As an example of what I feel is absolutely *perfect* interaction design,
I'd like to nominate the jetBlue kiosk for self check-in. I'm new to
this list/field, but these things look like the gold standard to me.
The first time I used one I felt like a kid at disneyland; it was fun,
friendly, but more importantly so easy to use that by the end I felt I'd
made a new friend!
I was one of the poor souls caught in New York during the "blizzard" a
few weeks back, but my annoyance level never rose above mild thanks in
large part to the experience of those kiosks (and the obvious attention
to customer service that the kiosk represents).
Krug talks in DMMT about the "reservoir" of goodwill, and how a full
tank can see a company through hard times. This was a perfect example.
Without those kiosks, i'd have been just another angry traveller.
I have many thoughts about what makes the kiosks so good, but i'd love
to hear others'.
>Kate's comments on my ATM email got me thinking about gas stations. >The pay-at-the-pump systems most gas stations use suffer from the >same types of problems: > >(1) The machine asks you for your zip code. Then it says "press enter >to continue." But the "enter" button is tiny. Why can't I press the >"yes" button? Why does there need to two buttons which perform the >same function? > >(2) Some machines ask you to swipe the loyalty card first, then the >payment card. Some machines ask you to swipe the payment card first, >then the loyalty card. But: if you get it wrong, it cancels the >transaction, and you have to flag down an attendant to reset the system. > >(3) Machines ask you if you want a receipt *after* you're done >pumping gas. Often, this happens while you're screwing the gas cap >back in, and you miss your window of opportunity and have to go >inside for your receipt. Ask me up front -- after the zip code. That >way, when I'm done, I can grab the receipt and go. (In cold climates, >this minimizes the time one spends with the gloves off.) > >Given the limited problem domain, small set of input devices and >narrow decision tree, why are these types of special-purpose devices >so inconsistent and difficult to interact with? > > >Paul > > > > >