You are here: Articles --> 2000 -->
Garbage in, garbage out: using affordances
Vous êtes ici : Essais --> 2000 --> Garbage in, garbage out: using affordances
by Geoff Hart
Originally published as: Hart, G.J. 1999. Garbage in; garbage out: using affordances. Computerworld Canada, Aug. 27:15. Republished as: Hart, G.J. 2000. Garbage in, garbage out: using affordances. http://www.techwr-l.com/techwhirl/magazine/usersadvocate/usersadvocate_affordances.html
In my previous column, I invoked the "garbage in, garbage out" law to explain the need for validity checking in data-entry forms. Unfortunately, as we develop and maintain increasingly complex databases, the complexity of the validity checks required can become daunting. Worse yet, certain types of data such as graphics haven't been studied long enough for anybody to have developed broadly applicable algorithms for building those validity checks.
In such cases, the trick is to make data-entry forms clear enough that workers understand what you require of them without having to ask. This understanding alone can drastically reduce the frequency of errors, but to turn that understanding into a payback, you'll have to design a label for each field that is truly obvious to the workers. Information designers call these clues "affordances", and if you're lucky enough to have technical writers or editors in your organization, you can probably enlist their aid in designing these clues. Most of these people are eager to be consulted and to expand their repertoires by taking on new challenges.
One key thing to remember is that most workers think in words, not numbers, and thus, you should use words to prompt them. Consider, for example, the difference between the following fields:
The first label will generate inputs ranging from 01-21-99 to 1999/01/21, with a wide range of abbreviated text thrown in for good measure, whereas the second will confuse some users, particularly if you're developing one form for use across Canada (e.g., French workers would use mm-jj-aa), and will generate Y2K errors from the two-digit year. Only the final label will get you the data you really want to receive. Since there's plenty of room to add the extra instructions directly on the screen, there's little excuse not to do so—after all, you can always use a second or third screen if you run out of space. (Just don't make people scroll; unseen fields that lurk offscreen might as well not exist for some viewers.) Whether you actually store the input data as a number or a string is irrelevant; it's the manner in which the workers input the data that's important. Guided by appropriate labels, they can focus on entering the right information rather than guessing how you want them to format the information.
The best way to develop affordances is to spend some time with your workers to learn how they would prefer for you to present the clues. When you do this, it pays to park your ego at the door: they don't speak the same language you do, and aren't interested in learning your language, so you'll have to learn theirs. The payback is twofold: they'll soon understand that you're on their side in the daily war against the computer, and you'll get cleaner data and fewer calls to technical support.
It takes a fair bit of time to develop useful affordances, but it's a worthwhile investment. In addition to reducing data errors that you'll have to clean up later, building relationships with the users of your forms will get you through the inevitable rough times, even as you reduce your workload and make those rough times a little bit easier for everyone else to get through.
[Author's note: The example of date labels is obviously simplistic, since you can certainly bury your readers in extraneous information. But that's the subject of a future column.]
©2004–2017 Geoffrey Hart. All rights reserved