Hello, My idea is to dictate programming phrases into an xTalk type language [SuperTalk, Hypertalk, old Lingo for Director 6]. That means sort of a pigeon english. However, there may be a way to fool the xTalk interpreter into parsing non printing characters and so the variables could be more like Johns+cool+variable where the + is a substitute for some non printing character in this post. Example put it into Johns cool variable [blanks are really some weird non printing character] [ SuperTalk currently requires Johns_cool_variable or JohnsCoolVariable ] John F. Richardson > >1. Recognition accuracy is far below VV at similar stages of training and >>use. Is this just me, or is this a common phenomenon? > >Recognition accuracy has been about the same for me. Neither has been >very high, but my primary microphone has a weak signal, and I'm dictating >strange word orders (programming code). >