There are several problems with old codepages, that is why unicode was invented in 1988.
-People who know different languages as GerdtR pointed out.
I can read/write dutch, french, english, which is all no problem in one codepage but I have also learned to read russian and even understand some of it. Remember Yogev Ezra? He speaks russian, hebrew and some more...
-Programs would need support for different codepages and even be able to convert from and to them.
For example, most of the internet nowadays is UTF8 (hooray!).
-We already have capability to print unicode (and so UTF8) text in kernel, so why not actually make things simpler and only use that instead of archaic codepages...
I understand you, but....
I think, for the run of universality of UTF8, we'll lost a lot of children in the bath.
More important things awaiting' us than rewriting all the old programs.
So, good support old CP's is better than new troubles, but
writin' new sysfn's and libraries must have mbcs updates in mind