Rethinking thinking and ambient displays
Monday, Oct 06, 2003
Probably my hottest passion in HCI is the concept of ambient displays. Ambient displays abound in nature. The way the spectrum of light changes a few minutes before rain starts falling, the way birds waking up at 4am let you know you've been studying too late again, your sense of balance. These are all examples of ambient displays.

In the HCI realm, the hot geek project de l'annee has been to create novel man-made ambient displays. Classic examples are the network activity dangling string and a thicket of waterfall and windchime-oriented projects. In the past few years most of these projects have been filed under the concept of 'ubiquitous computing' though I think this is a bit of a misnomer, since if an ambient display is truly ubiquitou, the 'computing' portion of it should be invisible to the user and therefore should no more have the label 'computing' than a car or DVD player should. The advancement of the field comes in the expression, not the computing.

At any rate, I've been wanting to create ambient displays at home for quite a while, but time, money, or other factors always got in the way. Now that I'm settling in to a new home, the desire to create an ambient informatic environment has risen anew, and I've spent the last several days thinking about two things: What form could these displays take, and what information do I want to display?

Though I don't have a shortage of answers for either of these two questions, I often find a disconnect between the two lists. Without any 'in the world' relationship between, say, traffic to and the sound of flowing water, that relationship has to exist in my head. Therein lies the problem, because there is a deliberate cognitive step that has to happen in my head when I hear the water surge briefly to understand what that display maps to in the real world. Further, someone who hasn't explicitly been told about the relationship between flowing water and my web site traffic (or in the linked example, the dangling string and the office's overall network activity), would never make that connection. This brought me to my first realization:

All ambient displays are learned.

Whether it's the flat sunlight before an imminent downpour, or the birds chirping at 4am, these displays only become effective as the user makes the connection (causal or otherwise) between the two phenomena. In the most effective ambient displays, this connection happens unconsciously, so that not only does the subject not know how they know it's about to rain, but they don't even notice that the light outside has changed.

In the network-string example, it's likely that the information needed to correlate the string to network traffic isn't available to the user, unless they start to realize that their web-browsing gets slower at the same time as the string gets more energetic. In the website traffic and water example, there is even less data to correlate because my website traffic is a metric completely hidden from someone sitting in my living room. The data that the subconscious brain needs to create this binding simply isn't available, and so explicit knowledge is required, negating the very nature of ubiquity.

To take it a step further, I believe that the linkage between the display and the underlying data should not only be available to the subject, but it must be available in a way where it is internalized inexplicitly. In other words, just having a sign saying "this string's activity indicates network traffic" won't do, because the knowledge of the linkage, while in the world, still has to be internalized consciously, and after the first handful of interactions with the display, the user will carry the knowledge in their head, but in their conscious attention.

This creates a direct obstacle to ubiquitous assimilation of the display's information, because a short-circuit to the conscious level has been created. When the subject encounters the ambient display, they think about the display and their explicit learned linkage, eliminating the opportunity for the display to affect them of its own accord.

It's like stopping hiccups: The most successful and difficult method to succeed is to think about something else entirely, only you can't, because you keep polling yourself to see if it worked, at which point you hiccup. By trying to use an ambient display ambiently, people will often try to see 'if it's working' which means it can't. When a linkage between display and data happens in the subconscious, there isn't that conscious recurrent check to see if it's working, because the conscious mind was never given a role in the experience.

So what makes an effective ambient display? What is effectiveness? Is ambience and/or ubiquity the most important facet? Or is it the fidelity to which the changing data is realized in the subject? It must be a middle ground, where explicit data is sacrificed for the sake of 'calm'. A cellphone ring is not an ambient display, while a static painting falls on the 'overly calm' side of the spectrum: a display that might have a deep meaning, but no change over time.

I'm still doing a lot of thinking on the subject, but rather than running headstrong into waterfalls and colored balls, I'm taking a step back and approaching from a research perspective. I'm going to start keeping a log of the ambient displays I sense every day, how I interact with them, and how I learned the relationship between the display and the information behind it.

My next step will be stretching a few of these displays a bit farther from their data, and see if they still work. For example, right now it's very quiet in my apartment because it's 1am. The ambient noise level is a display telling me very roughly what time it is. If I tied this kind of relationship to my radio, so that it grew softer as the evening wore on, and grew louder in the morning, mirroring the average change in ambient background noise, it might give me a better indication of the time of day, both when I'm staying up too late blogging, and when I should be getting up to start the day. In this respect it might serve as both an ambient alarm clock and 'time to sleep' notification, without any of the abruptness of a clock-radio. The most important difference here is that this radio doesn't attempt to tell you what to do or when, it simply gives you a better sense of the world around you.

Approaching the problem from the other end, I should take a look at the pieces of data I want that aren't adequately addressed by ambient displays. Then I need to find the right way to extend that data into the real world, as opposed to creating a display and an arbitrary linkage.

These are slow steps, but hopefully the results will have a greater utility to wow-factor ratio than most of the ambient work I've done so far.

If you like it, please share it.

Hi, I'm Kevin Fox.
I've been blogging at since 1998.
I can be reached at .

I also have a resume.


I'm co-founder in
a fantastic startup fulfilling the promise of the Internet of Things.

The Imp is a computer and wi-fi connection smaller and cheaper than a memory card.

Find out more.

We're also hiring.


I post most frequently on Twitter as @kfury and on Google Plus.


I've led design at Mozilla Labs, designed Gmail 1.0, Google Reader 2.0, FriendFeed, and a few special projects at Facebook.

©2012 Kevin Fox