|Image from vanRijn|
But the interview is not about intention. It is about capability. The Kinect is able to multiple identify people, both by voice and by vision, in noisy, chaotic environments. People are already starting to hack the Kinect to allow Minority Report interfaces and the like.
The second thing I heard was on Hacker Public Radio, which I started to get into after Indiana LinuxFest. The podcast itself is a pre-interview discussion between KDE spokesman Aaron Seigo and Jonathan Nadeau of Frostbite Systems.
Accessability questions have been floating in my head recently. There's lots of buzz about Section 508 in the circles I float through, enough so I've attended two compliance talks in 2011, and my wife has been trying to be a productive person while having severe pain in her right elbow and forearm that keep her from being able to type. She's been playing with speech recognition software, finally trying the Windows 7 native setup that came with her new laptop.
But honestly, when thinking about accessability, I wasn't really thinking about accessibility. I think that there's research at the end of the line for this train of thought that will serve to help accessibility, but it isn't foremost in my head here.
I was driving. I was thinking about cars.
As I think I've mentioned, I've been looking at some travel applications. I tend to have either Listen or a media player of one sort or another playing when I drive, rather than CDs or even worse, radio. (I have only one objection to Amazon Cloud Player, which is that it wants to run only on WiFi, and by and large, I want to use my phone as a media player when I'm mobile, away from WiFi, but maybe the coming of 4G or the next software fix from Sprint will fix that.) I have a TomTom GPS showing my location and speed and I sometimes use Sprint or Google for directions, too. Also, I've been looking into Vlingo, which gives a voice interface to phones. I noticed a bug, hopefully fixed soon, where I hit the Talk Now button while listening to a podcast, and told it to call my wife. Normally, when you tell it to call, it shuts up the media player, Listen included, but here it didn't. So, I'm trying to pause the player and/or mute the player, while on the Interstate.
I know. I know. The point of the exercise is to keep from having this sort of attention-sucking frustration from occurring while I'm driving, because I don't want to hurt myself, my stuff, or other people and their stuff, pretty much in that order. Haven't been too keen of Vlingo since. It's cool enough, but you shouldn't have to press a button while driving to say "I wanna talk now".
And I don't think the Vlingo is alone. Watch just a few high-end car reviews on Top Gear and you'll know that the UIs for even fancy cars are crap.
Ultimately, to do a driving interface right, you have to assume that there will be no vision at all. Most of the things you do while driving (besides avoiding obstacles and other drivers) occur without vision: you use your sense of touch to feel the vibration and you hear the engine to know when to shift. You see the red Check Engine light but you take it in when it starts to sound or feel funny. You can even get away with not often checking the speedometer by staying at a similar speed to your fellow drivers. If you're going to do much computing while driving, it'll have to be via voice. And once that shows up in high-end vehicles, that browser is going to rival Firefox and Chrome and IE at the top of the browser food chain.
(As a pure aside, I think Opera will go there. I don't like Opera and haven't for over a decade, but I know I'm biased on the subject, but it seems like a place that Opera would go, more than anyone else.)
But as we needed the keyboard and then the mouse to get the WIMP interface for computers into the office and the home, we need to rock a voice interface to get the computer into the car. Which is where the Kinect comes in. I kinda think that the car will be a big docking station for future computers, as seat settings and address books and the like are highly personal but a great deal of hardware is there for anyone who sits down. Kinect-based technology would be necessary to distinguish between the driver, the passenger in the front seat who talks with his hands and the loud kids in the back seat. The recognition between conversations between user and computer and conversations between driver and passengers is crucial, and one I think that Microsoft Research and Kinect are closest.
There's going to be a small LED screen for the backup camera in the New Car's rear-view mirror (and some cars of today already have it) and a Kinect as well. Of this I am sure.