John Cowan wrote:
> The new Unicode FAQ (like the old) supplies the panting world with
> John's Own Version of Unicode Conformance:
some of the old ones seem to be pre-unicode 1.1. should they not be updated?
> 1) Unicode code units are 16 bits long; deal with it.
this is true for the default encoding form, but not for utf-8 or utf-32. it is also
misleading for utf-16; yes, the code units are 16 bits long (or wide), but you need to
know that sometimes you need two for a code point. the above sounds like a statement
from ucs-2 days.
> 4) Loose surrogates don't mean jack.
this needs some explanation - they are illegal sequences, but should be passed through
for interoperability (i think that is what the book says). by itself, it is also
misleading.
markus
