Re: Equivalency, Languages, Checkpoint 2.3

Al Gilman wrote:
> 
> >#5. Related to issue #4 is the issue of why we exclude system-generated
> >equivalency relationships from the checkpoint. The original wording (23
> >October 2000 draft) does not exclude system-generated equivalency
> >relationships. I have some preference for including all equivalency
> >relationships, not just author-specified ones, though I think that handling
> >just the author-specified ones might be okay, if there is some good reason
> >to exclude them.
> 
> AG::
> 
> I would agree that 2.3 should probably apply to "user agent recognized"
> equivalents, not just "author specified" equivalents, if there is any
> difference.

Difference:

 A paragraph of text may end up being an equivalent for an illustration,
 and the user may know that by hearing it, but the user agent can't
 recognize it as such. So that is a text equivalent supplied by the
 author, but not specified as being one.
 
> Can anyone think of an example?  What was meant by a system-generated
> equivalency?  What part of the system is generating this relationship or
> content?  Is this captured in an issue that ever crossed through the issues
> list?

Yes, we've dealt with this before: author-specified v. user
agent-recognized.
We have both in the document, for different reasons.  For example,
checkpoint 2.4 includes "Respect author-specified synchronization cues 
during rendering." and that is meant to say: respect what the author
has specified. It is not appropriate in this context to say "do whatever
you can figure out."

Similarly, 4.1 - 4.4 include 'author-specified' because it's about
overriding
presentation that the author has specified. Same for checkpoint 9.3.

At times (e.g., checkpoint 8.1), at least to my memory, we use
"author-specified"
to narrow the scope of the requirement for the developer. For instance,
for
checkpoint 8.1, we say "author-specified relationships among table
cells"
so that developers don't think they have to invent, or infer
relationships
among cells.

Checkpoint 2.1 uses the expression "recognized equivalents" as opposed
to "author-specified equivalence" so that if the UA is clever (i.e.,
cleverer than what is strictly stated by the author), it still has
to do something with information it recognizes.

For consistency with 2.1, I agree with Al that we should use
"recognized" instead of "author-specified".

 - Ian

Previous discussion on this topic:

 April 2000 face-to-face meeting, issue 210
    http://www.w3.org/WAI/UA/2000/04/rfbd-20000410
    http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#210

 (I'm sure we've already changed some checkpoints in light of this
  discussion, but I don't have other URIs for the moment.)

-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel:                         +1 831 457-2842
Cell:                        +1 917 450-8783

Received on Thursday, 2 November 2000 13:40:48 UTC