These principle guide the development of the pattern language.
Ground patterns in practice. These patterns are drawn from examples of papers published in respected venues. We do not need to duplicate the extensive literature on research methods. That is, we try to capture the research patterns that are accepted in practice, not the abstract guidance about how research ought to be done.
Avoid duplication. This should serve as a sourcebook, not an attempt to be complete. First, there’s no point in duplicating work that’s been done. Second, we can’t possibly have enough energy. This implies developing a good annotated bibliography.
Remain tractable. Pick an initial scope that allows for critical mass — broad enough to be useful in some area — but narrow enough that it can be reasonably well fleshed out. For example, focus initially on conference papers, which are useful building blocks in developing a body of research, but tend to be much simpler in structure.
Become useful. In the same vein, flesh out the initial scope so that the important alternatives are included. As a bootstrapping strategy, within this scope favor plenty of pattern sketches with pointers to details over a small number of completely-fleshed out patterns
Provide guidance about choices. The emphasis should be on clarifying differences among research methods and on identifying factors that should guide the choice of a research method (and certainly on choosing a pattern that matches the research method actually used).