[XHTML-role] How to define roles still needs clarification

Hi

This is a comment for "XHTML Role Attribute Module"
http://www.w3.org/TR/2006/WD-xhtml-role-20061113
2006-11-13
Working Draft

Re:

> Note that current best practice is that the URI associated with that
> namespace resolve to a resource that allows for the discovery of the
> definition of the roles in the namespace.

As I suggested back in September,

http://lists.w3.org/Archives/Public/www-html/2006Sep/0030.html


Surely it should be an *absolute and fundamental requirement* (not
merely "best practice") for "the URI associated with" a role's namespace
to "resolve to a resource that allows for the discovery of the
definition of the roles in the namespace" so that user agents can always
learn new roles?

Re:

> User agents, search engines, etc. may interpret these relationships in
> a variety of ways

If browsers are to learn new roles, should it not be an absolute
requirement for role definitions to include machine-understandable
suggestions on how to interpret and render such relationships, aurally
and visually?

Should it not also be an absolute requirement for role definitions to
include a machine-understandable specification of whether the defined
relationships are:

1. Of primary importance and must be obviously exposed to end-users
(like ordinary hyperlinks).

2. Only of secondary importance, with access dependent on end-users
requesting more information (like the TITLE attribute in HTML4).

3. Unimportant to end-users (like the CLASS attribute in HTML4).

Should such styling and behaviours be entirely dependent on (potentially
disabled) stylesheets and scripting, and does that conflict with
accessibility requirements? This must be clarified.

Bear in mind, when considering this question, the example of the radical
difference in treatment by current browsers between the HREF attribute
of LINK, the HREF attribute of A, and the under-appreciated CITE
attribute of INS, DEL, Q, and BLOCKQUOTE in HTML 4.01. No rendering was
suggested and no importance was specified for CITE, and it has been
mostly ignored, undermining its potential to extend hypertext in
interesting ways.

(I'm cross-posting to the www-html list for discussion.)

--
Benjamin Hawkes-Lewis

Received on Saturday, 18 November 2006 00:08:21 UTC