Devices, Interactive Design, and Dancing

Devices, Interactive Design, and Dancing

I attended PAX West as an exhibitor for a game I, as part of a 4-person dev team, created for the new Bose AR technology.  Their latest hardware (two headphones, QC 35 II and N700, and Bose Frames) are equipped with accelerometers, gyroscopes, and rotation vectors. At the booth, I observed people playing our game for 12 hours over four days, with average play time between 5-10 minutes. I’ll focus analysis on the N700s as this was the device we had players use the most.

In our game, The Clairvoyant, players are psychics, able to hear and assist ghosts who haven’t yet passed on. Players listen to music and characters, and interact through yes/no gestures, turning, and performing specific actions, such as dancing. Our game is designed to be playable without looking at your phone (where the app is located).

Prior to the event, I tested the game several times on the N700s. From my end, it worked well! But, I had made two key assumptions in deeming the game “playable”:

1. Players understand the hardware
2. The way I move is similar to how many players would move

Naturally, these assumptions were challenged when faced with 100+ players trying our game, and the hardware, for the first time.

I witnessed what Don Norman calls “the paradox of technology”, where “the same technology that simplifies life by providing more functions… complicates life by making the device harder to learn…” (34). The N700s are designed to be intuitive, with touch controls and natural gesture tracking. In reality, these “intuitive” controls require specific instruction to understand. The touch control area is only on the front half of the right cap, a little difficult to find quickly when you are wearing the device, and with so much you can do in that small space, it’s easy to make a wrong gesture. For example, the game requires you to tap and hold, but a quick tap once will play or pause music. With touch controls, the headphones provide an array of affordances, but no signifiers on the device. The front right half of the cap looks and feels exactly the same as the back half and left cap, and verbal instruction was required.

Dancing!

As for the movement, the base code for yes/no gestures did not prove to effectively register the wide range of nodding and head shaking people used, making the gesture more difficult than it appeared. When I tested, I would do a curt nod once for “yes”, but the vast majority of players nodded quickly and many times in a row. Continual movement delayed feedback from the game, which caused frustration for players. We had more success with dancing, a gesture we created that tracks any movement above a certain threshold, designed to be easy regardless of how players choose to move. Even still, some people danced without moving their heads enough to trigger the device.

This experience certainly highlighted more design challenges to overcome!


Works Cited:

Norman, Donald A. The Design of Everyday Things. Revised and Expanded, Basic Books, 2013.