On 04.04.2026 06:19, Doug Ewell via Unicode wrote:
Sławomir Osipiuk wrote:
should we ever find ourselves in a world that needs a bigger Unicode
AND still uses UTF-16.
My hope is that, if the world ever does need more than 15 non-PUA planes for
whatever reason, it will have moved on from UTF-16 by then, and the choice can
be between -8 and -32.
IMHO in that case we should have a "new-Unicode". Let's be realistic: in
such case using actual Unicode will cause many problems (not just the
encoding): we may save some bits, but programs will be complex on
handling properties (with a lot of changes on each version, else we will
not reach such point). Without entering on the rendering (which the
render/shaping engine, and hopefully the font system, must know the
Unicode version of the text), we must have some canonization (which it
is not "stable" in respect not-yet published Unicode, if we find future
codepoint. Very common in our phones, computers, emails. Also security
needs to know some Unicode properties. Modifiers and combining
characters are also useful for programs (without need to know the
properties or renderings).
So (but not in my lifetime), I think we should have some properties
encoded in codepoints (e.g. a (or more) plane for LR, one for RL. Also
combining characters should be split. Modifiers (hidden one, like
variant selectors, ZW(N)J, etc.)
And compatible (maybe also binary for some encoding) with at least Basic
Plane. Good luck! Let's hope we do not ever need it. And also let's hope
that we will stabilize (far less additions) characters on some not so
far future. History is not so long (history as the period with written
records), there would be addition, but hopefully few.
giacomo
