”. int flarb[] = { 1, 2, 3 }; means “flarb is an array of integers, and it is large enough to hold the following items: ...”.
• Declarations may also come with various type qualifiers, such as static, const, and so on. At first, you can mostly ignore them. Their meaning will become more apparent (and more important) as you get more experience with C.

The above is the bare minimum background you’ll need to start skimming C code. I’ve left a lot out, for the sake of brevity and to avoid repeating material you can learn from more authoritative sources like Kernighan and Ritchie.

The biggest immediate difference between C and (e.g.) JavaScript is that the programmer must declare the names and types of variables before using them, and that there are both value types and pointer types.

Most of the rest of C grammar will probably look familiar to JavaScript programmers, and somewhat familiar to Python programmers.

You probably won’t understand everything right away, and that is OK. Like Robot, you will find your Kitten with patience and in due time.

So much for the grammar of C. What about the vocabulary? Like all programming languages, C has a standard library of functions and data types that programmers can use and build on. The standard library is documented in Unix’ manual pages (or “man pages” for short).

On macOS with Xcode, the manual pages are part of the Xcode install; on Ubuntu and other Debian-based systems, you may have to install a package to get them: sudo apt install manpages-dev.

To read the manual page for a given function, open a Terminal window and type man 3 foo at the shell prompt, where foo is the name of a C library function you want to read about. To page up and down, use the arrow keys, b, f, the Space Bar, and so on. To quit reading the manual page, press q to quit.

The 3 in man 3 foo refers to the chapter of the Unix manual: chapter 3 is for C library functions. Chapter 1 is for command-line programs, and Chapter 2 is for kernel system calls. (See man 1 man for more information.)

A good manual page to read is man 3 getopt. getopt is a C library function that parses the command line, and on most systems the manual page includes an example of how to use it. Also read use of getopt in robotfindskitten.c.

Here are some more good manual pages to read (all in chapter 3), that might help you when implementing changes to robotfindskitten:

• calloc (and free), for creating (and destroying) dynamically-sized arrays
• fopen, fread, fwrite, fclose, and getline, for working with files

You’ll notice when reading the code that identifiers (variable names, function names, type names, et c.) seem to come from multiple naming conventions. That is because they do. The difference is that ‘legacy’ code (such as the C standard library, and the ncurses library) comes from the 1970s and 1980s, when memory was expensive and text editors did not have fancy features like autocomplete. Thus, you get these all-lowercase and pathologically terse names like errno (error number), printf (print formatted string), and mvaddch (move and add character). Memory was so expensive back then that early C compilers were not required to recognize more than the first 6 characters of an identifier, so all identifiers in a program had to be unique in their first 6 characters. Bonkers!

Now that computers are cheap and text editors are good, we can afford to write code that looks like it means something. Hence the ‘modern’ naming conventions used in this code:

• GlobalName (including functions like GetRandomIcon and InitializeGame)
• local_name (including struct members and function-local variables)
• i and j for loop iterator variables

“Writing Programs With NCURSES” by Raymond and Ben-Halim describes the ncurses library and how to use it, in depth.

The Practice Of Programming by Kernighan and Pike is an excellent discussion of ‘modern’ programming practices.

Managing Projects With make (2nd edition) by Oram and Talbott describes the workings of the Makefile and the make program.

lex & yacc (2nd edition) by Brown, Levine, and Mason describes how you can parse complex input languages in C.

Expert C Programming: Deep C Secrets by van der Linden is an amazingly great (and hilarious) look into how C works (and, doesn’t work) under the hood.