I've got an early 60's edition of The American Practical Navigator. It amazes me that one young man, Nathanial Bowditch, wrote this book in the late 18th Century and helped to make America a Sea Power.
I read through parts of it from time to time and every once in a great while will have a mathematical revelation. The other 99% of the time I put it down utterly frustrated at my inability to comprehend it.
I, too, possess a copy of that book. I acquired it back in the eighties, faced with an interesting problem in database marketing. We had a list of several million consumers who were good prospects of our client, who had a network of retail outlets. The idea was, let's send out computer letters to the consumers, informing them where the nearest outlets were. But that would have needed a table of nearest stores by zipcode. Which we didn't have.
However, someone on our team latched onto a database linking five-digit zip codes to latitude / longitude pairs (trivial in the Google age, but required connections of the non-ISP variety in the eighties). Obviously, the answer was to compile a list of nearest stores based on zipcode and use that. Simple: just calculate the distance from the customer's zip to the nearest outlets' by zip. That's where Bowditch (and the IBM 360 floating point instruction set and Fortran 4H libraries) came in (with some assembly language trig simplifications in the interest of computational efficiency).
Of course, it didn't work out too well. Some of the results were positively embarrassing. E.g., folks on Long Island needed boats to reach those near stores in Connecticut. But, for most of the country, it was fine, and the marketing campaign was judged successful.