Re: headers attribute (was Re: Form elements)

On 2007-06-02 23:20:19 +0200 Maciej Stachowiak <mjs@apple.com> wrote:

> On Jun 2, 2007, at 2:01 PM, Leif Halvard Silli wrote:

>> On 2007-06-02 22:20:55 +0200 Laura Carlson  <laura.lee.carlson@gmail.com> 
>> wrote:
>> 
>>> On 5/31/07, Schalk Neethling <schalk@alliedbridge.com> wrote:
>> 
>>> Related quantity of cases blog comment by Ian Hickson:
>> 
>>> If something is
>>> rare, then it won't be supported. It really is that simple.
>> 
>>> Related blog comment by Gez Lemon:
>>> "...By definition, people with disabilities are a minority...
>> 
>> Exactly.
> 
> In context, I think Ian was referring to rare authoring needs, not  rare user 
> needs. In particular, the example cited where scope would  not adequately 
> describe the table but headers would, is fairly  obscure and not 
> representative of the kinds of tables found on the  web.

As authors, we are limited of the options we have. I have never used SCOPE or HEADERS. But if they had been available as CSS selectors, then I am pretty sure I would have used both long ago.

I don't like this approach «let's see what we can remove». Why remove it? If you have managed to create rules for for how SCOPE is used and calculated, then HEADERS is a peace of cake: Each cell will have a implied HEADERS attribute, and its implied value would be the value which the SCOPE operation told it. If most AT UAs support HEADERS, then it make most sense to «automate» how HEADERS is working.

When people manually inset HEADERS, they can overrule the SCOPEd value. (Or insert values for those that do not «get» the scoped value.)

> An author might have 
> the choice to lay out their table  differently but a user would not, once the 
> table is created.

What do you as an browser maker think about the idea of making SCOPE and HEADERS usable as CSS selectors? This would increase author incentive for using them. And that again would increase AT UAs incentive of supporting them.

> I'm more interested in the impact on existing screen readers of scope  vs. 
> headers, since that will be more relevant to common tables.

I support you on this. But I am a bit afraid if the purpose is to see which attributes we can take out.
-- 
leif halvard silli

Received on Saturday, 2 June 2007 22:13:49 UTC