Design

• Object Oriented Design
  • In this methodology we thing of the data we are going to be working with
  • Both the types of data and the operations we wish to perform on this data.
  • We then build objects.
  • We have encountered this with strings
    • A collection of characters
    • I/O
    • Find substrings, length, patterns, ...
  • The idea is to encourage the development of libraries of objects to support reusable code.
  • An Example: arbritrary precision mathematics
    • The GNU project
    • GMP
    • Do not try this on cslab103
    • #include <gmpxx.h>
    • Class type mpz_class, for integer
    • operations as per normal with integers
    • compile a program

      g++ -o progname progname.C -lgmpxx -lgmp
      	    

    • This also has other types and many functions
    • Find 15!
    • Compute the odds of winning powerball
  • Notice how, given a well designed object, we are able to develope code very quickly using those objects.
  • Also notice that we didn't need to know how they implemented this class (object), we just need to know the interface.
  • This is a key concept: ADT - Abstract Data Type, or a data type whose properties (domain and operations) are specified independantly of any implementation.
  • In other words, design what you need, and worry about implementation after the design is done.
  • Much more on this in 131.
  • This is the idea of the 80's (or at least became very popular in the 80's)
• Functional Decompisition
  • This is a much older method
  • Start with an abstract problem
  • Break it into a number of smaller, but perhaps still abstract problems
  • Continue until you reach a concrete problem which can be easily implemented.
• Read the end of the chapter
• Read Programming at Many Scales, page 178
• Read Testing and Debugging page 180