Your own project, and what comes next

What you will be able to do

After this lesson you can

• plan a project of your own with a method that you can repeat,
• choose the containers and the functions before you write code,
• test it with a table, and grow it in small steps, and
• know what to learn next.

The method

You have built tools, and a menu. Every one of them followed the same method. Here it is, in eight steps. Use it for any project.

1. The goal, in one sentence. "A program that keeps the marks of my class and tells me who needs help." If you cannot say it in a sentence, it is too big. Make it smaller.
2. The jobs. List what it must do: add a mark, show the average, list the students below 40. Each job will become a function.
3. The data. What does it know, and in which container? A list for an ordered row. A dictionary for a lookup by name. A set for who is there. A tuple for a fixed pair. Draw it on paper.
4. The expected answers. Make a small example, and work out the answers by hand. This is your answer key.
5. The functions. Write the thinking as small functions, each with parameters and return, and no input() or print() inside. Each has a docstring.
6. The tests. A table of inputs and expected answers, with the edges (empty, one item, ties, the biggest and smallest values). Run it after every change.
7. The user part. Only now, write the part that asks and shows: a menu, a loop with input(), the messages.
8. Grow it. Add one idea at a time. Run the tests. If a change breaks a test, you know where to look.

Step by step example: a unit converter

Suppose that your idea is: "a program that converts lengths between kilometres, metres and miles."

1. Goal. Convert a length from one unit to another.
2. Jobs. to_metres(value, unit), from_metres(value, unit), convert(value, from_unit, to_unit).
3. Data. A dictionary of how many metres in each unit: {"km": 1000, "m": 1, "mile": 1609.34}. A dictionary, because we look up by the name of the unit.
4. Expected answers. 1 km is 1000 m. 1 mile is about 1.609 km. 5 km is about 3.107 miles. 0 of anything is 0.
5. Functions. convert(value, a, b) is value * metres[a] / metres[b]. An unknown unit raises a ValueError with a clear message.
6. Tests. The four expected answers above, a test with the same unit twice (the answer is the value), an unknown unit, a zero, a negative value.
7. User part. A loop that asks: value, from, to, and shows the answer. A bad unit gives a message and asks again.
8. Grow. Add temperatures (not a simple multiplication! a new idea), and a list of the last five conversions.

Notice how little of the work is the keyboard. Most of the thinking happens before the user part, and is easy to test.

Do it: a few ideas

Choose a project that you care about. A project that matters to you is a project that you will finish.

Stuck? A short list

• It does nothing. Did you call the function? Print the values inside.
• A message in red. Read the last line first. Name the error. Find the line.
• A wrong answer. Work out the right answer by hand for a tiny input. Trace the program on it.
• It is too big. Do only the first job. Make it work. Then the second.
• You do not know how. Write the problem as a smaller problem that you do know, and solve that.
• Ask for help well. Show the smallest program that has the problem, the full message, and what you expected.

The last line is a good test of any pair of functions: going there and back should give the number you started with.

What comes next

You have finished the Essentials. You can store data in all four containers, handle errors, use modules, search and sort, think recursively, and work with text. The next stage opens bigger doors:

• Objects and classes: make your own kinds of thing, with their own data and actions.
• Files: read and write data that lasts after the program ends.
• Comprehensions and generators: shorter and faster ways to build sequences.
• Algorithms and data structures in more depth: stacks, queues, graphs, and how to measure speed.
• Python for science and maths: numpy for fast number work and matplotlib for graphs, with real data from physics, chemistry and statistics.
• Larger projects: a game, a data analysis, a small website.

Wherever you go, the habits of this book go with you: plan first, work out the answers by hand, keep the thinking in small functions, test the edges, and grow one step at a time.

Common mistakes

• Starting with the keyboard part. The thinking comes first.
• A project that is too big. Cut it to one job.
• No test table. You will be afraid to change anything.
• Never finishing. Finish a small version, and then improve it.
• Copying a solution that you do not understand. Type it in yourself, and change it.

Practice

Start with two quick questions. Each New numbers gives you a fresh one, so you can practise as long as you like.

Then the labs and the problems:

1. Warm-up. Open the lab Plan and test: a unit converter. You write the converter, and the checker tries the edges.
2. Build. Choose one idea from the table, and write steps 1 to 4 on paper.
3. Stretch. Build your project in the eight steps, and show it to a friend.

Remember these

• The method has eight steps: goal, jobs, data, expected answers, functions, tests, user part, grow.
• Put the thinking in functions without input() and print(). Test them with a table.
• Make a small version work before you add ideas.
• A project that you care about is a project that you finish.

Check yourself

Try each one on your own first. If you are stuck, open a hint. Then check with Show the answer.

1. Which step comes before the functions?

2. Where does the user part come in the method, and why?

3. What is a good first version of a big idea?