Tons and tons of good stuff in the article, but this section in particular caught my eye:
The difference between a user interface designer and usability specialist
The user interface designer takes information gathered about the nature of the tasks and goals, combines it with a knowledge of human factors, culture, psychology, graphic design, and many other areas, and produces an interface which is designed to allow the end user to achieve their goals in a clear, appropriate way. The usability specialist is the one who tests that design.
The role of the usability specialist is not a creative one in the same way that a user interface designer is creative. The usability specialist simply reacts to the design based on established guidelines borne of experiments and experience (heuristics), or has end users react to the design by having them use the design in a typical way and observing the results. In both cases, what the usability specialist does is provide a verification role.
IA and usability are starting to be seen as the same thing. A friend writes:
“Souls who describe themselves as “information architects” almost always mean ‘user-centered information architects’, so much so that I’ve stopped bothering to clarify it. ”
He defines himself as a “info-centered information architect, where the goal is not to make users happy but to drastically reduce development time and effort by making information structures fit a wide range of apps (including html display)”
I’ve always said that the usability specialist (or usability engineer or whatever.. would you people please decide on a title) studies and tests, while the IA designs structures. IA is a calculatedly creative act, while usability is research– during the beginning, middle and end of a project. IA’s should not test their own work any more than a writer should edit their own.. it’s just too hard to stay objective.
And usability folks need to not design if they are testing. It muddles things. I’ve seen more usability reports that look like text-redesigns. It’s hard when you are evaluating (I know!) not to recommend a few simple changes that might fix the site, but it must be done with great caution. If a usability person is there only during testing, they may not fully understand the constraints made by the requirements, and make specific recommendations that are impossible for technical or political reasons. “Move banner to top so users understand it is an ad” could contradict the VP of marketing’s desire to keep ads out of the branding space. “Most user’s were not certain what was an ad and what was not, and may leave site before completing transaction. Recommendation: make it clear clicking ad will take user to another site and abort the sale” By pointing out the consequences of not making the design change, but leaving the change in the hands of the client is a more effective way to get the message across.
Usability folks also may step on delicate egos with their recommendations and –I’m very sad to admit– have perfectly good insights discarded out of pique. A good recommendation is phrased with caution, using many qualifiers and kept general: “Most users had difficultly finding the search box, possibly because they were looking for a text entry field with the word search next to it. You may wish to explore ways to make the search box more visible, possibly by considering adopting standard web conventions such as placing it in the top left and/or using a standard format.” I’ve seen recommendations such as “make links blue so user understand to click them”, or “put box around text so users understand location of content.” This prompts some entertaining eye-rolling — Oh, those Jakob-types just don’t get design– but doesn’t help get the interface working better.
But I digress… I keep coming back to my friend’s quote. Should we really be splitting forces into info-centered and user-centered? Isn’t an IA who is trying to “drastically reduce development time and effort by making information structures fit a wide range of apps ” thinking of another set of users– those in the company? Isn’t it possible one could be user-centered in that one considers both internal and external users? Or is that too much knowledge for one IA’s head?