Python Finite Automata — ELI5

Think about a vending machine. It’s in a waiting state. You put in a quarter — now it’s in the “25 cents” state. Another quarter — “50 cents” state. You press a button, and if you have enough money, it gives you a snack. If not, it stays waiting for more coins.

The vending machine only remembers one thing: how much money is in it right now. It doesn’t remember what coins you put in yesterday or what order they came in. It just looks at its current state and the next coin to decide what happens.

A finite automaton works the same way. It has a fixed set of states (like “0 cents,” “25 cents,” “50 cents,” “enough”). It reads input one piece at a time. For each input, it checks: “Given my current state and this input, which state do I go to?” That’s it. No memory of the past, no looking ahead.

Where do we actually use these? Everywhere, surprisingly:

• Spell checkers — is this sequence of letters a valid word?
• Pattern matching — does this text contain an email address?
• Traffic light controllers — what light comes next?
• Simple game AI — the enemy patrols, chases, or attacks based on what it sees right now

In Python, finite automata are the engine behind regular expressions. When you search for a pattern like “any number followed by a dash followed by letters,” Python builds a little automaton that reads your text character by character and decides if the pattern matches.

The “finite” part means there’s a limited number of states. A vending machine might have 10 states. A regex automaton might have 50. But it’s always a fixed number — no matter how long the input is.

One thing to remember: A finite automaton is a simple machine with a fixed number of states that reads input one piece at a time — and Python uses them behind the scenes every time you write a regular expression.