Monday, January 7, 2013

Usability problem of the day (gussied-up machines that aren't infinitely fast)

In the courses I teach about human-computer interaction, I typically open each class with an example of a usability problem. I'm putting these online, in case others find them useful.

In 1987 Alan Dix published a paper titled "The Myth of the Infinitely Fast Machine," in which he describes the problems that can arise when software developers assume that computers respond instantanously to user actions. Dix, originally a mathematician, has devoted much of his career to formalizing what we know about HCI, and Human-Computer Interaction (Prentice Hall, 2004), by Dix, Finlay, Abowd, and Beale, is one of the best graduate-level textbooks in the field.

Aren't computers fast enough for us to think of them as executing instantaneous transitions between states, at least in the abstract? Not always. Dix identifies cursor tracking as a worst-case problem--the user moves the mouse and the cursor on the display can't keep up, leading to a move-and-wait, move-again-and-wait strategy for doing the simple task of target selection in a graphical user interface. This is not as common today as it might have been in 1987, but new software technology brings new problems...

One of the enterprise systems I use dynamically generates the pages it sends to my Web browser. (Approximately speaking, that is; it relies on Ajax.) When I visit the site, it typically looks like this as it loads:

Unfortunately, it loads very slowly. Further, different regions of the page load at different times. I've superimposed a red mouse curor on the screen shot of the Web page to illustrate a usability problem that I regularly encounter. Suppose I'd like to "Manage My Projects." I position the cursor over "Proposals and Project Management" and I click the mouse button. Oops! Just before I click, the rest of the page loads, and my target slides out from under the cursor before I can stop my action.

The cursor is positioned over something completely different. It's not unusual for me to click on an unrelated target and then have to go back (and wait) to find what I really wanted. As Dix puts it, "what you see is what you get" has turned into "what you see is what you had." I can't simply see what I want and click on it; instead I have to look around at the rest of the display to make sure that the system won't change in the next few seconds and only then take action.

It's a little bit discouraging that usability problems we knew about 25 years ago are recurring today.

And what is being loaded, by the way? Largely useless icons. I appreciate good graphic design, but as with all good things, it's painful to see it done badly. Here are four examples of icons on this page. Can you guess what they represent?

That's okay--I'd have no idea either, without the labels (and sometimes not even then). The first icon shows an open book with green pages next to a set of collegiate buildings: it's about online material for academic courses. In the second icon, a purplish homunculus addresses those same buildings with a word bubble showing a paper document: the function is "employment/income verification". The third icon shows a different building, with a green-and-yellow bank note beside it and an airplane (not a bat) overhead: travel authorizations, obviously. The fourth icon shows a small homunculus sitting behind a desk, talking with a standing homunculus carrying what may be a pumpkin or a large sack: this is faculty advising.The overall result is perhaps even more dispiriting than seeing a usability problem that's time-traveled from 1987 to 2013. We're not only failing to learn from the past--sometimes we're actively making things worse.

No comments:

Post a Comment