Search This Blog

Showing posts with label Siri. Show all posts
Showing posts with label Siri. Show all posts

Thursday, October 27, 2011

Siri, Iris and the dream of talking to our phones

The only scene I really loved in "Star Trek: The Voyage Home" was when Scotty tried speaking verbal commands to a Macintosh Plus. Keyboards always seem to get in the way of doing what I want to do -- and nowhere is this as apparent, or frustrating, as on smartphones and tablets.

With the recent launch of the iPhone 4S, Apple's not-really-new voice recognition system Siri has been getting a lot of attention.

Yet it took the India-based software company Dexetra just eight hours to create the initial version of Iris -- a blatant Android knockoff of Siri.

Even the name "Iris," which is the reverse of "Siri," stands for "Intelligent Rival Imitator of Siri" according to Dexetra's blog. (And yes, I'm just waiting for the trademark suit from Apple.)

Granted, Dexetra wasn't starting completely from scratch. This company had already been working on natural language processing and machine learning -- two notably thorny, complex technologies -- for more than a year.

A few days later, Dexetra made an improved version of Iris available as a free app in Google's Android Market and as of this writing it has been installed more than 50,000 times.
I put Iris on my Android phone this weekend, and it's amusing. For instance, here's a discussion I had with Iris yesterday:

Me: "What time is the John Scofield concert at Yoshi's Jazz Club in Oakland tonight?"

Iris: "I have no idea."

Me: "Who is John Scofield?"

Iris: "John Scofield, born 1951, the musician." (Shows a photo of Scofield performing.)

Me: "Where is Yoshi's, Oakland?"

Iris: "Right now being pulled in by a black hole."

So Iris is about as entertaining as Siri seems to be. (I don't own an iPhone, but for comparison I've been checking out the STSS Tumblr blog, a crowdsourced collection of weird and wonderful wisdom from Siri.)

Still, neither Iris nor Siri seems very useful so far.

Fortunately, there are better voice control options on both the iPhone and Android handsets than either Siri or Iris. And they've been around for a while.

For some time the Android mobile operating system has had pretty well integrated speech-to-text functionality (Google Voice Actions). I use this often for texting, searching, navigating and e-mailing on the go.

Also, whenever I bring up a keyboard in any Android app, there's a microphone option for voice entry. Generally it recognizes pretty well what I want to say or do. If Android guesses wrong, I can use the keyboard to correct it, and it does seem to learn over time.

Then there is Vlingo, a voice control app for all the major smartphone platforms. I've tried it, and for some tasks it works reasonably well.

On the iPhone, Siri does integrate with some of Apple's own productivity tools (such as the calendar). But as my CNN.com colleague Mark Milian pointed out, Siri can't yet execute many basic commands like taking a picture.

Most importantly, Siri doesn't integrate with any third-party iOS apps, such as Shazam or Tweetdeck. Given Apple's closed iOS ecosystem, it's an open question whether such integration will ever happen.

Besides Siri, there are other iOS apps and tools that provide some voice control. ExtremeTech recently published a list.

Such baby steps are important -- but on any mobile platform, we're still a long, long way from phones that you can just talk to and they'll do what you say.

This is frustrating from the consumer perspective. Smartphones are, first and foremost, phones. They're supposed to be for talking.

Typing on any mobile device, through a physical or virtual keyboard, is a chore. That's why QR codes are getting popular -- they eliminate the need to type on a mobile device.

The challenge of typing on handheld devices is exactly why the kind of voice control Scotty expected is such an alluring and intuitive idea. Eventually we'll probably get there.

Tuesday, October 25, 2011

Google's Ice Cream Sandwich, Siri talks back

On this week's Tech Check podcast, Doug Gross, Stephanie Goldberg and Mark Milian discuss the roll-out of "Ice Cream Sandwich," Google's delicious new version of its Android mobile operating software.

Google debuted Android 4.0 this week in Hong Kong, along with the new Samsung Galaxy Nexus smartphone, which will be the first gadget to run it.

The crew discusses its features, including Face Unlock, which is designed to recognize the user's face instead of a password.

Mark has spent a week testing Siri, the new voice-activated "digital assistant" on the iPhone 4S. We discuss her strengths and weaknesses, as well has giving in to the irresistible urge to ask her questions from "2001: A Space Odyssey."

Our Reader Comments of the Week come from a story about WireDoo, the new search engine being developed by a team working with MC Hammer. Yes ... that MC Hammer. (As if there could be two).

And we present to you a Tech Fail of the Week that was 14 years in the making.

Michael Dell, founder of Dell Inc., was asked in 1997 what he would do if he were the CEO of Apple. This week he got a long-awaited opportunity to revisit his fateful words.

To listen to Tech Check, click on the audio box to the left. To subscribe, you can add Tech Check to your RSS feed here. You can also listen, or subscribe, on iTunes.