Home > Uncategorized > The sound of code

The sound of code

January 15th, 2009

Speech, it is claimed, is the ability that separates humans from all other animals, yet working with code is almost exclusively based on sight. There are instances of ‘accidental’ uses of sound, e.g., listening to disc activity to monitor a programs process or in days of old the chatter of other mechanical parts.

Various projects have attempted to intentionally make use of sound to provide an interface to the software development process, including:

    People like to talk about what they do and perhaps this could be used to overcome developers dislike of writing comments. Unfortunately automated processing of natural language (assuming the speech to text problem is solved) has not reached the stage where it is possible to automatically detect when the topic of conversation has changed or to figure out what piece of code is being discussed. Perhaps the reason why developers find it so hard to write good comments is because it is a skill that requires training and effort, not random thoughts that happen to come to mind.
    Writing code by talking (i.e., voice input of source code) initially sounds attractive. As a form of input speech is faster than typing, however computer processing of speech is still painfully slow. Another problem that needs to be handled is the large number of different ways in which the same thing can and is spoken, e.g., numeric values. As a method of output reading is 70% faster than listening.

Unless developers have to spend lots of time commuting in person, rather than telecommuting, I don’ see a future for speech input of code. Audio program execution monitoring probably has market is specialist niches, no more.

I do see a future for spoken mathematics, which is something that people who are not a mathematicians might want to do. The necessary formating commands are sufficiently obtuse that they require too much effort from the casual user.

Comments are closed.