Designing the Real World


Email alert. Want the next column emailed to you? Add your email address to the list:  

Latest online column All previous columns The book


Select a column:


Left or right
Funny noises


Slider controls
Off and on
Powerful functions
Safety catches
Video conferencing


A user group of two
Marks and scratches
Take Away
Stereo  vision

Time and narrative

Real time

Interaction and specifications

International standards
People flow
Messy desk

Original SIGCHI title:
May 2003. Volume 35, Number 3
(This was the last column in the print verison)


First the bad news; this is the last real-world column in its present form. Now the good news; the column continues in the same bi-monthly format on the web at

And the even better news; on the web it will be supplemented with all the columns to date and other interaction design resources. First addition will be a selection of photos from my interaction design collection; everything from bad navigation in hospitals to ice-cream menus.

And now for this month’s column; ‘saying goodbye’, or more accurately; the bringing to a close of interactions.

Everybody is familiar with the awful feelings left when an interaction has not been finished properly. Having someone hang-up the phone or or walk out on you in the middle of an argument. Placing an order on a web site and suddenly finding yourself back at the home-page. The worst offenders are film and TV programs. Spinning their perfectly crafted narrative along until, all of a sudden; Bang! It’s the interval. Or even worse; Mulder has finally got through to to the secret room where the alien with the funny head is being kept when all of a sudden; ‘to be continued...’

Consideration of saying goodbye is important in interaction design because endings and beginnings are vital parts of any interaction. Also, saying goodbye is an important human interaction and designers of public spaces need to pay attention to how the meeting and departing of people are supported in a public environment.

My worst experience of this was the design of a high-speed train that had mirrored windows. I got on, selected my seat and then turned to the window to wave goodbye to my hosts stood outside. The mirrored glass meant that I could see out through a dark-pinkish tinge but from outside it was worse; they could hardly see in at all, and my last impression as the train pulled away was of them waving and staring unseeing in to my carriage like a couple of worried looking zombies.

Another key area is the arrivals hall at an airport. People waiting to meet relatives desperately want the earliest peek possible at the person arriving. Normally, you wait by the arrivals doors leading from the customs routes to the public area outside. I have waited at one international airport where there were small windows off to one side between the public waiting area and the baggage reclaim areas. This meant that half the people waiting were not stood outside the arrivals doors but were jostling around these little windows for a first glimpse of their loved ones arriving to get their baggage. Due to a bit of bad design the users were not using the environment as the designers had intended. There was increased traffic between these windows and the arrivals doors ‘he’s on his way to the doors quick lets run to meet him’. And a real ‘un-designed’ bustle around the windows ‘is that him way over there with the green coat on? Quick lift little Robbie up so he can see’.

In this way the beginning of the interaction with the arriver was scrappily managed, but the ends of interactions also need to be well managed, clarity is vital. Whether it be the decisive bringing to a close of a salesman’s call, the formal ‘over and out’ of a radio conversation or the closing bars of a piece of music; having a clear end informs the other party that that part of the dialogue is at an end and that it time for something else to happen.

The classic example from the world of the telephone is the part human/part computer directory inquiries phone-service where you ring up to get someone’s phone number. A human operator talks to you to find out who’s number you want and then when the system finds it, a computer generated voice takes over to read the number out digit by digit and repeat it if necessary. What this means for the user is that they begin an interaction with a human and then the interaction is suddenly terminated before they realize it and there is no chance to say ‘good-bye’, and more importantly no chance to thank the operator for helping them. I imagine that it must be worse for the operators carrying out thousands of such searches and not managing to get a word of thanks from the users.

Knowing about the end of an interaction is also necessary in interactions with systems that have an effect that is not undo-able. Making a purchase, placing a bid online, sending an email. The first of these is especially important; you want the user to be able to do lots of things before they place their order (review it, change the amounts etc.) and so it must be clear to them which action is the final, ultimate ‘really place your order’ action.

So it only remains for me to make one final point. The past years of writing this column have taught me many things, but I think the key thing that I have learned in this continual review of real-world interactions is what you might call the ‘golden rule’ of interaction design and I can state it quite simply and directly. The key thing I have learned is ...

’to be continued...’

on a web-browser near you.