JSON Structure

Perhaps we should create a JSON structure that acts as the translation source code. I don't really like XML, and that structure may be copyrighted anyway. If we use an autogenerator to patch the translation, we would want people to edit these JSON files and not the XML, so their changes are not overwritten.

The JSON structure could match what we already know and have: the KC3 translations. However, it has certain structural things that don't work great with applying to XMLs (different IDs, etc.)

KC3 to KCV JSON

The first step is to update our translations using the upstream KC3 translations. Usually it's the ship descriptions that get updated.

Perhaps the ones that are taken from KC3 can have boolean variables which tell the script to look for it elsewhere, and then new translations are in the JSON.

Any newlines need to be added systematically by the script.

KCV JSON to KCV XML

Then, the KCV JSON is injected into the XML. The compiler must provide the JP files themselves: that way, no copyrighted XML is used by us in the process.