- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Tue, 25 Mar 2008 10:33:05 +0200
- To: "Ben Boyle" <benjamins.boyle@gmail.com>
- Cc: public-pfwg-comments@w3.org, "HTML WG" <public-html@w3.org>
On Mar 25, 2008, at 05:34, Ben Boyle wrote:
> I wonder if where "ajax libraries" are mentioned, that may refer more
> generally to authoring tools? There are a number of tools -- content
> management systems and application frameworks spring to mind -- that I
> would also place in the same category: using classes to reinvent
> semantics that exist readily in HTML.
For the most part, ARIA doesn't make sense as something static, so
CMSs and application frameworks would have to generate JavaScript or
include a canned JS ("Ajax") library. That is, I don't expect
authoring tools to edit ARIA stuff directly but to insert canned
script-based widgets hand-developed in a text editor.
The part of ARIA that works as static, and thus could be relevant to
document editor-type authoring tools, are landmarks. I don't like
introducing landmarks at this stage. They are mostly redundant with
new HTML5 container elements like <header>, <footer>, <nav>, <aside>
and <article>. Implementing the new elements in a Web browser so that
they 1) parse as containers, 2) have display:block; in the UA style
sheet and 3) are reported as landmarks to AT should be about as low-
hanging fruit as reporting landmark role attributes to AT.
Just about the only good reason for having <div role=navigation>
instead of <nav> is that <nav> doesn't parse nicely in pre-Firefox 3
Gecko based browsers. But if significant styling isn't tied to that
element, the resulting behavior shouldn't be a disaster in Firefox 2.
Is this really reason enough to introduce an attribute-based way of
expressing document landmarks?
--
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Tuesday, 25 March 2008 08:33:50 UTC