Re: Registration of new charset

Chris Newman (Chris.Newman@innosoft.com) wrote:

>UTF-7 violates section 3.4 of the registration procedures because it
>is a character encoding scheme which includes a content-transfer-encoding.
>Base64 is clearly better thought of as a content-transfer-encoding.
>Because of this layering violation, UTF-7 can not take advantage of the
>8BITMIME infrastructure which is widely deployed in email.

When UTF-7 was first being designed, I asked if it should be done via a 
content-transfer-encoding, and the response by representatives of the 
IETF and the original authors of MIME was that a CTE which only worked 
for exactly one MIME type and character set was not appropriate. That's 
why it was done the way it was. You could argue by the same logic that 
ISO-2022 includes a CTE, since it encodes something with escape sequences 
and in 7-bit form and thus cannot take advantage of 8BITMIME. The fact 
that UTF-7 uses the Base64 algorithm for encoding is really irrelevant; 
the fact that a CTE uses Base64 and UTF-7 also uses Base64 does not make 
UTF-7 a CTE.

>In addition, UTF-7 is not significantly different from UTF-8 when used as
>a MIME charset.  UTF-8 is currently the preferred international charset by
>IETF policy and charsets with duplicate functionality are not desirable.

I don't think this registration contradicts that policy, any more than 
any other registration of a non-UTF-8 character set does. The Unicode 
Technical Committee is also planning to register UTF-16 for non-mail 
applications, by the way.

>Since UTF-7 is being deployed in products, I conclude that it should be
>labelled LIMITED USE per section 3.5 of the registration procedures.

I have no problem with labelling it limited use. Remember, it is used in 
PICS and (in variant form) IMAP4, not just mail. It is important to have 
a registration for this character set so people can use it (since they 
already are). I am in no way evangelizing people to use it for new 
applications.


David Goldsmith
International, Text, and Graphics Department Architect
Apple Computer, Inc.
goldsmith@apple.com



--Boundary (ID uEbHHWxWEwCKT9wM3evJ5w)

Received on Wednesday, 18 February 1998 11:08:29 UTC