Re: Checkpoint 7.2

thatch@us.ibm.com wrote:
> 
> Ian, I am terrifically dense. 

Nice beginning!

Please bear with me in this email, it may seem a little long-winded.

> 7.2 seems to say "provide access
> to content of an element that is selected."
> 
> 7.2 For dependent user agents. Ensure that the user has
>     access to the content of an element selected by the user.
> 
> You say that the intention is to allow the user to select and you go into detail
> about selection process.  The words of 7.2 certainly do not coincide with the
> intention you described.

You're correct, so clarification will be necessary.
 
> By my reading, 7.2 says if a user can select an element (paragraph, heading,
> whatever) then she should have access to the content of that element. If no
> element selection is possible 7.2 doesn't apply.

But why would no element selection be possible?

I agree with Charles, who said on this thread:

> We want something like ensuring that the user can make all the
> information provided for an element which is intended to be human-readable be
> rendered. (for example there is not much necessity to render the width of a
> frame, but it should be possible to get access to a title.

And then:

> I hope that someone can come up with some simpler wording which conveys the
> idea.

I hope so too.

Just a word of background. One reason this checkpoint is here is
to address access to table cell content on a cell-by-cell basis.
Users should be able to select an element according to the 
document tree and ask a dependent user agent to render it.
I believe that the first goal of this checkpoint
was fine-grain control over structured access to content
(to avoid, for example, problems exhibited by
some older screen readers for multi-columns or side-by-side
table cells).

However, structured access is just one technique.
Unstructured selection may also be desirable (e.g., text not
strictly within a single element). The dependent user agent should
still be able to render that selected content. 

More questions: Should the user only be allowed to select rendered 
content, wherever that content coments from? Or should
the user also be able to select an element and ask for
attribute values that haven't been rendered but are human readable?
This could be very useful (to all users, actually), 
but I think it would not be obvious when combined
with unstructured selection (which element do you mean if
your selection spans two partially?).

In my opinion discussion of structured and unstructured 
selection, presentation of attribute values, etc. belongs
in the techniques document. The underlying requirement is 
that the user be able to select some part of the document (in a 
device-independent manner!) and ask a dependent user 
agent to render it.

So, if "Ensure that the user has access to the 
content of an element selected by the user." is unclear,
let's simplify to:

  "Allow the user to select a portion of the document
   for rendering (e.g., a particular element, part of
   an element, human-readable attribute value, etc.)"

Then, in a more elaborate note after the checkpoint:

  "For instance, allow the user to identify 
   a table cell with the selection and render the 
   cell's content and (optionally) associated header 
   information. Or, allow the user to select part of
   a paragraph for rendering. Or allow the user to 
   select an element and request that the value 
   of an attribute known to promote accessibility 
   be rendered."

  - Ian



-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel/Fax:                     +1 212 684-1814

Received on Thursday, 26 August 1999 22:44:52 UTC