The evolution of records

If you’ve read the brief history of this blog you’ll know that this is the result of a diary project started several years ago. Here’s a quick note on that for those wishing to digitize their thoughts etc.

I started out recording daily events by hand in journals, the old-fashioned way. I found that as much as I deeply and obsessively love stationary, it just wasn’t worth the sheer time required to record even the shortest events. Perhaps someone with proper handwriting, working those elbows into neat, regular strokes, would have a different experience. For me, making the transition to a digital record of daily events meant a three- or four-fold increase in the amount of random crap I could deposit on a page.

Then I tried something kinky, in a literary sense. Speech Recognition. As fast as typing is, speaking is even more rapid, and being able to dictate to one’s computer and have that recorded as text would be a wonderful way, thought I, to put all my impressions of a day’s journey onto page. Turns out this was a temporary campaign.

Microsoft’s built in (with windows) Speech Recognition software is excellent. I was truly impressed with how capable the software was of transcribing my speech using both an increasing knowledge of my speech patterns (gained through regular training sessions where you teach the computer what you sound like) and a statistical analysis of word progressions. That said, it wasn’t good enough. After teaching the computer for a few days, I found that I could get it down to making one error per sentence, less if I was very deliberately dictating (but of course that defeats a good bit of the purpose of using speech). With this error rate, dictation was much faster than typing, perhaps twice as fast (surprisingly to me, it certainly doesn’t feel that way but with side by side studies I found that typing was indeed much slower). Still, it wasn’t worth it, the break in thought process of constantly having to go back and correct words was just not worth the increased speed.

Hopefully this information will be helpful to anyone consider trying new methods of interacting with their computers. As a caveat, from this experience I am confident that within a few years speech recognition will have reached a point where one can speak fluidly to the computer without a problem, depending on how much companies, and microsoft in particular, invest in the technology. Certainly if a point is reached where one can dictate to a computer just as one naturally speaks, our digital experiences will be significantly more efficient and perhaps richer.

As another note, Microsoft’s design of the non-dictation speech recognition experience i.e. navigating around the computer, the desktop space, and within programs, is also very well designed and highly efficient. I can recommend anyone with a microphone checking it out.