Twenty-four-hour dietary recalls are used to collect high-quality dietary data. Because they require highly trained interviewers, recalls are expensive and impractical for large-scale nutrition research, leading to the use of food frequency questionnaires. We are developing a computer-based, self-administered 24-hour dietary recall for use by adults. Our goal is an easy-to-use, low-cost, publicly available, Web-enabled instrument that will include elements of the Automated Multiple Pass Method developed by the US Department of Agriculture, which uses five passes to enhance recall. The initial pass is called the "quick list" and allows respondents to report foods consumed the previous day freely, in any order, and without detail. Using a crossover design, we conducted initial formative pilot testing among 18 adults in a self-administered computer environment. We tested two versions of a "quick list" (the first Automated Multiple Pass Method pass) for remembering foods consumed the previous day: "unstructured" and "meal-based." Respondents showed a strong preference for the meal-based version (13 of 18), although positive features of both were identified. Chronological reporting was most common, although many foods were sporadically reported out of order. Versions did not appreciably differ in number of foods reported, moved, or deleted. Usability issues and preferences were also identified. If these developmental efforts prove successful, the use of affordable automated recalls could be valuable in clarifying diet-disease associations in observational epidemiologic studies and measuring dietary compliance in clinical trials. This pilot work illustrates the usefulness of formative cognitive and usability testing for questionnaire and software development.