Re: LC28: HTTP Transfer Coding and 1.0

Hi Charlton.

* Charlton Barreto <cbarreto@webmethods.com> [2005-03-30 15:17-0800]
> In issue LC28 [1], it was raised that as the transfer coding feature 
> does not apply to HTTP 1.0, it is unclear how a processor would handle 
> a  {http version}="1.0" and {http transfer coding} claim. In general 
> such a claim should be safely ignored by the processor.
> 
> To resolve this issue I propose we update section 3.10.1 to have the 
> following language:
> 
> "Every Binding Message Reference component MAY indicate which transfer 
> codings, as defined in section 3.6 of [IETF RFC 2616], are available 
> for this particular message.
> 
> The HTTP binding provides a mechanism for indicating a default value at 
> the Binding component and Binding Operation levels.
> 
> If no value is specified, no claim is being made.
> 
> Any transfer coding specified for HTTP 1.0 Binding is ignored."
> 
> where the change is in the addition of the last sentence.
> 
> [1] http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC28

We have two ways to go about this: either ignore the value or force it
to be empty. I have a slight preference for the latter, but can live
with the former.

However, in any case, we don't define the concept of HTTP 1.0 Binding.
I would therefore like to propose a friendly amendment, using property
values:

  The value of the {http transfer coding} property is ignored when the
  value of the {http version} property is "1.0".

The solution with forcing it to be empty would look like:

  When the value of the {http version} property is "1.0", the {http
  transfer coding} property MUST be empty.

Cheers,

Hugo

-- 
Hugo Haas - W3C
mailto:hugo@w3.org - http://www.w3.org/People/Hugo/