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

Problems with guideline 4.1

From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
Date: Sun, 25 Jul 2004 19:37:51 +1000
Message-ID: <16643.32623.472041.233968@jdc.local>
To: Web Content Guidelines <w3c-wai-gl@w3.org>

Some issues with guideline 4.1:

1. "the site has documented". Where? How? This will probably need to
   be revisited if we adopt the idea of an "authored unit" (or
   similar) in defining the scope of conformance claims.

Proposal: it should be documented in, or linked to, the conformance claim.

2. Suppose the content is written, in part or in whole, in a format
   for which no specification exists. Does this mean it fails
   guideline 4.1? If the specification exists but hasn't been
   published, does this count? If so, how can I ever know whether an
   unpublished specification for a technology I happen to be using,
   exists?

pROPOSAL: "Specification" means a published specification which is
available to any member of the public for free or at a reasonable cost.

3. Suppose I write my content in an XML format that I have developed,
   but for which I have not written a formal DTD/schema, and details
   of which I haven't published. The content is transformed into a
   number of output formats for various devices and user agents
   according to the delivery context. Have I violated guideline 4.1?

This raises the larger issue of where along the delivery path the
guideline are to be applied. Note that the "authored unit" proposal
doesn't resolve this issue. For example, content written in my
custom-designed XML format is available in that format via a URI -
even if it is a FILE URI on my local filesystem, or perhaps a URI used
by my own software or a third-party gateway to retrieve and transform
the content for delivery to a user agent.

4. "Structural elements and attributes". This is very specific to
   markup languages and looks like an attempt surruptitiously to
   support the HTML techniques without being general enough. What we
   mean is that every feature of the technology is used as specified.
   Even if we do decide to say "structural elements and attributes",
   what of technologies that don't provide elements or attributes? Do
   we say that this requirement is satisfied by default?

The broader issue is that of the assumptions which each success
criteria requires to hold in order for it to be applicable, and how
these should be documented in the guidelines.

5. The level 3 success criteria. These should be at level 2. It seems
   troublesome that a Consortium such as W3C would be asserting, in
   effect, that conformance to specifications without exceptions is a
   level 3 requirement - a really good thing to do, rather than at
   least a level 2 requirement.
Received on Sunday, 25 July 2004 05:37:55 UTC

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