Use It or Lose It

This was written on November 9, 2000

It was fascinating. I sat in a tea shop the other day and watched, for about 15 minutes, while customers tried to pull open the front door, a door which can only be pushed. The big brass door handle that stuck out into the street, begging to be pulled on, was a ruse.

After a while, I noticed another effect of the design. Customers leaving the tea shop, carrying a piping hot beverage in one hand and the rest of their stuff in the other, had to pull the door open to exit the cafe. They would set their tea down on the counter, open the door and awkwardly make their way through, grabbing the cup before the door closed on their hand.

As I watched this scene, drinking some tea that I’d fortunately purchased “for here”, I thought of Donald Norman’s book, The Design of Everyday Things, in which Dr. Norman discusses the design failures of so many door handles and light switches. After reading his book, when I find myself pulling on a door that only pushes, I think, “This isn’t my fault–it’s poor design!”

If we have significant problems designing usable doors and switches, do we have any chance of designing a usable voting ballot?

“The ballot is very straightforward,” said Palm Beach County Commissioner Carol Roberts about the “butterfly” ballot that came under much scrutiny in the 2000 presidential election. “You follow the arrow, you punch the location. Then you have voted for who you intend to elect.”

In the eyes of the ballot’s designers, the Palm Beach County ballot was as usable as it could be. It had large print, to accomodate people who are blinds as bats. It also had big arrows pointing to the holes for each candidate. The process couldn’t be any simpler, right? Like the woman said, you follow the arrow and punch the hole. Unfortunately, it’s now clear that this design was never tested. The design process failed, and after the election we were left with an alleged 19,000 double-stamped ballots in that county.

I asked Dr. Norman about the ballot’s design. “There are those who say that the fact that people made a mistake is due to their stupidity,” he said. “This is really bad thinking. If one person makes a mistake, it might be that person’s fault. When thousand do, it is bad design.”

The ballot’s original design seems to have had a number of restrictions. For example, the voting booths and ballot cards already existed, so the ballot itself had to be designed around the booth before the voter could be considered. One column of holes was necessary, and those holes had to be a certain distance apart. Another restriction lies in the ballot cards: The ballot cards needed to be as small as possible, to save space and make it easier for the vote counting machines. This meant less space between the buttons and a single column punch card.

That may have been fine if the choices could be presented in one column on the ballot. But the ballot’s large print was deemed an important factor, and it required a two column design. It was a sacrifice of one form of usability for another–making the ballot easier to use for visually impared voters, but harder to use for every voter. Wanting to avoid confusion, they compensated with the big arrows pointing to the buttons, but the arrows simply didn’t work. Maybe some other visual design would be usable and accomodate for all the design restrictions I’ve mentioned, but that’s beside the point: the chosen design failed.

Ahh, The Politics of Usability come into play heavily. Or, should I say, the Usability of Politics? The user interface is often the bane of a technology, even one as old and familiar as the mechanical voting booth, with its circa 1960s hole-punched paper ballot. Lets hope they get it right in 2002.

A couple weeks ago, I was at Karl’s parents’ house, and the topic of the recent Lexington, KY plane crash came up. Karl’s dad is a pilot, and he said, “The Lexington, KY airport is one of a handful of airports where two runways start at the same place, so you’re never sure if you’re on the right one.” He said this isn’t the first time—the Lexington airport has confused pilots in the past.

I’d always wondered why airports with two runways were designed in an X pattern rather than a V. Yes, the X takes up less space, and runways are very long. But the V runway setup is also inherently confusing. Now, in the Lexington crash, there are questions of which runway lights were on and which were off, questions of how much sleep the controller had, and so on. But I think Karl’s dad is right on: the real culprit is poor design.

Here’s Lexington’s airport from above (via Wikipedia):

To get to the short runway, just follow the red line down that short path from the terminal. And to get to the long runway, just follow the blue line down the same short path.

So all planes taxi out in the same direction from the terminal.

Now here’s another example—from an airport down the road in Paducah, KY:

Now, I am not a runway designer, or even a pilot, but from an everyday usability perspective, I like what I see. From above, there are five, and probably more, very obvious distinctions between the short and long runways:

• They are almost at 90 degree angles to each other.
• The main terminal is between the two runways, so pilots taxi out in opposite directions.
• But if you’re asleep or dyslexic, or for some other reason it’s still not obvious, there are two 90 degree turns required to get to the shorter runway from the main terminal, while the long runway asks a fairly straight path. This is the “Are you sure you want to delete this file?” dialog box of runway design. You have to do extra work to get to the more dangerous runway. If someone accidently takes off from the long runway, chances are there won’t be any problems, so that path can be pretty straight.
• Those big white stripes on the long runway are pretty official looking. “You’ve come to the right place,” they say. Showing up at the small runway with a commercial jet, you will probably notice the lack of stripes.
• And probably the most obvious distinction is this: there are two separate terminals! One for small planes, at top, and one for bigger planes (with the bigger parking lot), at right. And you can’t even get from the small plane terminal to the long runway without first going past the main terminal. Aside from being safer, this reduces congestion that might happen on the road to the airport, in the terminal, on the tarmac near the gates, and on the taxiway itself.

In other words, this design is really two airports in one, sharing space. But there might be one problem here that might not exist in Lexington: during busy times, a plane at the end of the short runway can’t see another plane at the end of the long runway. There are buildings in the way. I wonder what are the chances of two planes thinking they’re cleared for take-off at the same time? And if they do proceed, what are the chances of a collision in the Paducah design? I think this is an unlikely scenario, as there is a sort of “two-phase commit” that happens between the ATC and the aircraft, for any take-off or landing, that does a pretty damn good job of serializing the events.

Anyway, I can’t think of a place where wayfinding is more important than on the runway. The consequences of confusion could not be higher. So the best way to follow up this recent crash in Lexington is with a reassessment of all airports to find and remove ambiguities in runway design, and to further segregate the small planes from the jets. Go check out your local airport on Google Maps and see where the taxiways are. If there is any doubt in your mind, make your complaint to the FAA. Of course, it’s probably too expensive to redesign the V runways, but the taxi path and the distinguishing marks for each runway, from the pilots perspective, could be reviewed, improved, and tested with both new and frequent users of the airport.