On 8/24/2011 10:48 AM, Richard Wordingham wrote:
Those are two different claims. 'Never say never' is a useful maxim.
So is "Leave well enough alone."
The problem would be in using maxims instead
of an analysis of engineering requirements to drive architectural decisions.
The extension of UCS-2, namely UTF-16, is far from optimal, but it
could have been a lot worse - at least the surrogates are contiguous.
All I ask is that we have a reasonable way of extending it
Why?
if, say,
code points are squandered.
Oh.
Well, in that case, the correct action is to work to ensure that code
points are
not squandered.
I think, however, that<high><high><rare
BMP code><low> offers a legitimate extension mechanism
One could argue about the description as "legitimate". It is clearly not
conformant,
and would require a decision about an architectural change to the standard.
I see no chance of that happening for either the Unicode Standard or 10646.
that can
actually safely be ignored when scattering code assignments about the
17 planes (of which only 2 are full).
A quibble (I know), but only 1 plane is arguably "full". Or, if you
count PUA, then
*3* planes are "full".
Here are the current stats for the forthcoming Unicode 6.1, counting
*designated*
code points (as opposed to assigned graphic characters).
Plane 0: 63,207 / 65,536 = 96.45% full
Plane 1: 7497 / 65,536 = 11.44% full
Plane 2: 47,626 / 65,536 = 72.67% full (plane reserved for CJK ideographs)
Plane 14: 339 / 65,536 = 0.52% full
Plane 15: 65,536 / 65,536 = 100% full (PUA)
Plane 16: 65,536 / 65,536 = 100% full (PUA)
--Ken
