IPP Mail Archive: Re: IPP> MOD - Separate 'document-format' and 'document-language'

Re: IPP> MOD - Separate 'document-format' and 'document-language'

Ned Freed (Ned.Freed@innosoft.com)
Wed, 01 Oct 1997 08:06:16 -0700 (PDT)

> Thanks for clearing up all my muddledness. And for pointing out the
> fact that MIME is written intentionally from an open-ended perspective.
> I'm sorry that my background is almost wholly in ISO and ITU-T specs,
> and I'll failed to suspect that MIME was written from a different
> slant. And I didn't realize that the IANA character set registry
> didn't exist as recently as RFC 204x series.

Actually it is worse than this. The registry exists, but there are no rules
governing it at present. We're working as hard as we can to get the necessary
rules in place to fix this. (I've posted two revised drafts this week alone,
and I'll probably be posting a third.)

> So as a co-editor, you have told us that any IANA registered character
> set may be used in the 'charset' parameter of 'text/*' MIME types.

That's what RFC2045-2049 say. However, it turns out that this can be somewhat
problematic, in that MIME imposes some restrictions on the use of certain
characters in MIME text, e.g. CR, LF, and NUL. Some of the charsets registered
currently aren't compatible with these restrictions (e.g. various forms of
EBCDIC.)

There are also problems with the names of some charsets and the restrictions
MIME imposes on the use of names (in particular MIME part III and the new
parameter value in RFC2184). Quite a few registered charsets have primary names
that don't meet these requirements. I recently added ABNF to the registry
document that spells out the restrictions quite explicitly for any charset
intended to be compatible with MIME.

All of this is being dealt with in the new charset registry work. Since we
cannot unregister a charset (nor would we want to -- there are more uses than
just in MIME for this stuff), we've added a "eligible for use in MIME" criteria
instead. Old registrations can then be evaluated as to their appropriateness
and can be flagged accordingly. We'll also be rearranging the names of
registrations to make as many of them as possible be usable within MIME.

> Could the character set policy that Harald is hard at work on
> possibly just SAY that, in the interim until the MIME RFCs are
> updated (perhaps quite a bit later), so us chickens would get the
> message clearly?

Well, the hope is to publish Harald's new document alongside the new
registration procedures document. So this really should not be necessary.
However, if for some reason we cannot get the registration document out I
agree this would be a good thing to do.

Ned