In a message dated 2001-06-25 20:19:18 Pacific Daylight Time, [EMAIL PROTECTED]
writes:
> (For instance, I
> don't see how it would be possible to encode a sequence of unicode
> scalar values corresponding to a low and a high surrogate; if you
> tried to map this back then you would get a single unicode scalar
> value outside of the BMP). Perhaps someone on the unicode list could
> elaborate?
This is the source of my remaining confusion about definition D29. It
requires UTFs to round-trip all Unicode code points, and by extension all
sequences of code points; yet if you use UTF-16 and start with the sequence
<D800 DC00>, you don't end up with that -- you end up with <10000>.
The way it was explained to me on this list made it sound as though UTF-16 is
the "master" UTF that other UTFs have to accommodate. That didn't make sense
to me, but I've been trying to cope with it.
Proposed UTFs that are based on UTF-16 code units, and are thus subject to
the same D29 limitations as UTF-16, really annoy me, though.
-Doug Ewell
Fullerton, California
