I connected all the pieces and built out two chat rooms to playtest on the floor. Both rooms use the structure of my Chat Woids project from last semester’s Computational Typography class, and offer a very plain interface:
The first chat room counts the number of words (or characters separated by spaces) and for each word plays a tone at a random frequency for a random duration. Due to the structure of my socket server, the random tone that the typer hears is a different from the random tone that another connected user hears. This was an afterthought, but I decided to go with it anyway in the interest of time. After connected participants press enter their words disappear and the tones start playing. Their screens remains blank. While it’s still deployed, try it here!
This next iteration sonifies words in the same way, except that this time typers’ words fly across in the screen in a flocking formation (visuals only demo here). While it’s still deployed, try it here!
Observations & Notes
I tested both sketches with two pairs of people on the floor. I brought two people close to one another and directed each of them to my server on their laptops. Then I told them to begin. That was it. No other instructions. Afterwards I asked, “What did you notice?” and “What did you think of?” Here are some rough notes from our followup conversations:
In both cases the expectation of a communication device was clear.
There was reporting of an initial curiosity and focus on how words were translated (the underlying structure). Different theories were tested. Was it sentiment analysis? Were their similarities in the sounds that might lend themselves to some shared word use? Was it related to the number or kind of letters? Was it related to the number of spaces? Or the number of words?
One person noticed and was thrown off that the resulting sounds were different on the different computers.
In both instances there was a immediate desire to communicate with their partner; they wanted their partner to understand them. Sometimes that meant pausing and allowing time for the person to respond.
One person reported that when they realized their words were not appearing, they felt tempted to share and reveal more than they normally would, but then held back in case there was going to be a “reveal” at some point and also out of concern that my server was recording their input.
This freedom of anonymity paired with a worry of surveillance was mentioned in both groups.
It didn’t take long for participants to realize that the futility of decoding their word sounds—less than a minute maybe? Some kept typing at a normal text-chatting pace, while others focused on making as many sounds as quickly as possible.
With the second iteration, the sound was secondary—again for both groups. There was confusion or amusement to the upside down and flying words. One group reported preferring Playtest #01 to #02; the other group decided that their preference would depend on the context (and our conversation shifted before we could elaborate on this).
Both groups chatted in tones for several minutes without interruption—which I did eventually out of respect for their time and our surrounding neighbors.
Next time let folks “chat” until they get bored. This might mean creating a designated area out of respectful earshot of others.
I realized that I (maybe a collective we) often take for granted that we’ll be (mostly) understood by another in conversation. And also that the desire to connect with others is a basic human instinct. So when presented with a communication device that obstructs this instinct, there’s an immediate push to figure out why and how to overcome the obstacle. There’s a tension in this disruption to explore here.
Also, a design question: how to address concerns that my server is recording participants’ inputs?