httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joseph J. Gunn" <>
Subject Re: DefaultType and bug 13986
Date Fri, 07 Feb 2003 16:34:00 GMT
On Friday 07 February 2003 09:32 am, you wrote:
> Can anyone (with a bit of spare time) give his or her ideas on
> ?
my 0.5 cents.

My interpretation of the RFC is that there should be a mechanism for the 
client to interpret the stream. This argues for the "No Content-Type" option.

Reading between the lines the SHOULD on the server side, and MAY on the 
client side, allows content to be interpreted on the client side. This is 
just not possible, without violating the RFC, if the server always forces the 

This doesn't take control away from sysadmins, the ones who can have tight 
control over the content can still specify completely the content types that 
their servers use. The ones that have less control over the type of data that 
is served can choose to use no Conteht-Type, or a specific one through the 

So the default configuration of Apache should include, as it does, a way to 
identify well known types. Then hands off on the guess, unless specifically 
overridden by a human.

from the RFC:
   HTTP uses Internet Media Types [17] in the Content-Type (section
   14.17) and Accept (section 14.1) header fields in order to provide
   open and extensible data typing and type negotiation.
note the "open and extensible".... most completely achived by not

View raw message