1 · Every character is a number
Code point in decimal and hex, its UTF-8 bytes, and the same letter in the other case. Upper and lower case are different numbers; Chinese has no case.
2 · The report filter that skipped records (binary order, byte by byte)
Upper bound "ZZZZZZ": on
Add
3 · Case and alphabet in real tokenizers
Same sentence, as written vs ALL CAPS (tokens):