RE: Fwd: Last Call: UTF-16, an encoding of ISO 10646 to Proposed

[Ira Mcdonald <imcdonal@sdsp.mc.xerox.com>]

Hi folks,

Harald did a good job of concisely listed disadvantages of
UTF-16 over the wire and one (possible) advantage.  

Since Informational has a long history of being used to publich
'for the benefit' of the Internet community' standards from some
other standards body or vendor, please let's move forward and
register it as Informational.

I'd be delighted to see the sense of Harald's list of
disadvantages captured in the actual IANA charset registration,
so that implementors are sure to see the 'BEWARE'.

In a local file store, UTF-16 has some questionable advantages.
Over the wire, it just decreases significantly the probability
of interworking.  The IETF and the UNIX operating system
community had it just right when they chose to standardize
on UTF-8 based interfaces.

I also agree with Martin's sentiment that we should just
get this done and published as an RFC.  There is a lot
of ambiguous XML leaking into the real world by unsafe
interfaces.

Cheers,
- Ira McDonald

PS - I think that Harald's point about having terrible
expansion for characters outside of Plane 0 (BMP) is
very important over the wire.  Unicode 3.0 has come
markedly closer to *filling* the spare space in
Plane 0.