Lecture 7 program development issues (supplementary)

1 of 31Module 7 : Algorithms and Program Development

Introduction to Computational Thinking

Module 7 :

Program Development Issues

(supplementary material)Asst Prof Chi‐Wing FU, Philip

Office: N4‐02c‐104email: cwfu[at]ntu.edu.sg


Topics• Basic Rules in writing programs• Program Readability• More on Problem Solving• Planning for testing, verification, and

documentation


A Good ProgramWhat makes a good program?

• A program is a reflection of the writer and their thoughts

• First, you must have some thoughts (computational thinking)!

• The difficulty for most people is to figure out what has to be done, the problem solving, before writing a program


Three RulesRule 1: Think before you program

* Understand & Analyze the problemRule 2: A program is a human-readable essay

on problem solving that also happens to be executable on a computer* A program -- Not just for computer to run,

but also for human to readRule 3: Practice! Practice! Practice!

The best way to improve your programming and problem solving skills is to practice



documentation


Program Readability• We will emphasize, over and over, that a

program is not just for computers to read, but also intended to be read by other people, even if “other people” is you in the future!

• Write a program so that you can read it, because sometime in the future you may have to read it again…

• So… Any guideline?


Readability(1): Naming• The easiest thing to do that affects

readability is good naming• Use meaningful names for the items

you create that reflect their purpose• To help keep straight the types used,

include that as part of the name. Python does not care about the type stored, but you do!


What does this do?

• What is a? What is b?


How about this?

• More specific names on variable• Hungarian notation: append types


Readability(2): Space

More space in the program helps reading also(my personal habit: align assignment op.)


Readability(3): Comments• info at the top, the goal of the code• purpose of variables (if not obvious by

the name)• purpose of other functions being used• anything “tricky”? If a piece of code took

you some time to write, it is probably hard to read and demands for a comment


Readability(3): Comments

Visual example in module 6.2

Helps understanding when programis not short or not straightforward


Readability(4): Indentation• A visual cue to say what code is

“part of” other code• This is not always required in many

programming languages (more freedom), but Python forces you to indent properly

• This aids readability greatly



documentation


Strategies: Problem Solving• Engage/Commit• Visualize/See• Try it/Experiment• Simplify• Analyze/Think• Relax


Engage• You need to commit yourself to

addressing the problem. • Don’t give up easily• Try different approaches• Set the “mood”

• Just putting in time does not mean you put in a real effort!!!


Visualize/See the Problem• Find a way that works for you,

some way to make the problem tangible.• Draw pictures• Layout tables• Literally “see” the problem somehow

• Everyone has a different way, find yours!


Try It/Experiment• For some reason, people are afraid to just

“try” some solutions. Perhaps they fear failure, but experiments, done just for you, are the best way to figure out problems.

• Be willing to try, and fail, to solve a problem. Get started, don’t wait for enlightenment!


Simplify• Simplifying the problem so you can get a

handle on it is one of the most powerful problem solving tools.

• Given a hard problem, make is simplier(smaller, clearer, easier), figure that out, then ramp up to the harder problem.


Think it Over/Analyze• If your solution isn’t working:

• Stop• Evaluate how you are doing• Analyze and keep going, or start over.

• People can be amazingly “stiff”, banging their heads against the same wall over and over again. Loosen up, find another way!


One More Thing: Relax• Take your time. Not getting an answer

right away is not the end of the world. Put it away and come back to it.

• You’d be surprised how easy it is to solve if you let it go for awhile. That’s why starting early is a luxury you should afford yourself.


Example: Cryparithmetic Problem

E L F+ E L FF O O L

E= ?F = ?L = ?O = ?

Remember:• Engage• Visualize• Try it• Simplify• Analyze• Relax



documentation


Program Development1. Problem Specification

e.g., identify requirements & goals2. Problem Analysis (how to solve it)

e.g., identify input & output, formulae3. Program Design

- write solution steps e.g., pseudo code and flowcharts

4. Implementation- translate your solution into program- build first program skeleton/outlinee.g., comments on major steps

5. Program Testing- use test samples to try your program

Problem Specification

Problem Analysis

Program Design

Implementation

Program Testing

It can be aniterative process!!!


Program Testing• A program is correct if it returns the correct result for

every possible combination of input values.

Exhaustive testing: use all possible combinations of input values and check the output is correct. This will take a whole year, or forever, to show the program is correct.

-> Impractical!!!

• What we can do: 1) use test data that causes every program path (e.g., in branching and looping) to be executed at least once; and 2) think and be creative!!!


Programming Errors• Syntax Errors

• “grammatical” errors• detected by compiler and found automatically• need to be fixed before the compiler can understand the

code• E.g., missing colon before a block in while or for loops

• Runtime Errors• execution error (e.g., divide by zero, program crash, etc.)• detected during the execution of code• error messages may be useful to help identify the

reasons and locations• sometimes not easy to fix


Programming Errors• Logical Errors

• due to error in designing the algorithm or implementation

• no compilation errors; no run-time error message• most difficult to detect• e.g., in a role-playing game, you killed a monster

but experience point is not given to you (but not supposed to!!!)… note: you can still play the game… no crash! But no level up!


Debugging• The process of finding and correcting

errors, especially logic errors (BUG!!!)• Strategies:

•Hand Tracing or Simulation•Program Tracing•Use print() function at appropriate program

locations to check:•Program control flow•Values of the variables

• Try different user input•Be Patient! Don’t give up!!!


Documentation• Documentation is needed for further modification

and maintenance• Proper documentation includes:

• problem definition and specification;• program inputs, outputs, constraints and mathematical

equations;• algorithms and logic, e.g., flowchart, pseudo code, etc.• source code with appropriate comments;• sample test run of the program; and• user manual for end users (how to use it)

• It should be done alongside with the program development but not the very end!!!


Take Home MessagesBasic Rules:- Think before you program- A program is a human-readable essay on problem solving- Practice! Practice!! Practice!!!Program Readability: Naming; Space; Comments; IndentationProblem Solving Strategies:- Engage/Commit; Visualize/See; Try it/Experiment; Simplify;

Analyze/Think; RelaxProgram Testing- need to design test data to try every program path- unit test: test your code piece by piece on their correctness!Programming Errors- Syntax Errors; Runtime Errors; Logical ErrorsOther issues: Debugging and Documentation


Reading Assignment • Textbook

Chapter 3: Algorithms and Program Development3.1 to 3.5

Note: Though some material in textbook is not directly related to the lecture material, you can learn more from them.

Lecture 7 program development issues (supplementary)

Technology

programmingand problem

hard problem

problem tangible

harder problem

humanreadable essayon

best way

different way

aids readability