Guest commentary from Bill Bryant: I have been in the software business for more than 20 years now, spanning some 25 companies (including such notables as Netbot, Visio, Qpass, Isilon, Medio, and more recently Opscode, Winshuttle and Z2Live). Across those companies, there have been exactly two times when I was completely awed by a technology demo – that moment where your jaw drops and your first question was: “HOW did you do that?”
The first time was in 1992 when Ted Johnson, a founder of Visio, demoed an early alpha version of Visio to me during a job interview.
He showed me two things — drag-and-drop (which was just beginning to be supported by Windows 3.0) and then drawing a freehand circle with a drawing tool. It was early days and the UI was still primitive but the magic of dragging a shape from a palette and dropping it into a workspace was amazing for its time. I knew at that moment that I wanted to work at Visio, and was fortunate enough to be part of its early success.
The second time was May 2008. I was serving as a judge for the MIT Venture Lab, a program that MIT Enterprise Forum periodically runs allowing very early-stage startups to demo to a panel of judges in the hopes of making it to a larger public showcase venue.
The typical cats and dogs had trooped through during a long afternoon. The judges were meandering somewhere between jaded and bored, looking at the clock and thinking about dinner plans.
The next presenter was announced as “Cliff Kushler, who is going to demo a technology he’s been working on for text entry on touch screen devices.” Yawn. Whatever.
Cliff walked into the room to set up his demo, which used an old Windows CE touch tablet. He called up an alternative soft keyboard display, took a stylus, and started to “connect the dots” or “fingerpaint” across the keyboard.
Magically, words began appearing in Notepad (which Cliff was using as the target application). Cliff’s stylus was going faster and faster, and like some kind of black art, the text began flowing and sentences started to take form. Whoa. Wow. “Do that again, please.”
And then it got better.
Cliff took his fingernail (it had to be a fingernail because the display glass was not the touch sensitive form factor we’re accustomed to today) and started to enter text by trace pathing the words using his nail.
After retrieving my jaw from the floor, I asked to try it and lo and behold, the tech worked flawlessly. Cliff quickly showed how to capitalize, do double letters, and other nifty shortcuts.
He then talked about how he envisioned this technology being used on the coming generations of smart phones and tablet type devices, as well as anything with a display that required text entry (public kiosks, in car navigation systems, Microsoft Surface type products – ANY device that had a display but no attached keyboard).
I was blown away by what I had seen and knew that what I had just seen demoed was the basis for a company.
After the MIT event, I approached Cliff about how I might get involved, and introduced him to Mike McSherry, a founder of Boost Mobile and Amp’d Mobile, who had recently returned to Seattle and was looking around for a new project in mobile.
Mike and Cliff bonded, a company was formed, we scrambled to find an appropriate public launch venue (which turned out to be TechCrunch in October of 2008), and three short years later the company has now been acquired by Nuance.
While Swype looks like an overnight success, what this timeline overlooks is the five years that Cliff and Swype’s co-founder, Randy Marsden, had previously spent working on perfecting the core algorithm that was underneath the hood.
Cliff’s algorithm incorporated something like 40 different variables (velocity of the finger through the keyboard, proximity of letters…) to accurately predict what the user intended to input. Cliff and Randy had to account for the fact that we are fat fingered and thus text entry using a finger was naturally going to be less accurate than using a keyboard.
And even this elongated timeline ignores the fact that Cliff conceived of Swype (and Tegic T9 before that) due to a decades long passion around assisted technology for individuals with various physical disabilities.
T9 was conceived as a way for people who couldn’t use a full keyboard but could still use a 10 number keypad (for instance, people with one hand or crippling arthritis) to enter text. The software has now been shipped on over three billion phones.
Swype was conceived when Cliff observed how paraplegics could still use an eye controlled laser pointer to enter text.
Every so often, entrepreneurs do well by doing good. Congrats to Cliff, Randy, Mike, Aaron and the rest of the Swype team.
Bill Bryant is a Seattle angel investor and venture partner at Draper Fisher Jurvetson. He was an early backer of Swype.
[Editor's note from John Cook: I too was on the judging panel back in 2008 when Kushler first presented at the MIT event, and recall all of the judges being equally impressed with the demo, which won the top prize for the night].