[context help] difficulty of distributed prompt construction

The orientation information for filling out a form is the general topic
for which "context help" is a subset.

Labeling is the base level of orientation.  That is to say, let's use
'labeling' to describe a level of orientation that is always or usually
presented to the user in the context of controls where they have the
opportunity to perform an input action.

The difficulty of parsing this out into 'content fragments' rationalized
in the markup structure of the document and 'dialog flow' constructued
by the user agent from those fragments is illustrated by a current thread
on a very simple form, the ubiquitous search tool:


But this is a good thread.  The two points that John Slatin raises are
key facts of web usability for the screen reader user.


Received on Wednesday, 10 December 2003 11:05:24 UTC