Re: [I18ndir] [art] Fwd: New Version Notification for draft-bray-unichars-06.txt

["Manger, James" <[e-mail]>]

Thought experiment: Unicode is invented is almost the same way, except UTF-16 is slightly rearranged. D800-DFFF are still high/low surrogates. The code-point to/from code-unit calculation differs.
  1 code unit X in [0000, D7FF] => code point = X
  1 code unit X in [E000, FFFF] => code point = X – 2048
  2 codes units X Y a surrogate pair => code point = (X – 0xD800) * 1024 + (Y – 0xDC00) + 0xF800
  Unpaired surrogate => ill-formed

The result is exactly the same number of scalars, but they are now contiguous %x0-10F800. So scalars = code points.

JSON is identical, including escapes using UTF-16 code units. You can still get "\uDEAD" as a JSON text. You still have to deal with that (and explain it if desired).

But you can no longer use “code point” as a label for unicode-plus-unpaired-surrogates. You have to explain its actual nature.

P.S. Of course I am NOT suggesting any such change. But I think it hints at how “code points” is misused to straddle the scalar/code-unit divide is a way that causes confusion.

--
James Manger


General