Student attitudes play an important role in shaping learning experiences. However, few validated instruments exist for measuring student attitude development in a discipline-specific way. In this paper, we present the design, development, and validation of the computing attitudes survey (CAS). The CAS is an extension of the Colorado Learning Attitudes about Science Survey and measures novice to expert attitude shifts about the nature of knowledge and problem solving in computer science. Factor analysis with a large, multi-institutional data-set identified and confirmed five subscales on the CAS related to different facets of attitudes measured on the survey. We then used the CAS in a pre--post format to demonstrate its usefulness in studying attitude shifts during CS1 courses and its responsiveness to varying instructional conditions. The most recent version of the CAS is provided in its entirety along with a discussion of the conditions under which its validity has been demonstrated.