a unicode interface isn't exactly necessary. You have the raw data in bytes and you can do conversion to whichever encoding you want with Windows API functions like
MultiByteToWideChar. For now, the entire scripting language isn't even unicode aware-- so sometime in the very faraway future, that will happen and Khaled can make a more convenient interface for dlls. Convenient, but by no means necessary.
To directly answer your question: Khaled doesn't usually divulge future plans, but I can bet you lots of money the DLL interface won't change until there is a real need to do so... that probably won't happen for a while.