I'm thinking out loud a bit myself this morning. Pulled a late night work on chapters about the QOC method and the use of cause and effect diagrams in UCD. A key issue here is that we often speak of "usability" as a single thing when it has many different attributes.
This is tricky since usability is a general term and usability has many different attributes or dimensions, with predictability being one of those. Usability or more broadly user experience attributes can interact with one another. For example, in studies of what makes a system usable, response time often bubbles to the top of the list (actually, it is often in the top 3). You could have a system that everyone can learn easily (initial learning) but poor performance (average response time or variability in response time) might make a system that is usable in one dimension, unusable overall. The ISO definition has effectiveness, efficiency, and satisfaction, the basics, but there are many more dimensions to usability that have been considered since the 1980s based on work by Whiteside, Bennett, Gould, and other early proponents of usability engineering, some that overlap broadly. So in the statement below, you could replace "usable" with any one of a number of dimensions and compare it to predicability and examine relationships among different usability attributes. As Itamar noted, unpredicability can be good in some circumstances and human factors engineers have done a lot research into factors that affect vigilance in sustained-attention tasks. Wylie and Mackie in the late 1980s suggested that a system be injected with artificial signals and then the operators would get feedback on their detection performance. One of the issues that hinders our fields is that we fail to think of usability in our day-to-day work as have many dimensions with some dimensions being more critical in a particular context than others. The simplest example is a system where some people use a feature once a month while others use it 50 times a day. Many people speak of consistency in a user interface as promoting usability but if you have a bi-modal distribution, you may need two interaction methods to be consistent with the way people work. The principles of being inconsistent to be consistent is a difficult one for many corporations who decree that "all our products will be consistent visually and have only one way to perform a task (which gets into another topic of redundancy which is another good debate). Grudin published what I think is a classic paper that got into the subtleties of consistency (often considered an attribute of a usable system) and noted that to be consistent with how people work, you might need to design systems that have inconsistencies (as well as multiple ways to do things). The Grudin paper uses old examples, but the issue is still as strong as ever. For a good read see The Case Against User Interface Consistency by Jonathan Grudin at http://research.microsoft.com/users/jgrudin/publications/consistency/CACM1989.pdf This is a good topic since I think that an explicit definition of "usability" is critical to each project that we work one; listing the attributes of usability that are most critical to success of a product (learnability, efficiency, prevention of errors, flexibility, predictability, fun, memorability (critical for systems used infrequently like tax software) satisfaction, performance, ....). I often hear people referring to "usability" and they are each defining it differently which can create problems throughout the product design and development cycle. Chauncey On Sun, May 25, 2008 at 1:25 AM, Robert Hoekman Jr <[EMAIL PROTECTED]> wrote: >> >> Not everything that's usable is predictable and not everything that's >> predictable is usable. > > > Examples? > > -r- > ________________________________________________________________ > Welcome to the Interaction Design Association (IxDA)! > To post to this list ....... [EMAIL PROTECTED] > Unsubscribe ................ http://www.ixda.org/unsubscribe > List Guidelines ............ http://www.ixda.org/guidelines > List Help .................. http://www.ixda.org/help > ________________________________________________________________ Welcome to the Interaction Design Association (IxDA)! To post to this list ....... [EMAIL PROTECTED] Unsubscribe ................ http://www.ixda.org/unsubscribe List Guidelines ............ http://www.ixda.org/guidelines List Help .................. http://www.ixda.org/help
