L2/08-021

Source: John Cowan
Date: 15 Jan 2008
Subject: Time to freeze NPC and canonical equivalency

I think the time has come in Unicode development to declare that there
will be no more characters encoded which are canonically equivalent
either to existing characters or to characters encoded in the same batch.
That means that the normalization forms NFC and NFD will be frozen from
now on.

This will be very useful for people (like the IETF) who want to
specify things in terms of NFC, but who are only annoyed by the largely
hypothetical possibility that new characters will be encoded in future
only to be NFC-normalized away again.  Indeed, I have seen several cases
where characters were proposed and informally rejected, then proposed
again with canonical equivalents, and finally dumped because the canonical
equivalency makes the character in effect useless.  (I can't cite cases
offhand, so I hope I am not talking out of my rear end here.)

As far as I can tell, the last time any characters with canonical
decompositions were added to Unicode was almost seven years ago in 3.1,
when we added 13 musical symbols and a bunch of Plane 2 CJK compatibility
characters (which despite their names have canonical decompositions).
I think it's time to say we don't need any more of those.