After careful consideration, I offer you the following example (note the fact that you can turn the pages and sort the entries, hence my suggestion of a doubly linked list which you mistook as a char[] array ..)

My original answer is below (Yes, I would rather find screen shots of Microsoft Bob than give you the full source code needed to satisfy your homework):
Well, lets think about all of the stuff that is interesting about people:
typedef struct person {
char *firstname;
char *lastname;
char *phone;
struct person *prev;
struct person *next;
unsigned fname_hash;
unsigned lname_hash;
unsigned full_name_hash;
} person_t;
Note a hash for first, last and full name which is how people commonly search an address book.
Then take a look at a very, very basic hashing algorithm (pulled right out of an ini file parser (basically just a dictionary) that I maintain):
unsigned simple_hash(char *key)
{
int len;
unsigned hash;
int i;
len = strlen(key);
for (hash = 0, i = 0; i < len; i++) {
hash += (unsigned) key[i];
hash += (hash << 10);
hash ^= (hash >> 6);
}
hash += (hash << 3);
hash ^= (hash >> 11);
hash += (hash << 15);
return hash;
}
The hashing function above is not collision proof. You will still need to compare strings when you think you have a match, just to be sure. However, employing that hash (or another) is the fastest way to search through a linked list. You will be comparing unsigned values, not strings (unless of course the hash matches, then you compare just to check to be sure its not a collision).
When you begin, allocate enough space for n entries in the list (3 - 4 should be fine). After that, you will need to re-allocate for each addition or removal.
Of course, most people would like their address book to appear alphabetically, so inserting and deleting nodes is just half of the battle. You will also want to sort the list. Its very difficult to help you there with what you provided.
What I can say is address books commonly contain company information, so you might encounter "1A Pest Control". In that case, you might want to have a look at a natural language string comparison.
Finally, If this has to work on wide character sets, you need to edit and re-tag your question.