Introduction

Nerve damage or any kind of nerve impingement can halt the job of a programmer until it heals. If it does not heal, then it can end their career. Solutions have been attempted via voice recognition but they are extremely rigid (i.e. one cannot customize the software to adapt to the user), are proprietary, or are unable to be run or compiled due to outdated or missing software/libraries. Therefore, to address these issues, we developed software with the intent to use a free speech recognition engine, that users can easily customize and become productive in a short time.

Approach

The application was designed and a set of users volunteered to test the software and provide feedback. In a preliminary trial, a group of five users were asked to use the program after a few minutes of instruction. Simple tasks such as coding an algorithm by voice or developing a simple data type were requested if the users did not bring their own idea to implement. Using what they learned, they carried the task out and after 30-60 minutes of usage, they were questioned on what they liked and didn’t like, and were asked for recommendations to the program.

Analysis

All of the users were impressed at the speed and easy usage of the application. The only negatives that occurred (which was consistent among all users) were:

  1. How the recognition engine would abort prematurely, and then start immediately processing the subsequent buffered text, yielding unexpected commands (likely due to the subsequent point).

  2. Voice detection requires the usage of trained software specific to the user, or else the recognition rate is too low and it becomes cumbersome to use the application.

  3. A default library of various languages should be provided as an option.

Conclusion

The software is easy to use and, combined with a trained recognition engine, can be just as effective as typing for large tasks. Users were able to effectively utilize the program with very little instruction and quickly create grammar definitions to yield solutions.

Acknowledgement

A thank you to my supervisors Danny Heap and Daniel Zingaro for assistance throughout the project.