RE: Proposal for an HTTP ERR method

>-----Original Message-----
>From: ietf-http-wg-request@w3.org 
>[mailto:ietf-http-wg-request@w3.org] On Behalf Of Henry Story
>Sent: Wednesday, 23 June 2004 19:23
>To: ietf-http-wg@w3.org
>Cc: Justin Chapweske; Atom Syntax
>Subject: Re: Proposal for an HTTP ERR method
>
>
>Sorry I should have responded in more detail to your alternatives.
>
>On 23 Jun 2004, at 18:47, Justin Chapweske wrote:
>> If you're going to specify anything, submit a proposal to advertise 
>> feedback URIs using a RESTful approach, perhaps in HTTP response 
>> headers
>
>Sounds like an idea. But if the HTTP header itself is broken...?

Why shouldn't ERR not be broken in this case?
IMHO, POST is more widely implemented and more likely to work correctly.
This would also speed implementation and correctness of the system. Why
reinvent the wheel when you can already use it?

If HTTP itself is broken, why actually use HTTP to report HTTP is broken?
This seems quite risky to me...

>
>> such as Link, or within the (x)html itself as a <link> tag.
>>
>
>This one is less good, because you are asking us to read the 
>body of the response, which may be broken.
>

To prevent annoying the wrong person, whenever possible. The responsibility
of the content might be with a different person than the responsibility to
keep the server alive.

- Joris

Received on Wednesday, 23 June 2004 13:46:51 UTC