The long S is also problematic for OCR: https://books.google.com/ngrams/graph?content=cafe&year_start=1600&year_end=2000&corpus=15&smoothing=3 https://books.goo

The long S is also problematic for OCR: https://books.google.com/ngrams/graph?content=cafe&year_start=1600&year_end=2000&corpus=15&smoothing=3 https://books.google.com/ngrams/graph?content=cafe&year_star... (Google ngram viewer is pulling the word "case" with a long S from the many legal documents in its corpus)