UI and DOM: sequel to discussions about AT

User Interface and Document Object Model are two totally
disjunct subjects. Roughly said thay are about "How to
access" (UI) and "what information" (DOM). The UI can be
exposed as an object-model, but it would be a UI-OM.

The UI is historically platform dependent. Java tried to unify
all UI's and coped with least common denomitor problems
and in the end came up with their own UI.

Documents are not platform dependent (cerntainly not on the
www, which is the main concern of the w3c).

Assistive technology has to tackle two problems: How to
make the UI accessible and how to make the documents
accessible. 
If one part can be platform independent this would be great.
DOM is the answer there. A technique like MSAA and java
accessibility solve both problems at the same time, which is
great indeed, but makes the total platform dependent (if you
can say that java is platform-dependent at all). I do not see
this change although there might be only limited work
involved to seperate them. 

The ideal D-OM would then complement a UI-OM. The DOM
level 1 indeed lacks necessary richness (like reentrance, full
featuredness, events). Nevertheless DOM is something the
w3c can and should promote in my opinion. 

I keep promoting this view for the following reasons:
1) I think it holds the biggest promise for the future. Although
practical solutions are not nearby, we do want the best in the
future (don't we?)

2) Like others, I do think that the UI is out of the scope of the
guidelines. UI-accessibility is an important issue but touches
very much proprietary implementations. The UAGL must say
what a UA should be able to do, but not by dictating a UI. The
UAGL must say what means of input should be used (e.g.
keyboard), but not what specific keystrokes.

3) DOM is too important not to keep it in the UA is some
form.

Hans Riesebos (HRiesebos@alva-bv.nl)
ALVA BV, The Netherlands

Received on Tuesday, 9 February 1999 11:50:11 UTC