Re: several messages

On Fri, 28 Aug 2009, Bruce Lawson wrote:
> On Fri, 28 Aug 2009 09:24:07 +0100, Jonas Sicking <> wrote:
> Perhaps we've all be brainwashed by those evil Web Standards People, but 
> it doesn't seem right that you mark up the peripheral stuff with their 
> own elements, but the main content -the purpose of the page- merely gets 
> a measley meaningless generic <div>.

The main content doesn't get a measely meaningless generic <div>, that's 
just a styling hook.

No, the main content gets a remarkably important parser-supported <body>.

It's the peripheral stuff that is unimportant enough that it has to be 
shunted off into other elements.

On Fri, 28 Aug 2009, Henri Sivonen wrote:
> Surely the reasonable thing to do here is to adjust either HTML 5 or 
> ARIA or both so that the following mapping holds:
> role=main == <main> [...]

I think role=main is as pointless as <main> for most documents. However, 
in certain cases, where the HTML markup is basically a mess of <div>s 
because the HTML5 elements aren't being used, I could see it being 
necessary. It's like role=button; you wouldn't ever need it in normal 
company, and you certainly wouldn't use it to mark up individual parts of 
an <input type=number> since you can't get to the spin buttons in that 
code, but nonetheless, it is there in case you ever want to make your own 
<div>itis spinbutton.

The point of role="" isn't to make well-written HTML accessible. That's 
already implicitly possible. The goal of role="" is to make it possible to 
make poorly-written HTML accessible to users of ATs. ARIA roles need to 
map to platform accessibility APIs. They don't need to map to markup 
languages that already exist. Doing the latter misses the point of ARIA. 
It would be like making CSS only capable of achieving styles that are 
possible with existing markup, instead of making it possible to achieve 
any appearance.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Friday, 28 August 2009 18:23:13 UTC