Monday, June 24, 2019

The 5th edition of Data Structures and Abstractions with Java was given the 2019 Textbook Excellence Award from the Textbook and Academic Authors Association.  My co-author, Tim Henry, and I were honored to receive the award at the Association's 32nd annual conference in Philadelphia, Pennsylvania on June 24, 2019. The award recognizes "excellence in current textbooks and learning materials." Our book is published by Pearson Education.

Wednesday, September 5, 2018

Walls and Mirrors, 7th edition., receives the 2018 McGuffey Longevity Award

The 7th edition of Data Abstraction & Problem Solving: Walls and Mirrors was awarded the McGuffey Longevity Award from the Textbook and Academic Authors Association.  My co-author, Tim Henry, and I were honored to receive the award at the Association's 31st annual conference, which took place in Santa Fe, New Mexico. The award recognizes textbooks whose "excellence has been demonstrated over time." Indeed, the first edition of this book was published in 1995. Our 6th edition also received this award in 2016.

The 5th edition of Data Structures and Abstractions with Java is published

What’s New?

This new, 5th, edition of Data Structures and Abstractions with Java enhances the previous edition and continues its pedagogical approach to make the material accessible to students at the introductory level. The coverage that you enjoyed in previous editions is still here. As is usual for us, we have read every word of the previous edition and made changes to improve clarity and correctness. No chapter or interlude appears exactly as it did before. Our changes are motivated by reader suggestions and our own desire to improve the presentation.

In this new edition, we
• Adjusted the order of some topics.
• Added coverage of recursion in a new chapter that introduces grammars, languages, and  
• Added additional Design Decisions, Notes, Security Notes, and Programming Tips 
  throughout the book.
• Added new exercises and programming projects, with an emphasis in areas of gaming, 
  e-commerce, and finance to most chapters.
• Refined our terminology, presentation, and word choices to ease understanding.
• Revised the illustrations to make them easier to read and to understand.
• Renamed Self-Test Questions as Study Questions and moved their answers to online. 
  We encourage our students to discuss their own answers with a study partner or group.
• Included the appendix about Java classes within the book instead of leaving it online.
• Reduced the amount of Java code given in the book.
• Ensured that all Java code is Java 9 compliant

And our table of contents:

 Introduction       Organizing Data
 Prelude               Designing Classes
 Chapter  1           Bags
 Java Interlude  1 Generics
 Chapter  2           Bag Implementations That Use Arrays
 Java Interlude  2 Exceptions
 Chapter  3           A Bag Implementation That Links Data
 Chapter  4           The Efficiency of Algorithms
 Chapter  5           Stacks
 Chapter  6           Stack Implementations
 Java Interlude  3 More About Exceptions 
 Chapter  7           Queues, Deques, and Priority Queues 
 Chapter  8           Queue, Deque, and Priority Queue Implementations 
 Chapter  9           Recursion 
 Chapter 10          Lists  
 Chapter 11          A List Implementation That Uses an Array
 Chapter 12          A List Implementation That Links Data  
 Java Interlude  4 Iterators 
 Chapter 13          Iterators for the ADT List 
 Chapter 14          Problem Solving With Recursion 
 Java Interlude  5 More About Generics 
 Chapter 15          An Introduction to Sorting 
 Chapter 16          Faster Sorting Methods 
 Java Interlude  6 Mutable and Immutable Objects 
 Chapter 17          Sorted Lists   
 Java Interlude  7 Inheritance and Polymorphism
 Chapter 18          Inheritance and Lists 
 Chapter 19         Searching 
 Java Interlude  8 Generics Once Again
 Chapter 20         Dictionaries 
 Chapter 21         Dictionary Implementations 
 Chapter 22         Introducing Hashing 
 Chapter 23         Hashing as a Dictionary Implementation 
 Chapter 24         Trees
 Chapter 25         Tree Implementations 
 Java Interlude  9 Cloning
 Chapter 26         A Binary Search Tree Implementation 
 Chapter 27         A Heap Implementation 
 Chapter 28         Balanced Search Trees 
 Chapter 29         Graphs 
 Chapter 30         Graph Implementations 

 Appendix A       Documentation and Programming Style
 Appendix B       Java Classes                                   
 Appendix C       Creating Classes from Other Classes 

Online Supplements:
 Supplement 1    Java Basics   
 Supplement 2    File Input and Output   
 Supplement 3    Glossary 
 Supplement 4    Answers to Study Questions

Sunday, February 12, 2017

Wednesday, December 28, 2016

A way to learn more when you study

"You can truly understand a topic when you have to teach it." Many teachers have heard and agreed with this statement. The Feynman Technique for studying is based on this advice. Check it out!

Wednesday, November 30, 2016

Tree traversals and C++11 lambda functions

One of our readers of Walls and Mirrors 7th edition had a problem with the traversal operation for TreeDictionary, which is described in section 18.2.2 of Chapter 18 (page 559). My co-author, Tim Henry, fielded the question, and I’m happy to share his response with you. The problem involves the data type of the traversal’s parameter, and the solution involves lambda functions, a feature of C++11 but one that we do not cover in the current edition of our book.


Recall that TreeDictionary stores its key-value entries in a binary search tree. The traversal operation is declared as follows:

/** Traverses the entries in this dictionary in sorted search-key order
    and calls a given client function once for the value in each entry. */
void TreeDictionaryValueType>::traverse(void visit(ValueType&)) const;

This function has a single parameter, which is a function whose only parameter is of type ValueType.

Our BinarySearchTree, as described in Chapter 16, stores entries of type ItemType, and declares the operation inorderTraverse as

void BinarySearchTree::inorderTraverse(void visit(ItemType&)) const;

TreeDictionary stores key-value pairs as objects of type Entry into the binary search tree entryTree. When you call TreeDictionary::traverse, the visit function you pass to it expects an argument of type ValueType. But when TreeDictionary::traverse calls BinarySearchTree::inorderTraverse using the statement


visit expects an argument of type Entry. Therefore, the wrong type of argument is passed, and we get a compilation error.


Find a way to “convert” visit into a function whose parameter is of type Entry  so we can pass it to inorderTraverse. Maybe a function something like this:

void someFunction(EntryValueType>& someEntry)
   auto someItem = someEntry.getItem(); // Must have an address for someItem

But how do we give the function access to visit without changing its signature?


Lambda functions! Here is the new implementation for TreeDictionary::traverse:

TreeDictionary::traverse(void visit(ValueType&)) const
   auto treeVisit = [=](EntryValueType>& someEntry)
      auto someItem = someEntry.getItem();

The meaning of the previous lambda syntax follows:

[] means a lambda definition will follow. 
Only [] indicates to not “capture” or access local variables from the parent local environment. 
[&] means to access local variables as pass-by-reference. (This works for our scenario also.)
[=] means to access local variables as pass-by-copy (value). 

After the square brackets is the parameter list for the lambda function. In our case, it is the parameter type our tree needs: 
(EntryValueType>& someEntry)

Then we have the function implementation inside a pair of braces { } followed by a semicolon.

This works when the lambda function does not capture any local variables. 

BUT . . . 

We need to capture the local variable/parameter visit.

So we must use a more formal definition for the function parameters in the tree classes. (We could do this in TreeDictionary also, but it works without that change).

In BinaryTreeInterface.h, add:


Then in the tree files


change the signature of inorderTraverse so that it has the following parameter in red:

inorderTraverse(std::function visit)

This change relaxes some restrictions on how the lambda function can be passed as an argument. 

In BinaryNodeTree.h and BinaryNodeTree.cpp, you should also change the protected function inorder:

void inorder(std::function visit
                   std::shared_ptr> treePtr) const;

Make analogous changes to inorderTraverse and postorderTraverse.

Tuesday, October 18, 2016