2.1.mp4

What is an algorithm?

Welcome to week two. We’re glad you’re still with us!

Last week, you spent time reviewing some essential programming skills, in particular, the use of functions. I shared how programmers use functions to build reusable chunks of code and manage complexity, and proposed an analogy with a machine as a useful tool to explain functions to beginners. This analogy was also useful when you looked at inputs and outputs of functions, or parameters and return values.

This week, you will explore what programmers mean by the term algorithm and how algorithms relate to the programs you write. There will be an opportunity to write some plain English algorithms, and then use your new-found list skills to turn these algorithms into programs. You’ll also extend your Python skills by learning to create and manipulate lists.

Once again, there are plenty of quizzes to help you check your learning, as well as opportunities to share and discuss with other learners on the course. I’d like to encourage you to take these opportunities in order to make the most of week two.

Algorithms

One of the challenges associated with learning (and teaching) computer science is the additional vocabulary it introduces; from complex terms for simple enough ideas (algorithm, decomposition, abstraction), to terms that have specific meanings in other areas of learning (‘variable’ means something subtly different in maths, science, and computer science). While it is important that learners understand terminology, more important is their grasp of the concept or idea in question.

An algorithm can simply be defined like this:

An algorithm is a set of instructions that precisely describe an activity.
The instructions should be such that any ‘actor’ could follow to produce the same output.
An ‘actor’ refers to anything, whether human or computer, that is capable of computation.

Algorithms are used every day all around us to do all sorts of things, such as finding a route from A to B, searching the internet for the most-appropriate pages, or recommending movies based on your past viewing.

While you could substitute the word instructions for algorithms, there are some important characteristics of algorithms that make them distinct. An algorithm:

Must be sufficiently precise and unambiguous: every step must be clear to the actor implementing the algorithm.
Must be repeatable, producing predictable outcomes based on the same inputs.
Can be represented in a range of forms, including natural language, pseudocode, flowcharts, or symbols.

Algorithms and programs

The term algorithm is often confused with program. Whilst the two are related, they have different meanings. An algorithm describes how to complete a specific task for any actor to carry out, whereas a program is the same algorithm implemented for a computer to carry out and written following the rules of a given programming language.

For example, the square-drawing algorithm below could be implemented using the pen tool in Scratch, the turtle module in Python, or even by a floor robot.

One problem, multiple solutions

One of the really interesting aspects of algorithms is that, for any given problem, there are numerous solutions. Whether they be simple, elegant, speedy, or inefficient, they are all valid, as long as they produce the desired outputs.

In a group of learners, it’s entirely possible to get a different, but correct, algorithm created by each of them. At this point it’s important to acknowledge their correctness, but also to encourage conversations and compare the merits of each approach. It’s this higher-order thinking that programmers need to evaluate the ‘best’ approach while being aware of alternatives.

Why are algorithms so important?

Depending on where they are used, the performance of your algorithm can have a significant impact on how useful it is. Different aspects of algorithm performance are important in different cases:

The algorithm that produces the route a satnav system will follow can choose from many different routes from A to B. It’s important to the user experience that they get recommended the optimum route: the user will not be pleased with a route that takes an hour longer than necessary. However, the user is unlikely to care too much if their journey takes a minute longer, as long as they get to their destination.
In contrast, the algorithms governing the behaviour of a self-driving car are taking split-second decisions based on the available sensor data. A bad decision by the algorithm could have life-and-death consequences, so these algorithms need to be robust and trustworthy.
Every time you do a web search, the search engine uses an algorithm to rank the most relevant pages for that search. In this case the user cares about the accuracy of the search results, but also the speed of the search. The quality of the search experience may impact upon the popularity of that search engine and ultimately the money a company is able to make from it.

Problem solved?

Algorithms are everywhere, as you’ve seen, but that doesn’t mean that they are fully understood. Far from it! While there are many problems which have been solved using algorithms and plenty of ‘standard’ algorithms, there is still work to be done. As a society we’re finding new problems we can solve with algorithms in science, medicine, arts, politics, and AI. New algorithms are constantly being invented and existing ones improved. To keep doing that, the world needs more computer scientists.

Over to you

Here I have mentioned some examples of algorithms, but there are many more. Using the comments, can you:

Write an algorithm to draw another 2D shape, at least 3 more shapes (e.g. a pentagon or star)?

<- previous

next ->