W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2008

Re: PROPOSAL: i74: Encoding for non-ASCII headers

From: Martin Duerst <duerst@it.aoyama.ac.jp>
Date: Fri, 28 Mar 2008 15:04:12 +0900
Message-Id: <6.0.0.20.2.20080328145346.0662f2d0@localhost>
To: "Roy T. Fielding" <fielding@gbiv.com>
Cc: Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>

At 02:19 08/03/28, Roy T. Fielding wrote:

>In any case, UTF-8 would have been chosen today if we could get
>four people to implement it instead of just talking on IETF lists.

I'd be glad to help with implementing. I'm just still not sure what
you mean by "iso-8859-1 is implemented". After repeted calls for
examples on this list, we have finally received a single example
(Spanish with a wrong accent). And you have said yourself that
HTTP mostly cares about the bytes, which means that in terms of
actual implementations, it's 99.9% or so byte-in-byte-out, and
maybe occasionally in some corner there is a tiny bit of iso-8859-1.

If it's about a server sending some iso-8859-1, it's easy to set
up a server that sends out UTF-8. If it's about a client interpreting
something as iso-8859-1, it's a bit more difficult but still possible
to change that client to use UTF-8. As my main proposal is to
allow definers of new headers to choose UTF-8, I'd probably
implement a new Foo: header or so just to prove a point.
If you can point to the location(s) in Apache httpd where
iso-8859-1 in headers makes a difference, please do so.


>The standard specifies iso-8859-1 because that is what all
>implementations implemented at that time, not because it was the
>most popular choice of standards mavens.  I don't care what the
>encoding of TEXT is in HTTP so long as there are implementations
>that are interoperable.

There is obviously quite a bit of circularity here. The first version
of a spec says it's iso-8859-1, that's what people implement
(if they do), then the second version says that's what we have to do
because that's what the implementations do, and so on. That never
allows e.g. a new header to choose something easier and simpler.
What would be your conditions to break out of this lock-in?

Regards,    Martin.


#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst@it.aoyama.ac.jp     
Received on Friday, 28 March 2008 06:17:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:37 GMT