FW: Updated status on RFC3023bis

I'm not sure I've fully digested this, but I want to make sure
others have seen it.

paul

-----Original Message-----
From: Chris Lilley [mailto:chris@w3.org] 
Sent: Friday, 2011 November 04 13:18
To: www-tag@w3.org
Subject: Updated status on RFC3023bis

Hello www-tag,

This week I briefly met with Henry and with Noah and explained some
recent and upcoming changes in RFC3023bis. But the details are better
committed to a public and archived email list.

As you probably are aware, there were objections by some implementors to
RFC3023bis deprecating text/xml on the grounds that 

- they still have to process it anyway (feeds, etc served as text/xml)
and 
-  their charset handling was not as defined in the spec (assume
us-ascii in the case of no declared charset at the http level) but was
the same as application/xml (assume what the xm document says to be
true, in the absence of charset info at the http level).

You will also be aware that we were unable to make RFC 3023bis do what
they wanted, due to the definition of the text/* types.

In the last year things have changed. I attended IETF80 and presented
RFC3023bis there, explaining the problems. Harald Alvestrand initially
disagreed with my reasoning and then, on closer examination of the
relevant RFCs, agreed that my analysis was correct.

That same day, HTTPbis resolved to remove the default charset for HTTP.
I suggested that one way forward (although not an ideal one) would be to
have different handling for http and for email. Alexey Melnikov,
outgoing IETF Apps area director, spoke with me on that and proposed to
fix the text/* handling so that http and email would once again be in
alignment. He also offered to become a co-editor of RFC3023bis, which is
great.

He has since published an internet draft proposing the text/* changes:
http://tools.ietf.org/html/draft-melnikov-mime-default-charset-01

Therefore, it seems that RFC3023bis could take a different approach:

- text/xml is no longer deprecated
- text/xml (and the other text/* types) would behave as application/xml
does

This would resolve the implementor objections and would also align the
specification with existing implementation practice in browsers and xml
parsers.

I still have to ensure that the spec meets the TAG requirements on
handling of fragment identifiers (I have followed the recent thread
regarding fragments in RDFa, for example) and they suggested that I meet
with you on Friday at noon to discuss it. But it does seem that we can
move forward here by removing the substantial and blocking issue.



-- 
 Chris Lilley   Technical Director, Interaction Domain                 
 W3C Graphics Activity Lead, Fonts Activity Lead
 Co-Chair, W3C Hypertext CG
 Member, CSS, WebFonts, SVG Working Groups

Received on Monday, 7 November 2011 14:30:37 UTC