(Lazarus – Original 09/13/2005)
While I tend to agree with Norman’s assertion that activities should be design’s focus, I disagree with the black-and-whiteness of the debates in the wake of “HCD harmful?”
I believe the real issue is not the inherent superiority of examining either activities or people. Rather I suspect the real issue is how to determine the balance between examining each and under what circumstance one should lead, or take priority over, the other.
Norman’s critique succeeds as a valuable reality check to those with too much faith in personas, and as a reminder that the designers job is to help people better perform activities.
However, his critique also has a couple rather weak points. First, it uncharitably represents personas and scenarios. They are indeed more than just communication tools. To be fair though, Norman’s representation might be a reflection of how he has seen them commonly used rather than their full potential in the hands of an experienced professional designer.
And second, his critique tries to be too universal. What is an appropriate balance between ACD and UCD will obviously vary — a nuance his critique elides. Here’s an example. Compare call centers with hospitals. Employees in both are unionized, follow highly evolved procedures and interrelated activities, are strongly hierarchical, etc. You could very successfully model a call center almost entirely on activities without personas, because call centers have evolved to maximize the interoperability of its people, turning people into little more than hot-swappable cogs, so to speak.
On the other hand, even the most complete activity analysis of a hospital would provide only a superficial model that would overlook or minimize the complex web of cultural and personal motivations, goals, and attitudes that make each hospital so unique and that personas and scenarios are so good at revealing.
As a result Norman raises an important issue, but fails to help us develop the tools to prioritize the balance between activities and people necessary to a systematically successful and repeatable design process. Perhaps that will be in his third article soon.