Embedded UIs, usability testing, and the user from hell

September 28, 2010

Bernard Cole-September 28, 2010

While cruising the show floor at last week’s ESC, I became much more aware of how important an effective, clear and unambiguous user interface (UI) is in embedded designs, despite the many articles I had read and edited for use on EET/Embedded.com and  in ESD Magazine.

Some of the more outstanding ones have included "Interfacing the User," "Usability for Graphical User Interfaces," and "GUI testing: exposing visual bugs," and "Design user interfaces for cooperating devices."

But despite the many articles we have run on GUI design and hardware design in general, I did not  realize just how seldom real users (including those "users from hell" who don't read documentation and believe a  few minutes of use is all that is necessary) are involved in evaluating such designs before they get to market. 

As I pointed out in “It’s usability testing, stupid,” I do not think enough effort is made to find such users in the environments where they normally use the devices and monitor how they really use them.  

But it was at the ESC show in Boston last week that I thought of a possible answer to this lack of meaningful user from hell input into such designs. There, a number of companies showed me tools they had designed to aid in the development of many of these new UI enabled embedded applications.

One that caught my eye, and caused an immediate visceral response to it, was the Mentor Graphics Ready Start Platform and its Inflection Interface runtime engine for 2D and 3D graphics and wired and wireless connectivity middleware. It incorporated ready to use applications and configurations designed to shorten significantly the time to develop medical, industrial, automotive and consumer devices.

The example they showed me was of the setup for designing a glucose monitoring device. I am a diabetic whose life is sometimes made miserable by badly configured and provisioned medical devices, and I itched to get my hands on a real design on the platform and make half dozen significant changes and many minor ones.

I do not expect  the tools such as I saw at the show, designed primarily for developers,  to be used in this way. But the companies developing such tools should  think about coming up with variations specifically designed to aid them or their customers in getting  real, meaningful input from the actual “in the wild” users of the many such GUI-enabled embedded devices we can expect in the future.

Some design articles and columns we have published addressed some of these issues, including   "Usability testing for design kits: A good idea,"  "The usability dilemma," and "Put the user in the driver's seat."

But it is an area that deserves further exploration and I would like to see more from you on this issue and  others relating to embedded user interface design and testing. It is a topic that will become increasingly important as embedded devices become more wired and wirelessly connected and permeate our lives even more. (EET/Embeddded.com Editor Bernard Cole, bccole@acm.org, 928-525-9087)

Loading comments...