On Fri, Dec 08, 2000 at 04:13:43PM +0000, Nick Ing-Simmons wrote: > There is no difference for the old model. There is a difference in the > new model. No, no, no, no, no, no. The "old model" is what's going on now. Latin1 <-> Unicode works fine. The "new model" I was proposing simply extends that to other character sets. (Not other *encodings*, which is why your BIG5 and SJIS examples are irrelevant.) This morning, I had a decent solution to getting EBCDIC on Unicode, and now I'm so bloody confused I don't know what the hell's going on. I give up. -- Momomoto, Famous Japanese, can swallow his nose.