W3C home > Mailing lists > Public > public-tracking@w3.org > June 2012

Re: tracking-ISSUE-150: DNT conflicts from multiple user agents [Tracking Definitions and Compliance]

From: Rigo Wenning <rigo@w3.org>
Date: Tue, 05 Jun 2012 19:38:19 +0200
To: public-tracking@w3.org
Cc: Justin Brookman <justin@cdt.org>
Message-ID: <15578458.9z8cIT0bIc@hegel.sophia.w3.org>
On Tuesday 05 June 2012 12:45:25 Justin Brookman wrote:
> For purposes of argument, substitute NACK with a different
> response  header for "Invalid User Agent" or "Non-Compliant User
> Agent"that Shane suggested.

IMHO, if a server can decline (NACK) for any reason, why would they 
speculate on whether a User Agent is invalid (whatever that means 
and who is determining whether it is a valid judgment). 

Because of stupid web developers, I'm regularly forced to fake my 
user agent string. I know I'm a browser-weirdo, but the browsers I'm 
using usually work. Now imagine, a site forces a certain string to 
be used (works only with IE 10) and then says: gotcha -> invalid 
user agent!

This is only one of many frictions and collateral damages that we 
create by implicitly trespassing a user preference into non-
refutable service(server) obligation and by trying to solve the 
undesirable effect by transgressing server-side concerns back into 
user agent configurations. Instead, we should go back to the roots 
of the meaning this all started with: 

user -> DNT;1? -><- DNT;1 <- Service

If one side is missing or NACK, the protocol is validly executed but 
yielded no conclusive result.

user -> DNT;1? -><- DNT;NACK <- Service
just says No. Or it would force the service to contract under the 
user's conditions. That would be a strange result.

All IMHO and as a pure contribution to the discussion. 

Rigo
Received on Tuesday, 5 June 2012 17:38:49 UTC

This archive was generated by hypermail 2.3.1 : Friday, 21 June 2013 10:11:30 UTC