W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 1998

Re: Referer [sic] Redirect question

From: Koen Holtman <koen@win.tue.nl>
Date: Tue, 15 Sep 1998 19:23:43 +0200 (MET DST)
Message-Id: <199809151723.TAA01322@wsooti08.win.tue.nl>
To: "Jaye, Dan" <DJaye@engagetech.com>
Cc: http-wg@hplb.hpl.hp.com
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/17
Jaye, Dan:
>How should the user agent populate the referrer field when issuing a
>request after receiving a 3xx redirect response?  Is the original
>referer maintained or is it now the URI that resulted in the redirect
>response?  There is no discussion of this scenario in the referer or
>redirect sections of the draft.

I think that the language "allows the client to specify [...] the
address (URI) of the resource from which the Request-URI was obtained"
in the draft clearly implies that it should be the URI that gave the
3xx response.  I would not be surprised however to see some 1.0 agents
do it the other way around.

>Reason for the question:
>Ad Servers sometimes delegate remnant ad impression inventory to
>"fire-sale" ad networks using re-direct.  Usually an ad server will use
>the referer field to determine on what page the ad will be appear.

Using referer in this case is a dangerous technique for the second ad
server.  It definitely breaks with browsers that do not send referer
at all.  It is better to have a setup in which the first ad server
encodes the necessary information in the request URI for the second ad

Received on Tuesday, 15 September 1998 10:27:26 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:43:05 UTC