Code as Speech
As an Engineer in AI Solutions, This is How I Think Programming Might Look Like in Future.
“Code As Speech(CaS)” is not very far away with voice recognition and machine learning gaining considerable momentum. Code originates first in the minds of programmers. Then it is typed into an IDE. Speaking is 3 times faster than typing (Research), so an IDE which converts speech to code will increase the efficiency of programming. If you ever used voice for typing, you know how convenient it is. For example, I ‘wrote’ the text in this article using speech recognition on Windows 10.
I imagine the architecture to be comparable to above. From the diagram, a missing piece for Code as Speech is a unit which can convert text to code. A good start could be a mapper, we would add the intelligence part later. I don't know about the pre-existance of such a unit, but if I were to design it, I would go for the easiest and the most straightforward way.
All programming languages have its own set of rules which we call syntax. Till now, syntaxes are fixed. For example, you always declare a variable only in a fixed set of ways, you perform looping only in a fixed…
