Dear Jonathan,

I work for a small British-based (New Zealand-owned) company called
Dolphin Computer Access Ltd.

You may like to know that the next version of our Windows-based screen
reader, Hal v2.01, is compatible with both Microsoft Office 97 and
Microsoft Internet Explorer 4.0.  A beta version of Hal v2.01 should be
available in the next fortnight.

Dolphin has designed Hal to be completely independent of Active
Accessibility for a number of reasons, including the fact that Active
Accessibility is not yet properly supported under Windows NT 4.0.

Hal is a "screen reader" in the truest sense of the word, as it is
primarily designed to operate on the visible information on the screen.
It does not care how the information got there.

In Windows there are many ways to draw something such as a button, but
all buttons look the same on the screen.  Some of the traditional Windows
screen readers rely solely on the class name or class atom of a window in
order to identify what type of control it is.  Therefore if a Windows
application developer uses local superclassing in their application then
this technique for determining a control's type will fail.

Unlike these other Windows screen readers, Hal recognises objects in a
similar way that your brain recognises objects, by looking for distinct
attributes, shapes, borders, highlights etc. This makes Hal independent
of whether an application  has "obeyed the rules" of Windows programming.
So by looking at the final result that an application has generated on
the screen, Hal works independently of how something was created (or
whether the application implements Active Accessibility).

The advantage of this is that once set up for one application, all
similar looking applications will talk correctly without any adjustment
to the settings.  This is certainly true of the new pull-down menu
systems implemented in Microsoft Office 97 and WordPerfect 8.0.  Hal
v2.01 works well with both of these new pull-down menu systems because
they look identical.

How well an "Active Accessibility powered" screen reader works is
determined by three factors:

1.  How well Active Accessibility is implemented in the third-party
application.
2.  How well Active Accessibility is implemented in the screen reader.
3.  How well Active Accessibility is designed/works.

Screen reader manufacturers have no control over factors 1 and 3.  For
example, factor 1 relies totally on independent software houses (with no
experience in the field of disabilities) implementing Active
Accessibility support properly in their applications.  Do third-party
application developers even have a screen reader to test out the quality
of any Active Accessibility support they may implement in their products?
 An "Active Accessibility powered" screen reader is only as good as the
information it receives from the third-party application.

This is why Dolphin feel that it is important not to develop a screen
reader that is solely reliant on Active Accessibility.

If you would like a beta copy of:

Hal v2.01 for Windows 95
or
Hal v2.01 for Windows NT 4.0 Workstation

then please let me know.  Hal v2.01 also includes a whole host of new
features including Braille display support and document read.  It is also
planned to include object labeling by the time it is released.

I hope that this goes some way to allaying your fears that Internet
Explorer 4.0 will be inaccessible to people with disabilities.

Yours Sincerely,


Gareth Collins
Support Manager
Dolphin Oceanic