All the other RC files I found are saved with ANSI (ISO-8859-1) encoding. Therefore we should save them all with this encoding. If a RC file still uses another encoding than ANSI, this should be changed.
That's not reasonable, you would loose all Unicode characters except Latin 1 ones. So please don't save as Latin 1 (aka ISO-8859-1) except for special cases (languages like EN, DE, and other european lnaguages).
Use UTF-8 or even more WinNT alike UTF-16. For eastern languages UTF-16 is the prefered way, as UTF-8 would waste more space.
Unicode in Unix is rather a "new" thing, and to keep the apps code almost the same, Unix coders choose UTF-8, where the first 128 bit are the same as the ASCII charset. Although Unix apps with UTF-8 support usually don't add the BOM as leading chars.
P.S.: we have several language files in ros that are stored with chars encoded in one of the Unicode Transformation Formats.