New Now Know How

Here’s my thesis abstract. Comments appreciated even (especially) if you have no idea what I’m on about.

Speech recognition software is often thought of as a easy to use interface that allows a person to talk to a computer instead of controlling it using their hands. Using speech recognition software is often presented in terms of “talk instead of typing”. Presenting speech recognition software as a simple matter of talking instead of typing ignores the difficulties involved in actually using large vocabulary speech recognition software in a productive work environment. This thesis began with the question: what would a speech recognition application for use live in the courtroom of the ACT Magistrates Court look like? To answer this question, a second question had to be asked and a thesis that I originally though would be a software design and implementation project became an ethnographically inspired one: What does productive use of speech recognition look currently like? Investigating the use of speech recognition software and discovering what made it usable became the focus of this project. This investigation led me to another shift in my thinking. When I started this project I thought that the usability of any technology was contained within the technology itself - that usability was inherent. Having investigated a particularly difficult-to-use technology I now believe that the usability or usefulness of any software product resides in the complex interaction of many factors both internal and external to a particular software product.

The basis of this research is two ethnographically inspired studies. In the first study, I observed and interviewed speech recognition users about their use of commercial off-the-shelf speech recognition software in various public service departments. Of the eight interviewees, six had some form of occupational over use injury and one was no-longer working. The major conclusion of that study is that the productive use of speech recognition software is dependant on many factors external to the software mainly contained within the work that is being performed and the environment in which the work is taking place. The second study concerned observing and describing part of the work process of the ACT Magistrates Court involved in sentencing. In this study I interviewed and observed many workers at the Court including Magistrates, Associates, Clerks and other people involved in the sentencing process. The outcome of the study of the Court was to recognise the sentencing process as one distributed in space and time between a large number of people rather than one that is performed by a single Magistrate sitting on the bench.

This thesis presents:

• A descriptive analysis of the use of speech recognition in two public service departments
• A descriptive analysis of the work of sentencing in the ACT Magistrates Court
• A novel design specification for a speech recognition system for the Court
• An approach to using various sociological approaches to study software and its use. In doing so I have shown that it is possible to use established theories from sociology to work with technology rather than inventing new approaches.
• A re-consideration of usability as contingent upon context.