CS 206 - Data Structures

Programming Design Guide

Now that you are no longer a new programmer, you are expected to pay attention to your code organization. Your programming should adhere to the standards explained here. While some of the rules are due to convention, many more are pearls of wisdom distilled from solid software-engineering principles and all of them are essential for making your code readable and/or easy to maintain and modify/reuse. This document explains guiding principles. For more specific formatting rules please refer to the formatting guidelines.

Intentional Coding

  1. Every line of code should have a deliberate effect on your program. If you are not sure why a line is there, it shouldn’t be. If your program stops working because you removed that line of code that you don’t understand, you need to understand it before you put it back.
  2. Within a single program, do not copy-and-paste code, only to change a few details. Instead, declare a methods or other abstraction and parameterize it. Do copy and paste code (changing as needed) between programs.
  3. Use public static final constants to avoid so-called magic numbers - numerical values that are scattered all over your program and are difficult to understand or modify.
  4. Avoid the use of static methods. There are some occasions where they are warrented. But these are rare. In almost all cases, static methods are a sign of lazy/sloppy programming
  5. Do not debug by trying all combinations. If you (or the debugger) pinpointed a line that doesn’t work, it is important to spend the time to understand why. The most valuable gain of programming proficiency happens right here. A bug is not truly fixed unless you
    1. understood why it happend in the first place
    2. understood why changing the code the way you did fixed it

Scoping and Encapsulation

  1. Declare variables in the smallest scope possible. That is, prefer local variables over instance variables.
  2. Never use public (non-final) instance variables. Use private and accessor methods (getters).

Class Design

In Java, instance variables store the data and keep track of the states of the objects. Thus it is very important to carefullly consider how to organize your classes and instance variables so that
  1. Classes are independent. There must be clear delinearization of responsibiities between different classes so that the work does not overlap. If methods in two different classes are doing the same thing, or data is stored in more than one place, you have bad design.
  2. Classes have good communication. Your classes need to work together, thus there must be ways for them to get what they need from each other.
  3. Methods should assume that all instance variables are always up-to-date. In other words, assume that all other methods will do what they are supposed to.
  4. Each method is responsible for one task and updating those related instance variables only.
  5. Instance variables are used in more than one place. Generally instance variables should have values that are set in a constructor and the used (an possibly) changed in one or more methods. An instance variable that is only used in one method should not be an instance variable.
  6. Static variables and methods should be avoided -- almost always.
    Exceptions:
    1. Static methods are OK only if the method naturally fits in the class and uses only the arguements to the method -- for instance Math.sqrt().
    2. Static "variables" are OK only to define constants. In this case it is also OK for the variable to be public. For example
                  public static final double PI = 3.1415;
              
  7. Main methods have two uses.
    1. Illustrate the use of the class and its methods. In this case the main could be quite long. It will naturally have a lot of print statements.
    2. Start the execution of the program. In this case the main method should be short. Generally it should create an instance of its enclosing class and pass the execution of the program to an instance method.

Data Structure Design

Often, there is more than one way to store and represent data.
  1. Do not store the same thing more than once, or in more than one place.
  2. Choose the smallest storage so that the above is true and accomplishes your algorithmic goals.