Neil's News

OU-M873 (UI)

22 January 2005

Recently I took a six-month course with the Open University titled "User Interface Design and Evaluation". Programmers are notorious for creating applications which bewilder the rest of the world. I took this course hoping to learn how to improve my user interfaces.

[TI-83 calculator]Here's an example of what this course advocates. To the right is a TI-83 calculator. The text book examines the calculator's interface, and the following recommendations were made (Unit 4.67):

  • Choice of input device. "If a large touchscreen with a pen had been used, it would have been possible to use only screen buttons, which might have made the device more flexible."
    Implementing this recommendation results in a PalmPilot; a device that costs three times as much, chews up so much power that it has to live in a charging cradle, requires the use of a stylus since fingers aren't accurate enough and breaks far more easily due to its large screen. This recommendation results in an inferior calculator.
  • Choice of output device. "A backlit screen might be an improvement."
    No thought is given to the natural consequence of backlighting: that the battery would have to be replaced every four hours. This recommendation results in an inferior calculator.
  • Use of text. "A sans serif typeface has been used for the buttons and a serif typeface has been used on the screen. The latter is a surprising choice, and to my eyes the screen is slightly less readable than the buttons."
    Like all calculators it uses a monospaced font on the screen since you want columns to line up. There is only one character on the screen which is in a serif typeface: "1" Changing the "1" character to be sans serif would introduce a gap in the number, and would make it appear that it was two separate numbers. It would also be indistinguishable from an uppercase "I" character. Using a sans serif typeface would introduce ambiguities in the output. This recommendation results in an inferior calculator.
  • Simplicity of design. (part I) "A simpler approach might have been to put the buttons intended only for advanced users under a movable cover."
    The first thing they teach in this course is to identify the users. Texas Instruments builds dozens of types of calculators ranging from basic four function devices for primary school to graphing devices for university. If you bought a TI-83 then you are an "advanced user". Hiding the advanced functions wouldn't benefit these users. This recommendation results in an inferior calculator.
  • Simplicity of design. (part II) "Alternatively an on-screen hierarchy could have been included, with the more advanced commands embedded in the hierarchy."
    When you want an "advanced" command on a calculator, typically you don't want it once, you want it dozens of times. With the keyboard each command takes either one or two button presses. With an onscreen hierarchy it would take five cursor key presses to reach an option (assuming a trinary menu). This recommendation results in an inferior calculator.
  • Simplicity of design. (part III) "The mapping for switching the calculator off is not as clear and direct as it might be, because it requires two key presses (the 2nd button followed by the On button). This is surprising, as it is such a frequent operation. A simpler solution would have been to make the On switch toggle between on and off."
    I simply cannot believe that the "usability experts" are suggesting that a button which causes total data-loss be placed 5mm from the "0" key. This recommendation results in an inferior calculator.

Unfortunately most of the course was like this. Applying their rules is a sure way to cripple any interface. The only valuable lesson from this course is to watch the users and see where they encounter problems. But don't use the course's approach of formal user observation sessions since it distorts the results (Heisenberg-style). If you want accurate information, do the observation discretely, covertly or remotely.

I went into this course eager to learn and apply structured principles of UI design to my projects. I came out of this course determined that nobody calling themselves a "UI designer" should ever come near one of my projects.

< Previous | Next >

Legal yada yada: My views do not necessarily represent those of my employer or my goldfish.