W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2004

RE: Meaning and scope of "authored unit"

From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
Date: Mon, 26 Jul 2004 10:27:26 +1000
Message-ID: <16644.20462.917993.497559@jdc.local>
To: Web Content Guidelines <w3c-wai-gl@w3.org>

Gregg Vanderheiden writes:
 > 
 > Thanks Jason
 > 
 > This is a very good leap forward on these topics.  I think we should use
 > your # 1 definition to work from rather than the two parter. Also I am not
 > sure about the "may be retrieved' so I suggest  "[may be / are]" til we
 > figure it out.

That's fine. My main reason for saying "may be retrieved" is that not
every user agent or other tool processing the content will retrieve
all of the subsidiary resources. To take a simple example, suppose I
supply a style sheet; some user agents won't retrieve the style sheet,
but we don't want this to result in the style sheet's not being deemed
part
of the authored unit for which conformance is asserted. Similar
examples apply to metadata that may be retrieved by some tools but not
others, and which is part of the authored unit.

 > 
 > RE content negotiation -- since not everything DOES content negotiation -
 > and it may not be clear which version to get with multiple disabilities - I
 > believe that there must be default version.  
 > 

I disagree here. Suppose I set up my server so that failure to
negotiate results in an error response. That is, if the client doesn't
specify its delivery context it receives an error response and no
content is retrieved. This is a perfectly valid way of ensuring
negotiation in which there is no default version. It might not exist
today, but may well arise later, and my proposed wording was intended
to take this into account.

 >  So that would give us:
 > 
 > 
 > <propose>
 > 
 > An "authored unit" includes a resource identified by a URI, together
 > with any other resources directly or indirectly associated with it,
 > that [may be/ are] retrieved in the course of rendering or otherwise
 > processing it.
 > </propose>
 >  

Acceptable but I still prefer "may be retrieved" for the reasons outlined.
 > 
 > <propose>
 > Where a resource identified by a URI can be retrieved in multiple forms
 > depending on content negotiation or other aspects of the delivery context,
 > the resource conforms to WCAG at some level if the version that will be
 > retrieved in the absence of content negotiation conforms to WCAG at that
 > level. 
 > 
 > NOTE: Additional form may be provided that are more or less accessible to
 > different groups but a conformance claim must be made based on a single form
 > not on a union of different forms. 
 > 
 > </propose>
 > 
I agree with the note but disagree (see above) about the default version.
Received on Sunday, 25 July 2004 20:27:29 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:17:58 UTC