NEXTVR KEYBOARD
NextVR's audience enjoyed free access to broadcast quality stereoscopic sports and entertainment for quite a while before we considered monetization. There had been no need for authentication (identification) or authorization (access control), but we needed to introduce accounts. How would people enter credentials?
NextVR's audience enjoyed free access to broadcast quality stereoscopic sports and entertainment for quite a while before we considered monetization. There had been no need for authentication (identification) or authorization (access control), but we needed to introduce accounts. How would people enter credentials?
This was still early days for VR, and most of the seven platforms I designed for didn't come with hand controllers or have system-level keyboards. Some weren't stereoscopic, and there was a wide range in terms of field of view, resolution, even weight. Engineering was building each platform from scratch; there was no middleware, so the design had to consider each platform's abilities. I started to think about friction and consistency.
Assessment
I wasn't allowed to do research with our audience, so I did interviews internally. I was surprised by the amount of passion around this issue. Vocal opponents thought this path would kill the company. They were terrified that it would be so difficult for people to type that acceptance of VR in general would be damaged. People imagined either targeting letters all over the field of view with their heads (there was no eye tracking) and tapping a headset button to select each letter, or waiting for a countdown timer to select each letter.
I reassured everyone, promised to keep them in the loop, and convinced them to reserve judgement.
Diving In
I believe in trying as many things as possible, rather than designing a few solutions in my head or running with one obvious solution. I researched existing solutions, then began a broad exploration. I prototyped a cylindrical alphanumeric band that encircled the viewer. I investigated voice input (Engineering deemed this a non-starter). I designed keyboard layouts around letter popularity, where heavily used letters were bigger and easier to target. I tried traditional QWERTY variations, and other configurations. I used depth and distance, then returned to two-dimensional panels. There were dozens. With each, I narrowed and refined the problem space, while warming people to the process.
▲ A few of the rough explorations
Evaluation
I developed a way to get static wireframes from Photoshop into VR to 3D prototype quickly. This let me put the most interesting solutions in front of people regularly, learn, redefine, then iterate. With each keyboard I measured two things: first, I asked people to pretend to enter their email address; I timed this, then asked them to rate how it felt on a 10-scale.
There were many comments about neck movement. Smaller keys led to less sweeping neck strain, but more frustration as people struggled with fine motor control. Looking up a bit was considered hard, but looking down less so. I concentrated activity into a 30 degree area, and moved 'center' to about 5 degrees below the horizon. Gradually, we zeroed in on something fast and comfortable enough.
Eureka
Designs eventually led to a fairly familiar keyboard with some distinctive features. For emails, the space bar was unnecessary, but it was a nice large target for a submit or accept key. A bit of curvature went a long way with testers (more distant targets were larger), as did moving the letters slightly forward in space. Color, contrast, and transparency were all tuned.
One more thing...
I identified one persistent issue that I wanted to crack, though, even though people were already satisfied. People would enter a character, then glance above the keyboard to check the entire email string. They didn't notice or report it, but I knew I could shave some time and friction off.
I ended up punching up the visual and audible feedback for highlighting and activating a key. I also snapped the large reticle to keys instead of letting it float completely freely. Most notably, I attached the string to the reticle, so that the most recent letters added were always in sight.
Success
The keyboard was navigable by head targeting, direction pad, or beam-pointer. It was quick and felt comfortable. It worked equally well, looked consistent, and was readable on all platforms. Even the most antagonistic detractors were converted to fans. The company liked the solution so much that when system-level keyboards were eventually available (and I recommended using them), everybody wanted to stick with ours.
