Recently I presented a lecture on the use of speech recognition to a computer club. This was the second presentation I have made on the subject. The two clubs I addressed have a partially overlapping membership.
I started with the observation that history is not complete. Computers are still changing, and the input devices are changing. Speech recognition is difficult for a computer, but with increasing power and more sophisticated software, using speech recognition for personal computing can be both efficient and enjoyable. Some of the old-timers in the club were obviously skeptical based on their exposure to early versions of speech recognition on early PCs.
So I deliberately demonstrated the speech recognition function included with Vista and Windows 7 on an older laptop with only 2 gig of RAM. It worked fine. (BTW, if you have not used the included speech recognition package, it can be found in on the accessibility page.)
The thrust of my talk was that the state of the art is powerful enough that one can easily substitute speech recognition for a keyboard and mouse with much less effort and training than is typically spent on learning how to type and use a mouse. In fact, for some applications, speech recognition is definitely superior.
At this point one of the members who had been at my previous talk interrupted. He said that he had been inspired by speech recognition because he had an old handwritten journal that was frayed and generally is bad shape and not suitable for scanning with OCR, but he wanted to convert it to digital text to distribute to friends and relatives of the author. Deciphering a sentence and then turning from the book to enter it on the keyboard was slow and prone to errors. The answer, which worked very well, was to set up his PC with speech recognition enabled and take the training tutorial. Shortly afterward he was able to concentrate on reading the old document and dictating what he read. Even as a raw beginner, the results were very satisfactory with few errors. He is a believer.
Stories like that make the work of preparing a presentation worthwhile.
However, in the interest of full disclosure, I have satisfactorily dictated both newsletters and blogs, but this one was typed the old-fashioned way. Will I ever change to primarily speech input? Probably I will. As I said, history has not stopped. The way we interact with computers is changing before our eyes. Check on a flight by telephone and you will be talking to a speech recognition computer. Play a new game, and it will even be watching you to take clues from your motions. That is, game playing now uses non-contact motion to program a computer. The main reason I type still is that I spent many years building up an unnatural skill to the point that it now seems natural. Even at that, with only minimal training, I can dictate about as fast as or faster than I can type.
Finally, I predict that keyboards will not vanish entirely. Just as we use a vacuum cleaner, but still have brooms, so I think new technology will displace older technology, but the older technology will not vanish entirely.