Eric Scheid wrote:
> On 31/10/05 4:59 PM, "Joe Gregorio" <[EMAIL PROTECTED]> wrote:
> 
> 
>>Section 9: """Clients SHOULD be constructed with this in mind and
>>SHOULD perform a GET on the member resource before editing."""
> 
> 
> editorial: move this requirement to the section discussing editing, not
> listing?

Noted for 07. The editors will try it out and see how it reads.


>>Adding a single header is not encouraging """a bunch
>> of implementation specific metadata being
>> passed around in the HTTP headers"""
>>
> 
> 
> but it does set a precedent, which is I think what James was suggesting.

Opinion: if we're going to use entries as general purpose metadata
containers (mnot's phrase not mine), we might as well be hung for a
sheep as a lamb and use them for carrying everything. It's clean from a
layering perspective. At the same time I don't see Title as setting a
precedent - I'm betting impl specific junk will end up in entities not
HTTP - the world's mostly not ready to treat HTTP headers as non-wire
data just yet. SOAP and XML-RPC are presented in evidence.

cheers
Bill

Reply via email to