Functional Request for DOM

Hallo W3C,

as discussed in the teleconference on march 28th I will send you a list of
functional needs Ai Squared has using the DOM:

Events:
*	focus tracking: we need to get an event when the focus changes and
we need to be able to identify the element that has the focus 
*	state change: we need to get an event when the state of a radio
button or check box changes
*	list tracking: we need to get an event when a different item in a
list is selected (list box / combo box)
*	text cursor tracking: we need to get an event when the text cursor
is moved in an input element and when text is selected, edited or deleted
*	download: we need to get an event that tells us that downloading of
a page is completed

Data:
*	We need to be able to get all text and active items (links, frames,
controls, graphics) on a web page.  
*	For each element we need to get the screen coordinates 
*	We need to get screen coordinates not only for a text block but for
at least each word in the block if not each character
*	We need to be able to tell if an element is scrolled into view or
not 
*	For text we need to get the foreground and background color, font
name, size, attributes (bold, italic etc)
*	For controls and frames we need to get the type, title, state,
selected item
*	If the author hasn't provided a title for a control we need to be
able to fast navigate to the previous or next text element (assuming that
this might be the title)
*	We need an fast way to find the control that has the focus
*	For graphics we need a meaningful description

Navigation:
*	We need to be able to enumerate the links, frames and/or controls on
the page in the tab order
*	We need to be able to identify functional blocks on a web page
(banners, navigation bar, document area)
*	We need to be able to find elements by screen coordinates (at a
point or in a rectangle)
*	We need to be able to distinguish between tables that are used to
internally structure the page and tables that are actual tables for the user

Action:
*	We need to be able to scroll an element into view
*	We need to be able to programmatically set focus to a link, frame or
control
*	We need to be able to programmatically execute a link, click on a
button etc

Currently we are using the DHTML Object Model of IE, but we are currently
investigating what we could win by switching to IE DOM. One issue was not
discussed in the teleconference: performance. My assumption is that
performance could be increased by defining a set of functions that are doing
exactly what AT vendors need, for example: GetFocusElement,
GetElementAtPoint, GetElementInRectangle, GetSelectedElement, etc

A lot of our requests are related to screen coordinates / rendering
information. I do understand that this is not part of DOM, but for a screen
magnifier is an essential information. Maybe DOM could describe a way how to
get rendering information. Maybe it would be possible to have another
standard defining how the rendering engine is providing information to the
outside world. 

Best regards
Jost Eckhardt, Ai Squared

Received on Tuesday, 2 April 2002 09:14:14 UTC