Monthly Meetings

February 23, 2005: Three Keys To Successfully Refactoring User Interfaces

Downloads

• Martha's presentation on Refactoring User Interfaces
• Architectural Design 2.0 of OOoBib User Interactions (PDF file)

Abstract

There are three keys to successfully refactoring how users interact with a system:

1. Defining types of user costs, such as unlearning, that result from user-interface changes
2. Creating a frame of reference that decreases user costs of changes without adding development time
3. Resolving tradeoffs for any proposed user-interface change

This talk integrates these three keys into a coherent approach to user-interaction design that (1) does not add significant time to development, and (2) is compatible with the iterative approach of eXtreme Programming. The first key to successfully refactoring of user interfaces is to understand the user costs of change. Computers have no memory of previous versions of refactored code, but users have conscious and unconscious memories of previous versions. The unconscious memories are particularly difficult to change because they have been encoded in muscle memory that governs physical actions before conscious thought occurs. Thus user relearning typically involves making errors, which is why the relearning process is a major source of user complaints about user-interface change.

There is a user-interface design process, called GAINS, that allows iterative changes (refactoring) without adding significant time to the development cycle. Thus the second key to successful refactoring of user interfaces is to quickly create an architectural framework for iterative design. This can be done only when a clear distinction is made between user-interaction and user-interface design.

The third key to successfully refactoring user interfaces is to define and resolve tradeoffs for any particular user-interface change. For example, a conceptual change may be easier for users to learn than changing a well-learned (muscle memory) action. Thus changing the name of a button when moving it elsewhere in the user interface might generate less user costs than just moving the button without changing its name -- particularly if the new name does not trigger the unconscious physical action associated with the old location of the button.

Martha J. Lindeman Ph.D. (President, Agile Interactions, Inc.) has been a consultant/trainer in rapid user-interface design since receiving her Ph.D. in cognitive psychology from Harvard University in 1985. She created and repeatedly refined her GAINS design process in projects ranging from small 'on-the-fly' projects to very large projects for companies such as Xerox, MTel/SkyTel, Prudential, Ford, and Bank of America. Her focus has always been on twin goals:(1) creating easy-to-learn, easy-to-use user interfaces, and (2) NOT adding time to the development process. In addition, user-interaction design often has to be done using project documentation rather than information drawn directly from expected end-users. The GAINS process has been successfully applied to the design of interactive systems in many different types of domains and to systems ranging from very simple to very complex.