Coding classes in West Yorkshire

The learner downloads four novels from Project Gutenberg, Jane Eyre, Wuthering Heights, Agnes Grey and The Tenant of Wildfell Hall, and a sentiment word list called AFINN, which gives 3,382 English words a score from −5 for very negative to +5 for very positive. The program splits each book into words, adds up the scores, and reports the result per thousand words so that long and short books can be compared. It then does the same chapter by chapter.

The headline looks sensible: Wuthering Heights is the only one of the four with a negative score, and its darkest chapter scores −44.6. But three problems appear as soon as the learner looks closer. First, coverage: the list knows only about 7 per cent of the words in each book, so more than nine words in ten count for nothing. Second, negation: in phrases like "not happy" the list still scores "happy" as positive. Between 5.6 and 7.7 per cent of scored words follow a "not", "no" or "never" within three words; flipping them moves Agnes Grey only from +24.6 to +23.4, so here it matters less than you might fear.

The third problem teaches the most. AFINN scores "miss" as −2, meaning to fail or to long for something. But in a Victorian novel, "Miss" is usually a title. Jane Eyre uses the word 310 times, 262 of them directly before a capitalised name, and every one drags the score down. The list reads a polite form of address as a small sadness, 310 times over. Removing it lifts Jane Eyre's darkest chapter from −48.7 to −39.4 per thousand words. A word can have several meanings, and a program that ignores context counts the wrong one.

Learning to code in 2026 lets a young person in West Yorkshire check what a language tool is really measuring. The longer argument is in why teenagers should still learn to code in 2026.