W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1997

Re: Please clarify * is legal in HTTP1/1 Accept-Charset Header

From: Koen Holtman <koen@win.tue.nl>
Date: Fri, 10 Jan 1997 19:53:59 +0100 (MET)
Message-Id: <199701101854.TAA02122@wsooti05.win.tue.nl>
To: Chris Lilley <Chris.Lilley@sophia.inria.fr>
Cc: koen@win.tue.nl, ftang@netscape.com, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, Alan_Barrett/DUB/Lotus.LOTUSINT@crd.lotus.com, Chris.Lilley@sophia.inria.fr, Ed_Batutis/CAM/Lotus@crd.lotus.comi18ngrp, bobj@netscape.com, mduerst@ifi.unizh.ch, www-international@w3.org
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/2283
Chris Lilley:
>I would urge you to add additional wording such that, in the absence of an
>explicit q factor, the wildcard has a default, low q factor such as 0.1.
>This applies to Accept and Accept-Language as well. It would correct a
>sitiation where browsers send a list of things they do accept, followed
>by *, and this is (currently) often taken to mean exactly the same as
>if they had only sent the *.

In my opinion, this sitiatiomn should be corrected by changing the browsers
to send good accept headers, not by changing the spec to mandate some kind
of workaround in the server.

We could add a note to the spec telling that many existing 1.0 clients
erroneously send `Accept: a/b, c/d, *' when they mean `Accept: a/b, c/d,
*;q=0.1', and that 1.1 allows servers to work around this error.  But that is
about as far as I would like to go.

>Chris Lilley, W3C                          [ http://www.w3.org/ ]

Received on Friday, 10 January 1997 10:56:42 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:40:19 UTC