As for the case of new character addition to legacy encodings you mentioned below, The real problem emerges when the older version of receiving application has received new-version-legacy-encoded IDN(IRI) from the new version of sender application and it tries to convert that received IDN(IRI) into unicode using the *OLDER* version of legacy-2-unicode convesion table, but surely it will fail in some cases.
Soobok Lee ----- Original Message ----- From: "Doug Ewell" <[EMAIL PROTECTED]> > > As for additions, they shouldn't cause a problem anyway, because they > don't break existing legacy-to-Unicode mappings. An often-mentioned > case of adding to a legacy encoding was when Microsoft retrofitted > U+20AC EURO SIGN onto their Windows code pages (mostly at previously > unassigned code position 0x80). The only people who suffered at all > were the ones who thought "unassigned" somehow meant that they should > map 0x80 to U+0080. Everyone else made it just fine. > > -Doug Ewell > Fullerton, California > (would prefer to receive only one copy of these messages) >
