UMBC CMSC202, Computer Science II, Spring 1998, Sections 0101, 0102, 0103, 0104 and Honors

Thursday March 19, 1998

Assigned Reading:

• A Book on C:
• Programming Abstractions in C:

Handouts (available on-line):

Topics Covered:

• Yet another implementation of the list ADT. This time, we use a doubly-linked list. Each node on the list has a pointer to the previous item on the list as well as a pointer to the next item. Thus, it is possible to traverse the list in both directions. In particular, we can then use the pointer to a node to represent the position of a node we want to delete. (In the singly-linked list implementation, we used the pointer to the node before the node we want to delete.) In this example, we also use make the list circular. That is, the last node of the list points to the header node instead of NULL. (See header file, implementation, main program and sample run.)

• How would you sort a linked list? It turns out that the easiest way is to sort an array of pointers to nodes on the linked list. In this example, we reuse the Mergesort function from before. Note that all we needed to change in the source code for Mergesort was two type definitions and the one line in the merge routine which does a comparison. (See header file, implementation, main program and sample run.)

• Next we considered a new ADT that maintains an ordered linked list. The header file shows that the new interface simply includes list1.h then provides a new insertion function and two new search functions. The search functions take advantage of the fact that the list is ordered. So, if we see an item that is alphabetically larger than the key, then we already know that the list does not contain the item. Note that it was necessary for us to use new names for the new search routines so they do not conflict with the search routines in list1.h. Also, the implementation of the ordered list ADT must have its own NewNode() and FreeNode() functions because these functions were private in the implementation of list1.c --- even though the implementation of NewNode() and FreeNode() are identical in the two files. It is a serious deficiency of C that we cannot have functions be "private" in certain situations but "available" in other situations. This deficiency and the naming deficiency mentioned previously will be resolved when we use C++. (See also: main program and sample run.)