Re: ACTION-211 Draft text on how user agents must obtain consent to turn on a DNT signal


On Thursday 14 June 2012 09:39:16 Dobbs, Brooks wrote:
> > trouble is that IE 10 is not non-compliant for all possible
> > cases. There are tools that are non-compliant for all
> > possible cases.
> I am not sure I agree with that statement.  If IE10's compliance
> job is to communicate user preference in a manner that is
> discernable to a server, when does it achieve this end? 

Our problem is precisely that we have no way to know whether a DNT:1 
reflects a user intent. But this cuts both ways. From a bit-counting 
perspective, the header is a preference that is discernible by a 
server. So we have no additional guidance by that phrase.

> As has
> been pointed out, IE makes it impossible to discern between a
> DNT:1 that is or is not an expressed user preference.

You conclude without assumption :) We argue what to do in case we 
have a protocol violation. I suggest NACK and some suggest to send 
"I hate a certain browser". For the sake of decency, I would prefer 
the simple message as suggested by David Singer here:

Going beyond would likely trigger more concerns and meta-troubles in 
the process downstream. 

BTW, I was at the OBA meeting in Brussels and Madelin explicitly 
supported the idea of having a user choice menu during 
download/installation/first start-up If we would all agree on this, 
that could be a way out of the trenches. 

I maintain that the system is not complete if the exception 
mechanism doesn't work. Again, this is also a way to say no.

Has anybody an idea whether IE10 will support the exception 


Received on Thursday, 14 June 2012 16:06:04 UTC