Re: review of content type rules by IETF/HTTP community

Hi Roy,

On Aug 25, 2007, at 3:23 PM, Roy T. Fielding wrote:

>
> On Aug 25, 2007, at 1:10 PM, Magnus Kristiansen wrote:
>> If you mean patches to add types, much of it has already been  
>> reported (in
>> vain). [1] [2] [3]
>
> Apache does not list unregistered types in mime.types.  How many
> times does that need to be repeated?  If you want Apache to do  
> something
> nonstandard, then there are many ways to configure your own server in
> a nonstandard way.  Otherwise, take the half hour required to submit
> a message to IANA requesting the assignment of a type.  It is a lot
> faster than getting Apache to add types that are explicitly forbidden
> by IETF standards to be used by distributed software.

Regarding the use of registered MIME types and the earlier reference  
Apache bug[1]:

Why not simply configure Apache by default to return a content type  
of 'unknown' or 'x-unknown'. I just changed my own local Apache on my  
Powerbook to return a default of unknown and then Safari tries to  
download the file (an HTML document).

I wonder whether this could be solved by simply registering the  
unknown MIME type with IANA; updating the default Apache  
installations; and  perhaps amending HTTP to permit UA  sniffing for  
"unknown" as well. Is there something to be gained by sending no  
content-type header as opposed to sending "content-type unknown'?  
Just curious especially considering how Apache seems afraid to fix  
this on their end. With a registered root MIME type of 'unknown',  
this would also significantly raise awareness about the issue for  
current installations

Its interesting here too in that Apple, ,who is both an Apache vendor  
and a browser maker,  includes 'DefaultType text/plain' in the  
default configuration.

Incidentally, I've also started a new wiki page on this issue[2]  
(though still in a fairly rough form). My goal with this page is to  
focus the discussion in an HTML5 centric way. In other words to  
consider how 1)  how HTML might change to address the use cases; 2)   
how HTML5 might provide UA norms to address the issue; or 3) how this  
WG might liaison with other organizations and vendors to address the  
issue. Anyone is welcome to contribute to the wiki page (if that  
needs saying).

Take care,
Rob

[1]: <http://issues.apache.org/bugzilla/show_bug.cgi?id=13986>
[2]: <http://esw.w3.org/topic/HTML/ContentTypeIssues>

Received on Sunday, 26 August 2007 17:11:10 UTC