Re: XAG point 4.10 and 3.2 (more)

[Earlier, I wrote:]

In 3.2, all it says is that the markup should create the structures that the 
navigation functions (explained under the 4.10 "export semantics" documentation)
operate on.

[continuing the thought...]

Although this distinction is vague as we have developed it so far, it is important.

Phill Jenkins has raised this as a concern, that the division of labor between
the format and author on one hand and the browser and AT on the other be reasonably 

Some overlap may be unavoidable but we do want to minimize waste motion.

While supporting skip-nav links as current practice, we should be simultaneously
moving to replace them with escape-block methods in the user agent, as extended by
the Assistive Technology if necessary.

Advanced example:

So-called 'search navigation' includes a match function, an order function, and a go-to function.

It is important to realize that lists are a subclass of bags, so that the match
function is applicable to any bag and is re-used within this "search navigation" method
without requiring a new and different function.

The "search navigation" method may eventually offer a "best match" for fuzzy queries
in addition to the current "next match" or "previous match" options.  This is achieved
by changing only the ordering function.  The match function now returns a richer type than 
Boolean, but it still applies to any bag of matchables.

A navigable structure is a class of structure in that it has to be fit for use in 
navigation operations; but we may meet this requirement by describing a virtual-graph
structure as if a data structure and the proper binding of the arcs in the graph to 
possible state changes in the transactions of the proper navigation method(s).

Unless the operational use of the structure is specified in the dialect definition, people
won't populate the data model of the format in a way which does work well when the desired 
method is applied.  We learned that in the tooltip wars.  We don't want to experience that 
all over again. That is why there are separate 3.2 and 4.10 clauses.  


At 11:39 AM 2002-06-25, wrote:

>I have (yet another) a question concerning point 4.10 and 3.2 in the XAG (
>They both look similar to one another.
>Point 3.2 : Define navigable structures that allow discrete, sequential, structured and search navigation functionalities.
>Point 4.10: Document navigable structures. Describe how discrete, sequential, structured and search navigation mechanisms should work.
>Here's how I understand it:
>3.2 implies that the author defines those navigable structures (perhaps by explaining in the documentation the possibilities certain elements have in making navigation possible) and 4.10 makes sure that whatever is defined is also described. Both sounds very much the same to me.
>Have I (yet again) misunderstand? Can anyone explain it to me?

Misunderstood, but only a little.

In 3.2 it does not assume that the navigation functions have been documented.
That is why we come back and say that in 4.10.

In 3.2, all it says is that the markup should create the structures that the 
navigation functions (explained under the 4.10 "export semantics" documentation)
operate on.

A named anchor is an example of satisfying 3.2, hyperlink navigation using a 
#fragment URI-reference is a navigation method (direct navigation) that operates
on this structural feature within the document instance.

A MAP wrapping a navigation bar is an example of satisfying 3.2; an 'escape' 
operation that is defined for any MAP element is an example of a navigation 
function that operates on this structure.

See the 'escape' provisions in the standard talking book


>Thank you. 
>Astrid Callista

Received on Tuesday, 25 June 2002 10:44:31 UTC