Re: skepticism about generic access solutions (was: issues with XWindow)

[Jason White <jasonw ariel ucs unimelb edu au>]

A few points need to be emphasized here:

1. It is a misconception to frame the discussion in terms of text vs. gui
interfaces; rather, the point is to develop useful abstractions which make
the application logic and data representation available, via a software
interface, to whatever input/output modules the user has chosen.
Structured data representations are starting to become increasingly common
in a variety of areas, as can be observed by reviewing the various markup
languages which have been developed over the past several years. On the
"application logic" side, T. V. Raman's work offers a number of
suggestions which capture, but in so doing, generalise, some of the basic
user interface functions: selecting a single option from a set; selecting
multiple options from a set, answering a yes/no question, and so forth.
Through the evolution of user interface design techniques it may well
become possible to encode these abstractions more directly, whilst still
offering control over their presentation in a particular medium (E.G. the
visual medium).

2. Software developers will, no doubt, always wish to implement new and
novel user interface components; the solution here is to make it as easy
as possible to incorporate therein the additional code required for
access.

3. I reject completely the assertion that such an approach represents a
centralised or authoritarian approach. To the contrary, it merely takes
advantage of the existing tendency to use and reuse prefabricated user
interface components in different applications, while at the same time
creating opportunities for experimentation, innovation and creativity on
the part of user interface designers.

4. Referring briefly to one of the specific questions that has been
raised, the Emacspeak user interface is an excellent example of an
auditory interface which relies heavily on its non-textual aspects. The
auditory icons are the most obvious example, but also, perhaps more
instructively, the different speech characteristics employed by Emacspeak
to convey structure and semantics, are not part of the text itself;
rather, they are parameters of the speech synthesizer which have intrinsic
significance only in the audio medium.