Re: Registration of new charset [ISO-2022-JP-2004]

[Erik van der Poel <erikv@google.com>]

Sorry, that was confusing. Implementors that are concerned about
ancient software will not be using the new escape sequences (and new
characters) anyway. They will probably use iso-2022-jp, which is
already described in RFC 1468.

I still wonder whether we should mention format=flowed in the
iso-2022-jp-2004 registration, as an alternative to quoted-printable
when lines (or paragraphs) are long. If we do mention format=flowed,
we may also wish to describe the Japanese line breaking rules to some
extent. Then again, maybe all this is too much for an IANA charset
registration -- maybe it belongs in an RFC?

Erik

> > Suitable for 7-bit use in MIME body-part as text/plain or text/html.
>
> This wording might be misinterpreted to mean that iso-2022-jp-2004 can
> only be used with text/plain and text/html. I believe this charset is
> suitable for use with any subtype of text (text/*).
>
> Also, I wonder if we should say anything about
> Content-Transfer-Encoding. The quoted-printable encoding does not look
> good at all in very old (pre-MIME) implementations (and is not
> necessary when the lines are short enough). The format=flowed
> transformation would work fine, even in ancient software. (Do we care
> about pre-MIME implementations?)
>
> > B-encoding is recommended for use in MIME header-part, because
> > ISO-2022-JP-2004 is a partial extension of ISO-2022-JP.
>
> The term "header-part" is not normally used for these. They are called
> encoded-words. How about: "The 'B' encoding is recommended for use
> with MIME encoded-words in headers, as is recommended for the related
> charset iso-2022-jp."
>
> > - ESC ( B      for ISO/IEC 646 IRV
>
> I wonder if we should add JIS X 0211 here. This set is similar to (if
> not identical to) ISO 6429 control characters in the range 0x00 to
> 0x1F. The most commonly used controls in the iso-2022-jp family are
> horizontal TAB, CR, LF and ESC. SI and SO _must_ not be used.
>
> Erik