Mozilla - like JAWS or like Hal?

["Saqib Shaikh" <me saqibshaikh com>]

Hi

I ve been following the thread regarding Mozilla 1.7 RC1 acccessibility.
I'd like to make some comments based upon my experience with two Windows
screen readers - JAWS and Hal.

JAWS effectively textualises the screen: it inserts the word "link" or
"visited link" into the text of its virtual buffer, and also inserts words
like "list of x items" or "table with x columsn and y rows".  If you copy
and paste from JAWS into a text editor you'll get this textual
representation.

In contrast Hal takes the approach of leaving the screen just the way it is,
and reading what is actually there.  It has a virtual focus mode, but this
is more like a reinterpretation of the graphical screen, not a textual
replacement.

Likewise in Mozilla's text browsing mode I'd like links to be coloured and
underlined, but no word "link".  Likewise Headings should be bold or
whatever, and tables/lists/frames should look like what they are.  So what
we have is a version of the main page, but with the ability to cursor up and
down, select text with the keyboard, do text finds within the document, and
also maybe have a list of links/headings/frames appear at the press of a
keystroke.  This is all quite general functionality that is acceptable IMHO
in a text browsing mode, but which doesn't make it a screen reader only
browsing mode.

Then Gnopernicus should be given enough semantic knowledge of the document
that when it comes across a link it should know whether it is a visited link
or not, and when inside a table, even though Mozilla's table navigation
commands will be used, Gnopernicus should represent the table in
speech/braillle in the appropriate fashion.

I think this is the best way to present this UI, but would appreciate any
comments.

Saqib