separating navigability of elements from sequential/etc. motion.

Jon asked me to put something more specific on the list.

Here is a little bit.

For example, in the current draft recommendations we see things like:
    
   [285][Technique: 5.4.1] [Priority 1]
          Allow the user to navigate sequentially among links.
          
   [286][Technique: 5.4.2] [Priority 1]
          Allow the user to navigate sequentially among elements with
          associated long descriptions.
          
   [287][Technique: 5.4.3] [Priority 1]
          Allow the user to navigate sequentially among form controls.
          
[etc.]

First, the strict way I would read those requirements are, for example,

	From a form control, the user should be able to move
	directly to the next/previous form control.

While what I think we agreed is the need is

	The tabbing sequence in sequential navigation should
	include (stop at) form controls.

The tabbing sequence or basic sequential navigation motion is
going to be moving in textual occurrence order or TABINDEX order
if stated and different among a set of items which is the union
of links, form controls, and a few other things.

That was my first comment.  The second is that sequential 
navigation is not even really required.  It is just a minimum.
A good hierarchical navigation would remove the requirement
for sequential navigation, IMHO.  Not, as Scott would point
out, a bad hierarchical navigation capability.

I think that we should look at different classes of things that
appear in the document and say which of these kinds of things
the user needs to be able to

	- read in isolation
	- discover X contextual information about
	- move to in navigation

Some of this may wind up being expressed in terms of existing
notions of selection and focus, or a new notion of point of
regard, but I think that the designation of element classes as
feasible navigation destinations should be separated an arm's
length from the design of navigation strategies and their tuning,
whether they be hierarchical, sequential, direct or whatever.

Al

Received on Thursday, 10 December 1998 18:32:20 UTC