The tools required to build an assessment depend on its purpose. One can build a one-question test as a provocative start to a class without a lot of infrastructure. The teacher can develop the question based on the topic of the lecture, deliver the question orally, request the students to chose one of two responses by raising their hands, count the responses to each alternative, and remember the results. A high-stakes assessment, in contrast, requires a set of test-task specifications, a procedure for selecting particular tasks from a pool of tasks developed to these specifications, a mechanism for individual delivery of test questions, a means of individual entry of responses, a method for collating and analyzing the responses, and a means of storing results for subsequent analysis. The types of tools selected, of course, depend on the purpose of the test and a number of practical issues such as their cost, availability, and adequacy for doing the job. When the tools involve the computer for every stage of the process, the issues are still more complex. In this chapter, we describe the types of software tools that make CALT work.

It would be most satisfying to be able to describe a complete, working, sophisticated system for the development and use of CALT, but as of 2005 what we have instead are more modest systems, plans for ideal systems, and works in progress. We therefore discuss today's reality of CALT tools in this sequence.