A class compares three solutions for a classroom code. Goal: quick to send and accurate to decode. Test evidence: A: random long code (slow, many errors); B: short agreed code (quick, few errors); C: changing code (quick, many errors). Which solution is best supported by the evidence?