RE: action item for the testability of 3.5 (annotating complex information)

Awhile ago, there was some discussion about the testability of 3.5.
Thanks for the comments and sorry for the delayed response!

Here's a list of the proposed success criterion and a list of related
issues. I'd like to discuss this checkpoint on an upcoming conference
call.

Success criteria 1:
My proposed wording is:
"A definition or link to a definition is provided for phrases, words,
acronyms and abbreviations for the first occurrence on each page."

Issue: 
1. Should the definition appear with the first occurrence on each page
or site? 
Proposed resolution - On each page, because a user will not navigate
through a site in any particular order.

Success criteria 2:
My proposed wording is:
"Provide a summary for tables with nested heading and/or cells that span
multiple columns or rows. The summary should explain the relationship
between the cells."

Issues:
1. We need to define what constitutes a summary. 
a. At a minimum, the definition could be that some reference to the
table be made within the text of the document. However, how can we make
the definition of a summary more meaningful? 
b. Can the success criteria state that the summary should be explicitly
tied to the table?
If yes, then one suggestion is for the HTML technique to use <CAPTION>
or <LABEL FOR >. 
Will these meet our needs? Are there other techniques for other
technologies?

2. Does the success criteria cover everything? What about number of
cells?  Is it feasible to define an exact number of cells that warrants
a summary?

3. Is a summary always necessary? For instance, if the whole page is
about one table - is a summary really needed, or does the entire page
count as the summary?

4. Should the checkpoint apply to both data tables and layout tables? 
Research is needed as to whether or not layout tables should have a
summary. (for instance, magnifier users). I'll look into this. 

5. In the current draft, there is third bullet to success criteria #2
about a relationship being "obvious".
Can we make this testable? Possibly the HTML algorithm, per Wendy:
http://www.w3.org/TR/html4/struct/tables.html#h-11.4.3

6. What other applications should be included in the success criteria?
Proposed resolution: I had originally proposed that visual
representations of data and diagrams have a summary. I retract this now,
and agree with the comments that this is covered by checkpoint 1.1. 

I don't have any other suggestions, does anyone else have some ideas? If
not, we should remove this as an issue from the current draft.

Thanks.

Received on Wednesday, 13 March 2002 22:33:25 UTC