Re: comments on draft-barth-mime-sniffing

>  ... they seem remarkably resistant to pressure..

The best way is to show them exactly what you wish to be done.
One step in the cycle is getting the servers to provide an expected or 
at least approriate content type with the content.
I think in some servers this may be done by some level of automation 
that follows the main MIME list. In general, the community process 
keeps track of mimes that people and organizations are interested in 
registering for the maintained list. Then, I think servers are free to 
update themselves re the 'official' list. Usually it seems to take a 
while longer for the mime to be offical than for content to be 
published. In this time, the server guy can easily configure the 
server to deliver most any mime for any file extension as the web 
author wishes.

Browsers know about MIMEs and the connection with file extensions or 
other content model definitions when some are native and some are 
added by all sorts of web features and applications that get loaded 
into people's computers. I think that when content publishers are 
convinced that the W3C host browsers care and are actually using the 
things to add some production value (or even security and reliability) 
they will be responsive. But these are  seasoned producers and you may 
have to go there and tell tell tell them using actual representative 
working examples of flexible and highly reusable cross-browser 
behaviors using HTML 5 code and content.

For instance, I think text/html uses the mime definition to help 
because the mime helps predict that the sniffer may find a content 
type string somewhere up close to the beginning. Tool makers and 
authors use special and standard mimes, extensions to mimes, and file 
extensions for many reasons, even to distinguish 'public' from 
'special' content.

Thank You and Best Regards,
Joe
"There is good stuff built into the MIME scheme of doing things that 
we don't even know about anymore."

Received on Friday, 19 June 2009 18:57:38 UTC