Get a grown-up, a piece of paper and a pen. You are going to tell them how to draw a cat, they are going to do exactly what you say, and the cat is going to be terrible. That is the point.
You give the instructions. The grown-up holds the pen.
Here is the only rule, and they have to promise: they do exactly what you say, and nothing else.
No helping. No guessing what you meant. No drawing the sensible thing.
Go on. Tell them your first instruction.
You say: draw a circle.
So they draw one enormous circle, right across the whole page, corner to corner. There is no room left for a cat.
And they are right. That is what you said. You never said how big.
You say: put two eyes near the top. They draw two eyes near the top of the paper, floating in the air outside the circle.
The brilliant part is that you can see every mistake. It is sitting there on the page, in pen, and it is your fault.
A computer does what you say. Not what you meant.
Well. A bit annoying. But they are showing you something true.
A computer is not clever. It will never work out what you probably wanted.
A computer is not stupid either. It did not make a mistake. It did the thing you asked, perfectly.
A computer is exact. That is the word.
And it never gets bored and it never says are you sure? If you tell it something silly, it will do that silly thing again and again and again, all day, and never once ask you about it.
Start again with a clean sheet. Press the button to give one small instruction at a time, and watch a cat actually turn up.
You say: draw a circle. The grown-up fills the entire page with one giant circle. Why did they do that?
Here are some drawing instructions. Some of them a computer could do straight away. Some of them need a person to guess what you meant. Tap each one into a box.
Tap an item, then tap where it belongs
There is a cat, a house and a tree on the page now. Your next instruction is: colour it in. The grown-up stops dead and will not move. What is the trouble with that instruction?
You’re previewing as a parent — nothing here is recorded.
That’s the last one in this channel for now. More are on the way.
Back to all episodesGiving a partner spoken instructions to draw a picture they cannot see, and following those instructions literally, is a long-standing unplugged computing-education exercise used to show that a machine executes instructions exactly as given rather than inferring intent (CS Unplugged, University of Canterbury; Code.org CS Fundamentals unplugged lessons, e.g. 'My Robotic Friends', which uses the same instruct-and-execute structure with cups). A computer program is a sequence of instructions carried out exactly as written, with no capacity to infer the author's intent — which is why an ambiguous instruction produces an unintended result rather than a sensible guess (Encyclopaedia Britannica, 'computer program'). 'Decomposition' — breaking one large instruction into smaller ones a machine can carry out — is one of the named core skills of computational thinking in school computing curricula (Computing at School / BCS, 'Computational Thinking: A Guide for Teachers'; CSTA K-12 Computer Science Standards). No figures are quoted anywhere in this episode. Nothing here rests on a number.