Saturday 31 August 2013

Leap Motion’s Struggles Reveal Problems with 3-D Interfaces

Hype surrounding Leap Motion, an $80 3-D gesture-control gadget touted for its exceptional finger-tracking accuracy, reached fever pitch in the weeks before its July launch. Hundreds of thousands of people ordered the device ahead of its release, and a flashy demo video on YouTube was viewed millions of times.
Yet after one month and a raft of “meh” product reviews citing problems like difficulty controlling apps and tired arms, the sardine-can-sized gadget—which connects to a computer’s USB port and tracks the movement of your hands and fingers as they move above its sensor—seems to have lost its steam.
What happened? Much like the computer mouse and touch screen before it, it may simply take some time for app makers to get comfortable making applications that make good use of a completely new form of interaction—at least, that’s what Leap Motion and its developer community hope.
“Things involving human-computer interfaces often move extremely slowly. It may take a while before the Leap reaches its full potential,” says Robert Jacob, a computer science professor at Tufts University who studies user interfaces and new modes of interaction.
Developers say they like the app-creation tools that Leap Motion provides, and that it’s not particularly hard to build apps for the platform. What complicates things is the need to think about building apps in three dimensions, and to invent motion controls that users will understand how to use.
This can be especially confusing because there is no standard set of Leap Motion gestures, so actions like selecting or grabbing an on-screen item can vary from app to app.
“It’s something really difficult to grasp—how to teach someone to use something completely new and use it well,” says Adam Gastineau, a 17-year-old high-school senior and app developer from Dayton, Ohio, who built a $1 gesture-control app for the Mac called Swish.
User expectations may have been too high. Product reviews were middling at best (see “Look Before You Leap Motion”), and Leap Motion’s forums are littered with complaints from customers such as “thus far, I’ve had nothing but headaches with it” and “control is spastic to say the least.”
“I use mine and have fun with the games available, but it hasn’t really revolutionized my use of the computer as I hope it will once the software gets to the right point,” one user wrote.
Microsoft’s Kinect quickly became popular after its launch in 2010, but it was touted solely as a gaming controller for the Xbox, and wasn’t initially open to developers. With Leap, many consumers were looking at it as a more multipurpose device that could replace a keyboard and mouse. And while Leap Motion believes this will eventually be possible, it will take a while, says Michael Zagorsek, Leap Motion’s vice president of product marketing.
There were buyers with outsize expectations, he says, but it wasn’t the buyers’ fault. Leap is not releasing sales figures, so it’s unclear how many people have purchased the device in the last several weeks.
“I think we hadn’t realized how many people were tired of their mouse,” he says.
The problems experienced by users don’t indicate that Leap Motion wasn’t ready to be released, he says. Rather, it shows the difficulty in launching any new interface technology.
“We felt we had something people would get really excited about, and they did. But at the same time, we knew there was a lot more to discover,” he says.
As such, Leap Motion is working on improving its developer tools, some of which software engineer Kevin Horowitz showed me this week during a trip to its San Francisco office.
Waving his hand in front of a Leap Motion device connected to his PC, Horowitz demonstrated upcoming developer features like an improved skeletal model of the hand that can better predict where fingers are even if it can’t see them—like when you curl your fingers into your palm—and defines individual fingers as specific finger types (like pinky and thumb, which can be helpful for, say, virtual keyboard apps). He says pinching and grabbing capabilities have also been improved. The changes will be coming this fall.
Horowitz understands the frustrations some users and developers may feel with the device, as all the options for controls may become overwhelming.
“You have 25 degrees of freedom per hand, and nobody knows what they’re supposed to be mapped to. What is ‘selecting’? Nobody knows what it’s supposed to be at all,” he says.
Leap Motion is trying to standardize some interactions, he adds, so they will become more consistent across applications.
Developers, too, think this could help. Geert Bevin, a Manage, Belgium-based developer who created Geco, a $10 music-making app, and GameWave a $4 game-controlling app, says there’s still a “vocabulary to figure out” as a development community. Additionally, he believes developers have the responsibility to figure out the best use of the bevy of data Leap Motion can capture, and restricting it in ways that are optimal to the app.
“Everything is being experimented and invented right here and now,” Bevin says. “You have to do a whole lot of research.”
Investor attention may also aid the Leap Motion platform. Highland Capital Partners announced the creation of a $25 million Leap Motion investment fund in June. The fund recently announced its first investment: $3 million for Syntellia, which built the Fleksy virtual keyboard.
Manish Patel, a partner at Highland Capital responsible for the Leap and Syntellia investments, says the fund will put on some events with Leap Motion to encourage startups to speak with the company’s development team, and is building a set of mentors and advisors that will help startups build their interactive apps.
That may help with part of the equation. While Leap Motion says users have downloaded more than a million apps so far, there are just over 90 apps total, Zagorsek says.
For Bevin, this has translated into some fast cash: he has made about $28,000 so far from downloads of his two apps, after Leap Motion took its 30 percent cut of revenue.
Others aren’t doing nearly as well, though. Around the July 22 launch day, Gastineau says, his app was getting about 200 to 300 downloads a day. Now this has cooled to about 10 per day.
“The market’s just stagnating, and it really needs something to pop out and say, ‘Here’s why you need to buy this device,’ ” he says.
If history is any indication, it may not be that easy. The mouse and touch screen were invented in the 1960s, but the mouse didn’t catch on commercially until the 1980s, and the touch screen took decades longer.

0 comments:

Post a Comment