Corpora are often referred to as the ‘tools’ of corpus linguistics.
However, it is important to recognize that corpora are simply linguistic data and
that specialized software tools are required to view and analyze them. The functionality offered by software tools largely dictates what corpus linguistics research methods are available to a researcher, and hence, the design of tools will become an increasingly important factor as corpora become larger and the statistical analysis of linguistic data becomes increasingly complex. In this paper, I will first discuss how separating the data from the tools resolves various issues that are hotly debated within the field. Next, I will offer a critical look at the development of four generations of corpus tools, discussing their strengths and weaknesses. Then, I will discuss the role of programming in corpus linguistics tools creation and present a model for the development of future corpus tools. Finally, I will show a real-world example of a next-generation corpus tool that was developed for use in language learning.