Re: RFC 2616, partitioned

On 2007/10/22, at 7:26 AM, Roy T. Fielding wrote:

>
> On Oct 19, 2007, at 12:09 PM, Robert Sayre wrote:
>> The only downside I see to this partitioning is that the header  
>> definitions are split up. I found that listing in 2616 pretty  
>> convenient, but I suppose documents like the header registry are a  
>> better reference these days.
>
> Yep, that is part of why I wanted to separate the message parsing
> requirements from the semantics -- there is no point in trying to
> present the lists as if they are complete.  This way, the requirements
> on parsing can be stated once in general and only specifically
> mention the known exceptions (e.g., cookie headers have to remain
> as separate fields for historical reasons, ...).
>
> Part of the readability fixes needed for this revision is to
> integrate the IANA considerations and pointer to the registries
> that were defined post-2616.

It's also a good opportunity to update the registries to point to the  
document itself, rather than the registration document (which is less  
than useful, especially if people are going to start relying on  
that). Ah, good; that's already <http://www.w3.org/Protocols/HTTP/1.1/ 
rfc2616bis/issues/#i40>.


--
Mark Nottingham       mnot@yahoo-inc.com

Received on Monday, 22 October 2007 02:33:31 UTC