Re: UTF-7 is dead

From: Joerg Knappen (
Date: Wed May 26 1999 - 07:00:41 EDT

Pete Resnick schrieb:

> a) Most environments are now 8-bit safe, and more are moving that way
> all of the time.

It is less than clear, what the term "8-bit-safe" means. In practice it
often means: 8-bit-safe except C0 controls AND C1 controls. Fully 8-bit-safe
text transports are still rare (and will remain so, I'm afraid).

> b) In those cases where only 7-bit is available, UTF-8 (which is a
> charset, not an encoding) can itself be encoded as quoted-printable
> or base64.

Not a real solution -- in this case one can as well encode UCS-2 as base64.

There are some proposals, however:

Unicode technical report No. 16 (
contains a proposal backed by IBM for uft-8m without C1 controls and
utf-8-ebcdic for passing utf-8m octetts thru EBCDIC hosts.

There is also utf-7d5 (see:
and The main difference
is, that utf-7d5 only aims at ucs-2, while utf-8m covers full ucs-4 (31bit).

--J"org Knappen

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:46 EDT