- From: David Singer <singer@apple.com>
- Date: Tue, 18 Sep 2012 09:45:17 -0700
- To: Rigo Wenning <rigo@w3.org>
- Cc: David Singer <singer@apple.com>, public-tracking@w3.org, "Roy T. Fielding" <fielding@gbiv.com>, Nicholas Doty <npdoty@w3.org>
On Sep 18, 2012, at 1:23 , Rigo Wenning <rigo@w3.org> wrote: > On Monday 17 September 2012 14:13:22 David Singer wrote: >> I dealt with most of this issue by requiring that exceptions are >> granted in their entirety or not at all. > > This creates a conflict with pre-existing preferences (web-wide). Why? I think if you read the spec., it's clear what header gets sent. > We > either need a rule to override them (1) or we need a means for the > site to respect them(2). > > (1) I would suggest to follow the more than thousand year old rule > that the specific trumps the general (roman rule). Means a site-wide > exception or DNT:1 would trump all web-wide preferences. This is a > very simple rule for Browsers. Exceptions trump general preference, in general. There is no 'reverse exception' at the moment, so if the site has either a web-wide exception, or a site exception under the current site, they dnt:0, else they get the general preference. > (2) Allow the first party to access site-wide information so that > she can respect the preferences and not send info to a given third > party that gets a DNT:1 This is more friendly to Shane's use case. We could have an API that allows finding all granted exceptions for the current site, sure. David Singer Multimedia and Software Standards, Apple Inc.
Received on Tuesday, 18 September 2012 16:46:14 UTC