The basic ideas of programming

A computer calculates faster than any human and never gets tired doing it, but it understands nothing on its own: it follows instructions, exactly the ones that are written down. Programming therefore means describing a task so precisely that someone who neither thinks along nor asks back can carry it out. There is a handful of basic ideas for this that come back in every programming language. This lesson introduces them without a single line of code, with pictures from everyday life, and the lessons after it show you what each of them looks like in Python.

A program is a recipe

A recipe is a sequence of steps in a fixed order: first melt the butter, then stir in the sugar, then the eggs. A program is built the same way, except that the cook is a computer that takes every step literally. If the recipe says "butter on the bread", it puts the whole pack on top; "a pinch of salt" is too vague for it, and it does not notice a missing step. Such a precise sequence of steps that reliably solves a task is called an algorithm, and a program is an algorithm written down in a language the computer can read.

From this follows the most important habit in programming: when a program does something other than what you want, the computer has almost never made a mistake. It carried out what is written, and the question is what you need to write differently.

Input, processing, output

Think of a bottle deposit machine in a supermarket. You put bottles in, which is the input. The machine recognises each bottle and adds up the deposit, which is the processing. At the end it prints a receipt, which is the output. Almost every program follows this pattern: it receives something, such as a number you type or a file, does something with it and returns a result, on the screen, in a file or as a signal to a device.

Variables: containers with a label

A program has to remember things: your name, an intermediate result, the number of bottles. That is what variables are for. Picture a variable as a container with a label stuck on it that carries a name. Inside the container lies a value, and through the name on the label the program finds it again whenever it needs it.

Three things belong to this picture. A container always holds exactly one value: if you put a new one in, the old one is gone without being kept anywhere. The new value may be calculated from the old one: the variable points works like the scoreboard in a stadium, which looks at what it shows with every goal and puts up one more. And the name on the label is meant for humans: the computer does not care whether a container is called age or x, but whoever reads your program, including you yourself in three weeks, understands age at once.

What lies in a container can be of different kinds: a number, a text or an answer such as yes or no. The kind decides what you can do with the value. You can add two numbers, but you can only join two texts together, just as you weigh flour but count eggs.

Lists: a train with numbered carriages

Sometimes it is not one value but a whole row of them: all the shopping, all the readings of a day, all the names in a group. That is what lists are for. A list is like a train: each carriage carries a value, the carriages stand in a fixed order and are numbered, and you can couple one on at the end or uncouple one. The number gets you to a particular carriage. Many languages, Python included, start counting at 0, so the first carriage has the number 0.

Branches: the bouncer

A program often has to decide what happens next. Think of a bouncer outside a club: they ask every guest the same question, namely whether they are at least 18, and depending on the answer they let them in or send them home. Such a branch always consists of a question that can be answered with yes or no, the condition, and two ways, of which exactly one is taken.

In everyday life you make such decisions all the time without thinking about it: if it rains, you take the umbrella, otherwise the sunglasses. In a program you have to write them down explicitly, including the second way, because whoever forgets it has a program that simply does nothing when the sun shines. If there are more than two possibilities, you line up several questions, the way a traffic light first asks whether it shows red, then whether amber, and otherwise green.

Loops: the same once more

Many tasks consist of doing the same step again and again. For a salad you peel each onion in the bag, one after the other, and stop when the bag is empty. A loop writes such a step down only once and adds how often or how long it is repeated.

There are two ways of ending a loop. The first is the one in the picture: once for each piece of a list, so as often as the list is long. The second runs as long as a condition holds, like stirring a batter, which only stops when there are no more lumps; how many rounds that takes, nobody knows in advance.

The second kind hides a trap that is printed on many a shampoo bottle: "Lather, rinse, repeat." Whoever takes that literally stays in the shower forever, because nowhere does it say when to stop. A loop whose condition never becomes false is called an infinite loop, and it is one of the most common mistakes at the start.

Functions: the coffee machine

A coffee machine can make a cappuccino, and you do not need to know how hot the water has to be or how long the milk froths. You press a button, perhaps say whether the cup should be large or small, and get a cappuccino. A function is such a machine inside a program: a sequence of steps that gets a name, such as make_cappuccino, so that you can call it with one word instead of writing the steps out again every time.

What you give the function when you call it is called a parameter, and what it gives back to you is its return value. A function is thus a small program with its own input, processing and output. That has two advantages: you write the steps down only once and can use them as often as you like, and whoever uses the function does not need to know its inside, just as with the coffee machine. With functions you also cut a big task into small ones that you can solve and check one by one; how to go about that is shown in Cutting problems.

When it does not work

No program runs on the first attempt. Finding mistakes, called debugging, is therefore not a sign that you are doing something wrong but the largest part of the work. You go about it as with a cake that did not rise: instead of baking everything again, you check step by step where the result first differs from what you expected. To make that easier, you write comments into the code, notes for humans that the computer skips, like the pencil note in the margin of a recipe.

What comes next

Keep the pictures in mind when you start with Python in a moment. The language is new, but you now know the ideas behind it.