Talking to Myself

Talking to a computer takes a little getting used to.

I started using voice recognition software a couple of years ago, when an occupational therapist told me she used it to dictate her notes. I figured if somebody used it who had the option of not using it, it must be ready for prime time. I could still move my fingers then, but they were really numb. I had never been a hotshot typist anyway and was going downhill from there, so I didn't have a whole lot to lose.

After spending the last couple of years getting used to it, I have to say it's way better than the alternative. My fingers are still mostly numb, and now I can't move them very much, either. That whole keyboard thing is just not workin' for me anymore.

I noticed a pretty gradual transition from using a keyboard except for when I was really tired, to using voice recognition except for when my hands are working particularly well. (In my case, 'working particularly well' means being able to poke keys one at a time with the end of a pencil. I'm not talking 10 fingers, 120 words a minute. Not that I could ever do that anyway.) As my hands got worse, I begin to rely on the software more and more. The more I used it, the better I got at it. Funny how that works.

I'm still mildly incredulous that people who are perfectly capable of using a keyboard would choose to use voice recognition. The people who sell the software say they sell it to doctors and lawyers and such, and I know they couldn't say that if it weren't true, could they? I may have mentioned that although I was never any kind of speed demon at the keyboard, I'm reasonably certain I could get words on a page faster that way. No matter. That train has left the station, so let's start from where we're at.

Voice recognition is better for dictating text than it is for editing or revising. Even though my job title was technical writer, I always spent a lot more time on the job tinkering with existing text or revising a draft than I did coming up with new words. You can use voice recognition for that, but it's just like work. Unless, of course, the alternative is poking keys with the end of a pencil.

Voice recognition is most accurate if you speak clearly. It also helps if you know what you're going to say before you say it. Late in the afternoon, when I'm tired, a little fuzzy, and my voice gets weak, recognition accuracy goes straight in the can. The good news is that I never misspell anything. The bad news is that typos look like typos; everybody knows one when they see it. Recognition errors are totally invisible until you click Send, and they make you look like a total bonehead.

I bet even the doctors and lawyers who use voice recognition for dictation don't use it for moving around a desktop or a webpage. You can use voice recognition for that, too, but I suspect the only people who do are those with a disability that prevents them from using a keyboard or a mouse. Slow and awkward ain't the half of it. If you can still use a keyboard or mouse at all, don't throw them away.

My introduction to voice recognition took place in an office, with a door.  Amazingly, it was only after I retired and began to use voice recognition at home that I realized it's a lot like reading all your correspondence out loud. For an audience. And even if they don't mean to, when I start talking anyone in earshot invariably says:

"What did you say?"
"What?"
"Did you say something?"

I don't blog, IM, or respond to e-mail on weekends, and won't until we get an office with a door.