Object Orientation “Thinking” Object Oriented, Further Constraints, and Documentation

An Aside To some of you, I imagine that some of C++’s syntax and structure may be pretty foreign, to say the least. – In particular, some people have never worked (heavily) with OO before. – This is because there’s a whole different way of thinking about programming tasks in OO.

Programming Paradigms In the world, there are over 6900 different spoken languages. Some of these languages are more similar to each other than others. – Consider Spanish, French, Italian… – Consider Hindi, Urdu… or perhaps Mandarin Chinese.

Programming Paradigms Just as there are families of real- world spoken languages, there are multiple families – or paradigms – of programming languages.

Imperative Chances are, most of your programming experience is in the imperative paradigm. – Think C, for example. – Code is executed one line after another, step by step. – The focus is on the order of code execution.

Imperative Imperative-style programs make little attempt to enforce good data organization habits. – This must be handled by the programmer. – Data is often in a global scope, accessible anywhere, at any time, by anyone. As are functions.

Object-Orientation Object-orientation is quite different. – As we’ve seen already, part of its design is to enforce the organization of data into logical, conceptual units within the system. – Each object keeps its data private (ideally) and seeks to enforce constraints to keep itself in a proper form.

Object-Orientation Object-orientation is quite different. – Work gets done by objects interacting with other objects. As such, the exact flow of execution in the program may not be easy to track. – Object orientation aims to avoid making anything truly global. Java doesn’t even allow “truly” global variables… though it’s easily possible to compensate for this. C++ allows them.

Coding in OO So far, we’ve seen a few examples of OO code, but I haven’t really broken down what’s going on behind the scenes to make everything work. Let’s look at an example and talk through it.

A Fraction Object class Fraction { private: int numerator; int denominator; public: Fraction add(Fraction &f); }

Exercise 1 Code up the Fraction object Include constructor with two ints Include no-arg constructor Include accessor and mutator methods Make a main() that instantiates a fraction, sets its member variables, and prints it out

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); }

Exercise 2 Now add the Fraction::add() method Add to main() another fraction instantiation, add the two fractions, and print out the result Compile and test

Ways to Create Objects Can instantiate f1 by initialization – Fraction f1(i,j); Can instantiate f1 by new – Fraction *f1 = new Fraction(i,j); – But f1 is not a Fraction… – … it is a pointer to a Fraction object

Ways to Reference Objects When f1 instantiated by initialization – Fraction f1(i,j); – Access members by “.” When f1 instantiated by new – Fraction *f1 = new Fraction(i,j); – Access members by “->” – Use object by “*” dereference

Coding in OO First, let’s examine this line of code. f1.add(f2); //Both are Fractions What is this setting up and modeling? Secondly, what is going on in add() ?

f1.add(f2); //Both are Fractions This line is basically saying “Call the “ Fraction.add() ” method from the perspective of f1. Coding in OO

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); } So, that line of code has an implied reference to what was previously called “f1.”

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); } This “implied reference” is known as this within C++. It’s understood to be implied on any “unqualified” field names in the method below.

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); } The use of “ numerator ” and “ denominator ”, when not preceded by “ f. ” here, are with respect to this.

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); } What about when we do have “ f. ” preceding numerator and denominator ?

A Fraction Object public Fraction* Fraction::add(Fraction &f) { int num = numerator * f.denominator; num += f.numerator * denominator; int dnm = f.denominator * denominator; return new Fraction(num, dnm); } In such cases, the perspective shifts to that of the object f, from which it then operates for the field or method after the “. ”.

f1.add(f2); //Both are Fractions Even though the add() method is operating with two different Fraction class instances, the code is able to keep track of which is this and which is the parameter f. Coding in OO

An Aside There also exist other programming paradigms. – We’ll visit some others later on, but I’d rather keep things here for now and avoid confusion.

Questions?

For the rest of this lecture, we’ll switch gears to examining documentation and a bit more in regard to analysis.

Documentation Documentation is the “plain” English text accompanying code that seeks to explain its structure and use. – Some of this documentation is typically in comments, directly in the code. – Other documentation may be in external documents.

Documentation Types I expect four kinds of documentation – Source comments – enough to follow code – README file – how to install and configure, info on version #, etc. – User documentation – man page on how to use – Project report – description of purpose, theory, design, analysis, verification, and validation In addition, source code should include – Header files(s) – Implementation file(s) – Application (driver) file(s) – Makefile

Project Documentation Project report – should give the purpose, the overall structure of all the system, as well as the theory of operation, analysis of efficiency, proofs (if any), test plan and results, and known bugs – Comments in code necessarily follow the actual layout of the code, which may not be the best for understanding it

Code Documentation For complex code, it can be very helpful to place inline comments on a “paragraph” level, explaining what purpose that block of code is accomplishing. – A line-by-line commentary may clarify what the code is doing, but rarely indicates why. Note the purpose of your code – its goal.

Code Documentation We’ve already noted two different ways to comment within C++: // This is a one-line comment. /* This is a block comment, spanning multiple lines. */

Code Documentation In producing documentation for a method, it is wise to place some form of the “relationships” criterion within the description. – Generally, the conceptual purpose which a method, field, or class serves.

Code Documentation One should also include an explanation of the method’s preconditions, if it has any – Preconditions: the limitations a particular method imposes on its inputs – If a method is called with arguments that do not match its preconditions, its behavior is considered to be undefined – These are very useful for proofs also!

Documentation As there exists a notion of preconditions, there also exist postconditions – Postconditions: the effect a method has on its inputs (any unaffected/unlisted input should remain untouched), any generated exceptions, information about the return value, and effects on object state – These are also very helpful for proofs!

Benefits Documentation helps other programmers (or yourself at some later time) to understand the role of each accessible field and method for a given class Inside of code, documentation provides great reference material for future maintenance efforts It can serve as a starting point for project report, but should not have everything a project report has