There's a thread for that... but I can't find it. So, from memory, it's something like this: For each file, divide the compressed size by 1024 and add 1 if there is a remainder. That's the number of pages in that file. Add up all the pages for all the files, and that's the number of pages for the book. Then as you read through the book, the number of Unicode characters in the current file is divided by the number of pages in the file to get the number of Unicode characters per page.
Actually, that is all described Adobe's little book EPUB Best Practices, which I downloaded from the Adobe site some time ago. It's the section on 'Synthetic Page Numbers'.
So it is pretty artificial, and the font size is not taken into account at all. These are totally synthetic, virtual, (nearly useless?) page numbers.
|