Re: Content Sniffing impact on HTTPbis - #155

Adam Barth wrote:
> On Fri, Jun 5, 2009 at 7:10 AM, William A. Rowe, Jr.
> <wrowe@rowe-clan.net> wrote:
>> Server misrepresentation of Content-Type will cease, once browsers stop
>> misrepresenting the content type.  Until web authors and administrators
>> (including mass vhosters) become aware that they have misrepresented the
>> data they are serving, they will continue to generate the 3% (IIRC) of
>> mislabeled content.
> 
> This is unlikely to ever occur given the market dynamics of browsers.

Which is what makes the entire discussion so entirely laughable.

I concur with Roy and Julian, leave Content-Type null with an undefined
content type.  while user agents persist in nonsense (such as decoding
UTF-7 when presented an explicit charset), this is out of our server-side
and authors' hands, and no spec is going to correct their misbehavior or
unsafe practices.

For some tiny minority who care, the very absence of Content-Type conveys
new, useful metadata that should not break 2616 definitions.