He fingered and fingered the computer it simply amazed Melrose that the machine supposed to take the pain out of all sorts of niggling jobs took more time to perform a simple one than it would have taken Bub to do by hand ten times over.
?span>Martha Grimes, The Stargazey (a detective novel)
Many qualitative methods and heuristics are useful for analyzing and understanding interface design. These methods form the majority of the content of most books on the subject, including those cited in the references for Shneiderman, Norman, and Mayhew. For example, what an experienced interface designer can learn from passively observing a test of a new interface with a few subjects can be as valuable as what she can learn from any quantitative analysis. My concentration on quantitative methods is not meant to denigrate the importance of qualitative techniques but rather to help even the balance by emphasizing the numerical and empirically testable methods that are not yet widely used. Quantitative methods can often reduce argument to calculation; a further, and most important, benefit is that understanding why the quantitative methods work guides us to understanding important aspects of how humans interact with machines.
One of the best quantitative analyses of interface design is the classic model of goals, operators, methods, and selection rules (GOMS), which first gained attention in the 1980s (Card, Moran, and Newell 1983). GOMS modeling allows you to predict how long an experienced worker will take to perform a particular operation when using a given interface design. After discussing the GOMS model, I present quantitative methods for determining interface efficiency, cursor movement speed, and the time cost of decision making.