The venerable mouse, created by Internet pioneer Doug Engelbart, has been unchallenged since the dawn of modern computing. But rapidly maturing touch and speech technologies are threatening
to dethrone the mouse as the dominant computer input device.
Looking ahead five years, as the idea of a computer changes from a box under your desk to a device on your car dashboard,
or on the bathroom mirror, or in your pocket, the mouse will become less important – maybe even a distant memory.
These new touch and talk technologies are already making an impact. For example, CNN uses a touch display to present information;
Intel used one to showcase its new processors at the recent CES show. Ford uses speech recognition in its cars, and FedEx operates
its phone lines with a computer voice — with a distinctive chime that most of us recognize immediately. Apple's iPhone sparked the touch-screen revolution. Since the iPhone debut, companies such as HP and Dell have decided to bring that easy finger control to the desktop. And companies such as Nuance and TellMe are taking advantage of improved computing power and better statistical models to make speech technologies more viable than
"The magic is in grasping the human objectives of an interface and combining the best technologies simultaneously to address
those objectives," says Don Richards, the creative director at Foghorn Creative, the company that designed the Intel touch
wall at CES.
Of course, nobody is predicting that the mouse will totally disappear. Tests based on Fitt's Law of human-computer interaction
have proven that the mouse is the optimal pointing device. Engineers and programmers will continue to use them. Yet, these new products demonstrate how touch and speech technology have gone from an interesting idea to legitimate technologies.
Here are some examples.
* HP Touchsmart IQ816
The HP Touchsmart really is smart. There's a separate touch interface that runs on top of Windows. You can flick through photos,
download an egg-timer app that helps you whip up dinner, or play chess with your fingers. The model I tested is a 25-inch
monster with an Intel Core 2 Duo processor, so it runs fast enough for games and Photoshop editing.
Yet, it's the potential of this machine that has me hooked. The Touchsmartcommuinity.org showcases some interesting apps,
including one where you can interact with YouTube.com videos. HP is committed to make this touch platform work. One example:
you can sync the Touch calendar to your Google Calendar.
Another touch device worth noting is the Palm Pre, which could be the ultimate savior of Palm. The new WebOS will replace
PalmOS entirely, relying on touch control instead of a stylus. The main benefit: smartphone users prefer touch because it means operating a phone with one hand, interacting in a more physical way with the screen,
and browsing through media faster. During a hands-on test, it became obvious that Palm is serious about touch: the screen
responds quickly to any input, such as flicking a photo or ending a call. (Competing models from Samsung and HTC rely on haptics
technology for a slight buzzing sensation, confirming each click.)
Like HP, Asus is set to take the world of touch to a whole new level. They are introducing the Eee Top desktop model this year and a convertible
notebook PC called the T91. The resistive screen does not support multi-touch, which requires a capacitive screen. But Asus
deserves extra credit for the Linux operating system, which offers a painting program, a touch browser, and a nifty circular
interface that you can flick through with a finger. The most impressive feature: a table-top interface for browsing photos
where you can spread out images, ala the Microsoft Surface table.
* Dell Latitude XT2
With technology provided by N-Trig, the Dell Latitude XT2 is the very latest — and most advanced — touch computer on the market. (HP also just released the
Touchsmart TX2, but we were not able to test it out quite yet.) The capacitive touch screen supports two-fingered gestures,
such as clicking down on the screen and rotating a photo at the same time. Browsing the Web is amazingly easy: you can just
flick up and down through text, mimicking the readability and easy control of the Amazon Kindle reader. The real kicker: the XT2 may be the only touch-enabled device that should work with Windows 7 out-of-the-box when it ships later this year.
Speech recognition was a buzzword, then it was a misnomer – now it's just another technology. Speech recognition expert Bill
Meisel, president of TMA Associates, says the first companies to offer speech recognition promised too much and delivered
too little. Today, it is all about the statistical models. With each new year, companies like TellMe and Nuance accumulate
more data to learn speech patterns and dialect. Most critically, speech systems have now proven their worth by limiting voice
commands to just a few terms and slowly expanding.
"The more people talk to computers, the more data speech technology companies have to create the models, and the more parameters
they can use in these models," Meisel says.
Here are some examples:
* Microsoft Sync 3
The next version of Microsoft Sync, which will be available in Ford vehicles later this year, will use a new voice called Samantha that sounds less computer
generated. TellMe, the Microsoft-owned subsidiary, is also working on new mapping features to make Sync work like a GPS. And,
a Bluetooth phone connection could mean tapping into the Internet for voice search. Most importantly, the company uses an
entire data center to analyze voice commands. In a road test, the "recognizer" was highly accurate. TellMe has learned a few
lessons about speech: they power the 1-800 services for FedEX and many others.
* GraceNote CarStars
Imagine driving in a car on a country road. You lean back, wind in your hair, and suddenly get the urge to play an old U2
song. Instead of fumbling with a satellite radio station or pulling out a scratched up CD, the GraceNote CarStars system actually
lets you speak to your stereo. In a hands-on demo, I was able to play songs by speaking the name, and listen to the actual
artist make suggestions. GraceNote is still working on the product, but the idea is that if you're a fan of U2, then Bono
would be the one who talks you through music options.
It makes sense that a GPS device would support speech recognition – it means focusing on the road and typing with your fingers
less. The Garmin Nuvi 855T is easily one of the most advanced speech-recognition devices available for your car. You can speak
any term shown on the screen, use your voice to search for local business names, and find other points of interest. There's
still an occasional glitch: when looking for a new address, I decided to go "back" but the speech analyzer kept thinking I
meant Bakken Street. Fortunately, the GPS usually repeats back what you say, so there's also a back-door to start over or
just try saying things a different way.
* Nuance Dragon Naturally Speaking 10
Speaking to a computer still feels a bit awkward, especially if anyone else is around. Yet, as more people interact with a
PC this way, it could become more natural. Nuance has accumulated a vast 'speech vocabulary' over the past several years and
does a much better job of detecting words. Processing power – especially on dual and quad-core processors – also helps. There's
still a cumbersome training process that takes a good 30 minutes, but Dragon software is getting there. In dictating a page
of text, the program only typed a few minor errors. Sounds-like words are still an issue – the clothier Herbergers sounds
too much like hamburgers to a computer.
A touch and speech smartphone?
Heralded as the first Google Android-powered phone, G1 has revealed itself as a proving ground for innovative apps. The mapping applets – which let you swipe
on the screen to pan around local city streets – is worth the price of the phone alone.
Just in the past few weeks, two key speech applications show how your voice can be a far better input device than a finger
or stylus, especially if you're driving in traffic. TeleNav GPS Navigator for Android supports speech input for finding a
location. You just click a button such as intersection or business name, then click another button to say the term, such as
"bakeries" or "bank". You just hold down the green talk button, since the service just uses your phone microphone. With the
Android 1.1 update for G1, there's also a new voice search app where you speak a term and see the results at Google.com. I
tested at least 20 addresses and search terms and the G1 understood the meaning almost every time. I thought, why type an
address ever again?
Touch and speech technology are not the only ideas that could replace the standard desktop mouse. For example, Microsoft is
set to announce a new "touchless" technology where you motion to the screen to control the interface, skipping a touch display
and speech input altogether.
Logitech MX Air Mouse
Although this device has been out for a while, it's packed with several features that make it ideal for living room control
of a PC across the room. For one, it supports gestures such as a side-to-side movement to control volume, and has an accelerometer
(similar to the iPhone) that senses when the device is at rest or ready to start tracking mouse movement. Best of all, Logitech
skipped a complex calibration set-up process – the Air mouse actually learns on the fly.
Here's another alternative input concept. At CES, Hitachi showed off a "wave remote" technology that you can use with family
room HDTV. There's no handheld device at all, and no display to touch. Instead, to change a channel or turn up the volume,
you just move your hands. For example, to bring up the interface, you wave at the screen. The technology uses IR sensors and
a 3D tracking camera that would be built in to the television, scanning your movements in the same way the Apple iPhone accelerometer
senses movement when you turn it to the side.