I'm not sure if this is a comment, a question, or a suggestion. I do know
that Vista voice recognition still has a few problems. Certainly for me. I
use it to translate from Dutch into English. The easiest way to do so is, of
course, to dictate into an existing document, replacing the Dutch text with
English text. This means I can for instance leave unusual words or proper
names, deleting everything in front and then carrying on afterwards.
I do however notice an incredible problem developing with Vista's voice
recognition in this context. You see, somehow Vista decides it wants to read
everything that's in the word document, including all the Dutch text. This
means that when translating a text about Neoliberalism, I suddenly find the
word "Neoliberalisme" - with an E on the end - because that's the Dutch
spelling and that word was already in the text. :(
In addition, I use UK English, but for some reason this there seems to want
to mix and match UK and U.S. spelling. Recognise and honor!

Very strange...

Martin