Is there any documentation on this? One of the reasons I need the transliteration map file is to compare my export library (saved in original Arabic) with Calibre's base library (saved in not so great a transliteration, if you don't mind me saying).
It's clear the transliteration to ascii has to happen, and I guess they have to be one to one mappings, but are the particular mapping choices set in stone or can we work on making them make more sense for Arabic unicode.
fyi:
I am working on a tool to romanizes Arabic records into standards such as the
Library of Congress standard, the ISO standard, and so on, which is how many libraries store their Arabic collections, sometimes even losing the link to the original Arabic records and keepings just the romanization. I'd be curious to know if anyone is interested in incorporating such a conversion tool into calibre. something that would make all naming variations easily retrievable.