W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1996

Re: Non Latin1 charsets (draft-holtman-http-negotiation-00.txt)

From: Nickolay Saukh <nms@nns.ru>
Date: Fri, 01 Mar 1996 22:42:36 +0300
To: Daniel DuBois <ddubois@rafiki.spyglass.com>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Message-Id: <16295.825709356@Alyona.office.nns.ru>
> Charsets are not to appear in mime type tags in URI/Alternates headers.
> They ave their own slot.

Holtman paper says (4.6 Alternates):

	The type, language, encoding, and length attributes of an
	alternate description refer to their Content-* header
	counterparts.

Content-Type has charset for text/html entities (for iso-8859-1
it is implicit). Where is own slot for charset? As an extension
postponed till HTTP/1.2?

> If that is not sufficient, we now that we have Koen's method for reactive
> negotiation, by which you will be able to precisely pick which varaint you
> want, which might be useful if you recieve the the language you wanted, but
> not the charset you wanted. (You recieve TheProject.ru.html, but you realize
> there was also a TheProject.ru2.html, so you ask for it by name.)

An example

	Accept-Language: ru, *;q=0
	Accept-Charset: iso-8859-5, koi8-r, unicode-1-1-utf8

Server has alternates with all charsets.  By current papers all my
alternates has the same quality factor.  With what charset I would
receive document, if any, with preemptive negotiation?  In what order
alternates will be present to user agent for reactive negotiation?  Per
drafts the order is significant, because the first alternate is the
best one. Why not to have quality factor charset? Like this

	Accept-Language: ru, *;q=0
	Accept-Charset: koi8-r, iso-8859-5;q=0.8; *;q=0
Received on Friday, 1 March 1996 11:48:19 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:47 EDT