W3C home > Mailing lists > Public > public-html@w3.org > November 2008

Re: a/@ping discussion (ISSUE-1 and ISSUE-2), was: An HTML language specification vs. a browser specification

From: Julian Reschke <julian.reschke@gmx.de>
Date: Tue, 25 Nov 2008 09:32:19 +0100
Message-ID: <492BB813.8070203@gmx.de>
To: Ian Hickson <ian@hixie.ch>
CC: Jim Jewett <jimjjewett@gmail.com>, public-html@w3.org

Ian Hickson wrote:
> ...
>>>  The PING method is used to report to the origin server that the 
>>>  client has navigated from one resource to another on behalf of a 
>>>  user.
>> I wouldn't use the word "origin", since the host being pinged may not 
>> have sent anything to the client before.
> 
> I'll let Julian fix the text to be whatever would be appropriate for the 
> spec.
> 
> 
>> Other than that, it meets all objections of which I am aware.
> 
> Ok Julian, the text you asked for it above.

"The PING method is used to report a page navigation. The Request-URI 
represents a link auditing listener (terminology?). Source and target of 
the page navigation are reported as (... - insert what's used, headers, 
or request body)."

...statement about general transmission requirements -- both request and 
response can have a message body, to be delimited as per RFC2616. If 
PING doesn't use them it should state that they are currently unused and 
either should be ignore (must ignore) or nor (must understand)...

...statement about cacheability -- state that for tracking to be 
accurate, the server must not indicate that the response is cacheable; 
no caching headers at all should be sufficient...

...state that the method is safe in the sense of RCF2616 (from the 
user's point of view).

Finally, I'd probably choose a less generic term, such as "AUDITNAV" or 
something like that; must people understand "ping" as something different.

BR, Julian
Received on Tuesday, 25 November 2008 08:33:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:27 GMT