Coding classes in Antrim and Newtownabbey

Programmers call a dead-end line a stack: you can only add or remove at the open end. The learner models 24 aircraft, all invented, parked in fingers of different depths, from 24 separate single spaces to one long line of 24. Each aircraft leaves in its own order. To get one out, every aircraft between it and the open end must be towed out and back in. The program counts those blocking moves, repeats the whole thing 4,000 times with random parking, and averages.

The simulation matches a neat formula. With random parking, any two aircraft in the same finger are equally likely to be in the wrong order, so each finger of depth k causes k times k minus 1 over 4 blocking moves on average, and 24 aircraft in fingers of depth k cause 24 times k minus 1 over 4. Deep fingers pack more aircraft into less space, which is why the record says they allowed much larger numbers to be stored, but the towing grows with depth. One long line of 24 costs 138 moves on average, and 202 on the worst run.

Then planning changes everything. If you know the departure order, park the last to leave deepest and the first to leave nearest the open end, and nothing is ever blocked: zero moves at every depth. Real plans are rougher. When the program guesses departure order with an error of about four places, a single line of 24 still needs about 40 moves instead of 138. Learners also test the classic slip: counting a finger from the wrong end, which turns a perfect plan into the costliest one.

Questioning the cost of a design that already works is a sound reason for an Antrim and Newtownabbey teenager to keep coding in 2026. The longer argument is in why teenagers should still learn to code in 2026.