I plan to support multiple languages with my OS.

Unfortunately, I have no idea what is required in order to do this. I know almost every text file and html page that I come across in English uses only 8 bit characters.

To support the multiple language, I figured I would go 16-bit characters for everything(I think this is called unicode). But, I am not sure if this will do the job, or if I should use 8 bit characters and code pages (whatever that is).

Any help would be appreciated. I even plan to do language support for my function call names, so it will be pretty engrained into the OS.

...Gomer73

Maybe you could use UTF-8 char format? it uses different char width: normal (english) characters are encoded as usual, while the other characters (like ą, ś, ń, ź in Polish) are encoded in two or more bytes. This system is good, as it saves place, and makes life a lot easier when working with "standard" characterset. However, things like StrLen routine are more complicated... anyway it is a good system, and you should consider implementing it.
I can't help you much with more detailed explanation, as I was only using routines that deal with UTF-8, and never coding my own. Take a look at Allegro Library http://alleg.sourceforge.net/, it has a rich set of UTF-8 routines (AFAIR they are coded using assembly.

regards,
Mateusz

Thanks, UTF-8 looks like the way to go with my OS.

I can still use 0 for string termination. Also I didn't realize that unicode was greater than 64k, so even two bytes isn't sufficient.

This way I can code my source in any editor and not have to worry. Otherwise it would be kind of a pain trying to save in two byte format.

Takes up a little more space for Asian characters, but oh well.

UTF-8 would be the best choice, I think. It's the application task to support 8-bit code pages when editing/viewing text files (there are more text files that are still encoded in native code pages) . UTF-8 is perfect. However to write a correct FAT12/16/32 driver you'll need in-kernel code pages support in order to encode old dos 8.3 file names with 8-bit code pages, use something like DosCP variable, as I plan to use in my kernel.