There is no need for disassembly, besides it's illegal. I've been granted access to the AoD code by BL-Logic.
I do know what non-unicode is but I wasn't on the previous release teams. Also non-unicode, like ACL Desktop or AuditExchange non-unicode versions, don't support unicode (that's why it's non). But the biggest thing thing to know about non-unicodes is that they don't support multiple language specific characters. AoD supports 8 languages (English, French, Italian, Spanish, German, Polish, Portuguese, & Russian). The best thing to do would be to go to UTF-8, which is an unicode encoding.
I'm not willing to redo all of AoD to use UTF-8 without a dedicated team that has a lot of free time. It's a massive undertaking and I've already spent way too much time with AoD fixes, upgrades, technology researching, etc. Maybe in a future release with more people rewriting and translating all the events, descriptions, unit names, etc. the developer can then change the code.
EDIT: SDL does character recognition at the hexadecimal level. Instead of doing hex edits, why not just do all of the text files (in /db and /config directories) in UFT-8...it's a much easier way and supports all of AoD's languages.