- From: Matthias Schunter <mts@zurich.ibm.com>
- Date: Thu, 19 Jan 2012 15:03:10 -0500
- To: "Karl Dubost" <karld@opera.com>
- Cc: public-tracking@w3.org
Hi Karl, The points I tried to make here is that A) it is testable whether a response header has been sent B) it is testable whether this response header contains a field like 'I promise to implement DNT and I will adhere to the requirement put forward in the compliance spec'. C) It is not testable from the outside whether a site in fact adheres to these promises. Eg if we were to require 'no recording if IP addresses', compliance is hard to test from the outside, Matthias -original message- Subject: Re: SHOULD or MUST for responses to DNT;1? From: "Karl Dubost" <karld@opera.com> Date: 2012/01/19 19:52 Le 19 janv. 2012 à 13:41, Matthias Schunter a écrit : > It is not testable in any case. Without response, even a claim to follow is > no longer visible. huh? :) A MUST conformance statement has to be testable. If you are saying that an implementer can NOT test the MUST, then the MUST is meaningless. Or I have misunderstood what you are trying to explain. -- Karl Dubost - http://dev.opera.com/ Developer Relations, Opera Software
Received on Thursday, 19 January 2012 20:03:53 UTC