The problem of non-English characters has been discussed here for many years ("ad nauseam", to quote Artful Codger). The current "solution" has some advantages and drawbacks, as stated many times. If it is guaranteed to be in force till the end of our days, postings in German, French, Spanish and Italian are safe.

Most other languages miss out, even if based on Latin characters such as Polish. Their users are not any worse off than they used to be, though, if they do the right thing using HTML escapes. However, those who sinned earlier will now no longer be able to read their own old postings correctly, and those of their compatriots (in terms of codepage).

The same applies to genuinely "special" characters such as ♫. (I produced this by entering "♫"; if I enter the character directly it still results in: ?.) This notably includes the automatically prettified apostrophes from some "smart"phones. From this point of view, switching to UTF-8 (as Max tried out and rejected before) was a better idea.

For other ideas see those old threads such as 135626.