Ideas on how to map spoken name to actual, in code?

[Jason Veldicott]

A few ways to dictate a name in code:

* Use voice commands as a virtual keyboard, spelling out individual chars and symbols.  This can be time consuming for a just single name, especially when components of the name are not full words that DNS recognises.  Not a preferred way to do it.

* Say first character in IDE and rely on popups. I'm not sure this always works. But even if it does, the act of choosing from the popup involves selection, and this interrupts speech flow.  

* Rely solely on speech, speaking the name in some fashion that is mapped somehow to the actual name.  

The last would seem to be the most desirable approach as speech flow would not be interrupted by popup selections.  

e.g. the name "ref1A_Article" might be *spoken* (as distinct from recognized words) as "reference one a article", or "insWordIntoDocPtr" as "insert word into document pointer".  

The AST of Eclipse could be use to grab names in the current scope.  But how exactly the mapping from spoken words to actual names would occur is unknown. 

Has anyone considered this problem or even tackled it?

I believe in VoiceCode this has been implemented.  A brief look at its source code did not reveal much as to how this mapping is done.

Jason

[Mark Finker]

Hi Jason,

I'm trying to get a voice setup going to do coding with as well, and it's some way from perfect.

It would be handy being able to speak identifiers, although I cannot offer much help as to how it could be done. Maybe someone in the VoiceCode group on Yahoo might know.

Mark