Re: 'role' should be property

2007/5/31, Mark Birbeck:
>
> Hi Thomas,
>
> > If you have no control over the document, you cannot add a link to the
> > "rolesheet".
>
> Of course you could, in just the same way that you can choose your own
> personal stylesheet. The problem today is that there is nothing in
> between writing a full-blown GreaseMonkey style of script, and
> changing stylesheets to get some CSS. There's lots of space in
> between.

Yes, such as a UA-specific feature to add a processing tool (use CSS
syntax if you like, or extension properties to CSS, or something else)
in between the "original page" and the one that is rendered (visually
or aurally or else).

> > Just use a transformation tool (XSLT* or another tool of
> > your own using CSS-like selectors) to add the role= attributes where
> > you think they should be and let your accessibility interface or
> > metadata storage system consume the transformed version instead of the
> > original one.
>
> Thanks. I'm well aware that you can hack anything if you put your mind
> to it. And you would prefer to use such a hack rather than Dmitri's
> proposal, because the letters 'C', 'S' and 'S' seem to imply to you
> 'presentation' rather than something else?

No, that's not the reason.

The reason is that if a web page uses <font size=7> to markup titles
(instead of <h1>) then you cannot infer that it's a title. Just like
<i> and <b> don't necessarily mean <em> and <strong> respectively.
If you want to standardise some additional properties to CSS or a new
language with a syntax similar to CSS, go for it, it's outside the
scope of this group.

But I believe that role= is worth being in the markup for HTML5 (even
if the future "rolesheet" language then defines a "*[role] { role:
attr(role); }" default rule, similar to the "b { font-weight: bold; }"
default CSS rule in browsers).

In brief, I'm not arguing that we should have "rolesheets", I'm
arguing that we should have role= within the markup (if proven that we
need it), eventually in addition to "rolesheets" (and rolesheets are
far more "virtual" for now than a role= attribute in HTML5)

-- 
Thomas Broyer

Received on Thursday, 31 May 2007 16:56:55 UTC