When Nuance acquired Seattle startup Swype last year, it wasn’t just another startup buy for the Burlington, MA-based mobile-input juggernaut. It was actually a kind of reunion, bringing together a core group of mobile veterans who had worked at Tegic Communications in the mid-1990s.
If Tegic’s name doesn’t ring a bell, think back to the T9 keyboard input software. That particular innovation expanded the usefulness of the old-style cell phone keyboard, allowing a user to tap away on the phone’s nine numbered keys and predicting the words they were trying to spell. (For you whippersnappers, that was a big deal because texting otherwise required several taps of each key just to decide which letter you wanted to use. The horror.)
When Nuance acquired Tegic from AOL, it got a bunch of patents and people related to the innovations in T9. But it didn’t get Cliff Kushler, the co-inventor of T9. Kushler then went on to found Swype, which brought a whole new way of navigating touchscreen keyboards.
Rather than hunting and pecking across a touchscreen keyboard, Swype allowed users to trace a path between letters in the word they wanted to type. Swype’s technology analyzed the paths that the user’s finger traced, and used that data to predict the word that you were trying to type—with remarkable accuracy.
Nuance had its own version, called Trace. And the former Tegic guys, at Nuance’s Seattle office just down the street from Swype, had a pretty healthy little rivalry going. But these days, they’re in the same office—and as of today, the new version of Swype is now available in beta for Android devices.
It combines the finger-tracing input of Swype with traditional tap-typing, Nuance’s Dragon voice recognition, and the ability to draw letters on a touchscreen (which is particularly useful for character-based languages).
Aaron Sheedy, the former Swype COO who is now a vice president at Nuance, says there are more fun combinations going on behind the scenes. At the time of the acquisition, both Nuance and Swype were working on ways to improve their respective products’ ability to predict the words that a user was trying to type.
The Nuance folks were approaching the problem from the perspective of improving their software’s advanced language model, which is pre-built and sits on the device, ready to work when the user gets the application installed. Swype was approaching the problem by trying to build a dynamic language model, which arrives as essentially an empty database that can then learn from the user’s input and start predicting what word strings they might use.
“It was actually incredibly fortuitous,” Sheedy says. “We took both models and were like, `This is perfect.’”
In a weird way, combining the two former Tegic teams also allowed both sides to crank out keyboard innovations a bit faster, Sheedy says—they weren’t having to constantly make sure they weren’t infringing on one another’s patents.
“Now, you can sort of take the gloves off. Both teams can work as aggressively as they need to, to build the product the best possible way given the foundation of IP that we have,” he says. “That’s actually refreshing, to not have to work under those constraints.”
Nuance may be a big dog in mobile, but as computing continues converging into smarter and more portable devices, the company sees its predictive input technology taking a big role.
“Of course, our aspiration is that it starts to propagate across devices as well,” he says. “So in some of the future versions that you see, we’ll start taking some of these language models and really start moving them around. So as you move form your television to your laptop to your tablet or your phone, these things can start to follow you around as appropriate.”
So when will it move to iOS devices? That’s a bit of a tricky question, of course—as is anything dealing with perfectionist, sometimes paranoid Apple.
Here’s Sheedy’s appropriately careful take: “It’s not hard to state that we would love to explore how we bring our keyboard technology to the Apple ecosystem,” he says. “And certainly, Nuance’s relationship with Apple is healthy.” (Nuance provides the speech recognition technology inside the Siri personal assistant app.)
