Re: Modifying a DNT Header (ISSUE-153, ACTION-285)

On 9 nov. 2012, at 20:54, Shane Wiley <> wrote:

> 'For the sake of moving this process forward.'
> I believe that's the core issue you're missing.  If there is no way to solve for this issue (disagree this is a Halting Problem as we're only suggesting a UA protect its own settings, not attempting to understand another's application/settings), then DNT will mostly likely NOT be implemented in the real-world.

First of all, I have suggested language about what reasonably can be expected from a UA in this regard. It is just that I see no point in going further, because you cannot guarantee an unaltered DNT signal leaving the endpoint. You can make an effort to verify changes you notice, but can't do anything about what you cannot even notice.

Asking for more will result on a standard that is impossible to comply with for UAs. Given Yahoo!'s position that it feels entitled to ignore DNT:1 signals from non-compliant UAs you may be creating a Catch-22 here.

Regarding the impact of this problem on industry's willingness to implement it: this process is based on the assumption that a DNT:1 signal will be honoured. That requires trust from users in servers. Trust is only sustainable when mutual. It therefore stands to reason that a server trusts a DNT:1 signal it receives. 

For the record: I suggest that this principle of mutual trust is to be incorporated in the standard and that we move on to things that are feasible.



Received on Friday, 9 November 2012 20:09:10 UTC