- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Tue, 8 Apr 2014 09:50:34 -0700
- To: David Singer <singer@apple.com>
- Cc: Nicholas Doty <npdoty@w3.org>, "public-tracking@w3.org (public-tracking@w3.org)" <public-tracking@w3.org>
Regarding: >> A user agent that allows extensions to directly make or modify HTTP requests MUST provide a corresponding API to those extensions for determining the user's tracking preference. On Apr 8, 2014, at 8:31 AM, David Singer wrote: > something I am not clear about — was this introduction of a ‘must’ the consequence of a decision we needed to implement, or something you noticed and believed needed fixing? Something I noticed and have mentioned repeatedly since the very first meeting. In order for DNT to work, it needs to be expressed on every HTTP connection; otherwise, there will be a very big hole in our protocol. That's why we had a section on plug-ins. The WG did not make a decision to require anything. However, it also didn't make a decision to allow plug-ins to track independent of the browser's DNT setting. So, knowing how this stuff works, I picked the solution that I think the majority of the WG would prefer. I also think it is the easiest solution to implement (easier than having each plug-in maintain its own DNT and UGE settings) and does not require a specific API. Is there a hardship here that I am not aware of? I am also fine with anyone raising this as an objection so that the WG can make a formal decision with each of the alternatives adequately explained. However, the objection needs to be about the requirement itself, not about whether or not we can add it before the first last call. The addition is a consequence of our other decisions to require DNT be sent, once enabled, on all HTTP requests. Plug-ins are a part of the UA. ....Roy
Received on Tuesday, 8 April 2014 16:50:58 UTC