Answer

Background

Multidimensional arrays in C were originally designed for ease of use with tables and matrices. The classic two-dimensional array is an array of arrays, allowing simple syntax for working with elements of tabular data. Over time, the approach has evolved, especially when dealing with variable-length arrays.

The Problem

Incorrect declaration and initialization of multidimensional arrays lead to compilation errors or logical failures. When passing a multidimensional array to a function, many developers get confused due to specification requirements — all dimensions except the first must be explicitly specified.

The Solution

Declaring a two-dimensional array:

int matrix[3][4];

Full initialization:

int matrix[2][3] = {
    {1, 2, 3},
    {4, 5, 6}
};

When passing to a function, all sizes except the first must be explicitly stated:

void printMatrix(int m[][3], int rows) {
    for (int i = 0; i < rows; ++i) {
        for (int j = 0; j < 3; ++j)
            printf("%d ", m[i][j]);
        printf("
");
    }
}

With the introduction of the C99 standard, functions accepting variable-length arrays can be declared:

void foo(int rows, int cols, int a[rows][cols]);

Key Features:

It is necessary to explicitly specify the sizes of all dimensions except the first when passing to functions.
A multidimensional array in C is an array of arrays; elements are stored in memory in row-major order.
Partial initialization initializes unspecified elements to zero.

Tricky Questions.

1. Can a function be declared that accepts a two-dimensional array without specifying the second size?

No, C requires that all sizes except the first be known at compile time. This relates to pointer arithmetic when accessing elements.

Example of an error:

// Error:
void process(int arr[][], int rows); // Not allowed

2. What happens if not all elements of a multidimensional array are initialized?

Remaining elements will automatically be filled with zeros if the array is static or global. For a local array with partial initialization, implicitly initialized elements will also be zeros.

int a[2][3] = {{1}, {4}}; // a[0][1] and a[0][2], a[1][1] and a[1][2] will become 0

3. What is the difference between an array of pointers and a two-dimensional array?

A two-dimensional array is a single memory block. An array of pointers is a set of pointers to separate (possibly separately allocated) one-dimensional arrays. This is important, for example, when allocating memory for "ragged" arrays.

Common Mistakes and Anti-Patterns

Errors in declaring functions with multidimensional arrays (omitting the size of "inner" dimensions).
Mixing up rows and columns when working with row-major layout.
An array of pointers being substituted with an actual multidimensional array and vice versa.

Real-life Example

Negative Case

Attempting to declare and pass a two-dimensional array to a function without specifying the second size, resulting in a compilation error. A surface-level fix by replacing it with a pointer leads to undefined or incorrect behavior during further calculations.

Pros:

Simplicity of function declaration (at first glance).

Cons:

Compilation errors, incorrect index arithmetic, data corruption.

Positive Case

The developer explicitly states the sizes of all dimensions, clarifying in documentation the order of element storage in memory, thereby reducing the number of errors in further maintenance.