W3C home > Mailing lists > Public > public-webapi@w3.org > September 2007

Re: Comments on XMLHttpRequest -- From CDF WG Review

From: Anne van Kesteren <annevk@opera.com>
Date: Tue, 25 Sep 2007 14:08:32 +0200
To: "Steve K Speicher" <sspeiche@us.ibm.com>, public-webapi@w3.org
Cc: " CDF WG" <member-cdf@w3.org>
Message-ID: <op.ty7lcic564w2qv@annevk-t60.oslo.opera.com>

On Tue, 25 Sep 2007 11:29:47 +0200, Steve K Speicher <sspeiche@us.ibm.com>  
> Is this stating that a server responding to an XMLHttpRequest load or
> such, must know if the user agent supports XML (including support for
> namespaces) and respond with null or is it up to the conforming user  
> agent to decide if it supports XML and whether to always sets the  
> response to
> null?

The user agent, as Maciej indicated. I made this a bit more clear in the  
draft but as noted in another e-mail that has not yet been comitted due to  
problems with the spec generator.

> 2) #dependencies for "DOM", "Window", "HTTP":
> I find the phrasing of: "must support some version of DOM Events and DOM
> Core " to be very confusing.  Then there is a reference (normative?) to
> DOM3 Core and Events.  So the "some version" really DOM Level 3 or is DOM
> Level 0 ok, as it is *some verson* of DOM?
> This wording exists for: DOM, Window, HTTP which seems a bit vague.  I'd
> suggest that it clearly define a minimum level to support: say DOM Level  
> 3 Core, etc.

It's intended to be vague, as full support is not needed at all in order  
to claim conformance to the XMLHttpRequest specification. I'm not sure how  
to say that in an accurate way though, suggestions welcome.

> 3) #terminology
> In what context is this matching scheme used or needed?

open(), setRequestHeader() etc.

> This section seems to discuss string casing and equivalency of characters
> in ASCII only charsets, it seems that is should consider matching defined
> by some internationalized context [2]

The strings it deals with are all related to HTTP and HTTP does  
case-insensitive matching per ASCII conventions.

> 4) #xmlhttprequest
> This section uses terms such as "successfully invoked" and "user agent
> successfully acknowledged the request", though was unable to find
> definitions of how success was defined.

That's because these are used as descriptions there, not as normative  
definitions. The actual state transitions are detailed in the algorithms.

Anne van Kesteren
Received on Tuesday, 25 September 2007 12:08:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:16:24 UTC