Bruce's post about Chuck Runquist's note on the future of WSR reminded me of something I've been meaning to ask. Has anyone been playing around with voice-driven apps on their cell phones? If so, what's your experience been?

Back in November, I bought an iPhone. Since then I've been experimenting with 3 applications that use or are tied into voice:

Jott -- either call a phone number or tap a button on your iPhone, and then you can record up to 15 seconds of text that are transcribed by combination of computer and humans. Jot is set up so it's easy to integrate with a variety of other Web 2.0 friendly websites/apps. So for example, I use jott to add to-do's to Remember the Milk, a terrific task management program that you can access either on the web on your iPhone. So, I can say, "Pick up David's present on Saturday" and Remember the Milk via Jott saves the task and gives it a due date of Saturday. Jott is easy to use, and for tasks 15 seconds is plenty. The voice recognition isn't perfect, but for most tasks it gets it right -- and when it's wrong, I can usually figure out what I meant to say.

Reqall -- a similar service. Reqall lets you record up to 30 seconds, and it has a bunch of other features that I haven't used (although being able to grab the notes I transcribed using Reqall via an RSS feed is pretty nice). However, integrating it with other services is tricky, so I'm mostly using it for notes. Sometimes the voice recognition is as good as jott, but sometimes it gets flustered and produces gibberish -- and unlike jott, it doesn't appear to let you spell out words you know it's going to have trouble with.

Google Mobile -- I love, love, love Google Mobile. It has a number of nice features, but the best one is that you can give them an address to search by voice. Like Jott and Reqall, the translation is done back on the mothership website. So far, it's been pretty damn good. And that's great news, because it's much, much easier saying an address than type it using the tiny iPhone keyboard.

After years of using NaturallySpeaking and Macros -- first naturallyspeaking's, then Vocola -- it feels a little weird using voice services where you have less control than I'm used to. But I'm pretty impressed what they've done so far, and they're just getting started.

The one catch: it's quite possible that Microsoft will decide to focus its voice-recognition efforts on competing with mass consumer services like these and won't really develop the features WSR needs to give NaturallySpeaking a run for its money. I hope not; what Microsoft's done with WSR is pretty impressive.

Thanks,
Anders