Indexes to the corpus of texts (fieldnotes, tapes, transcripts, materials directly produced by informants, etc.) are essential. They are best developed as the field work proceeds and before the corpus is too large.
Note that what I call "indexing" other call "coding" or "categorizing." In some instances the words can be taken to be mostly synonymous. Each however carry a possibly quite different theoretical load that should be clarified:
Categorizing and coding are not inherent part of ethnography: many ethnographers never categorize or code. But all, in one way or another index what they collect. Indeed anyone working with fieldnotes and other material collected through participant-observation, interviewing, video, etc., must start with an index designed for retrieval.
Several companies sell programs that are supposed to make the indexing process easier. Some students I know have used one of these programs with some success.
It is also possible to use list programs in WordPerfect or Word to achieve some of the functionality of these programs. It might also be possible to use HTML as I am doing as an experiment in these pages.
Each of these have been found useful by some students I know. They are all controversial as it is evident in a recent conversation (September 2003) among some of the top ethnographers of education. In this conversation many fundamental points are made about coding, premature closure and other essential issues.
As mentioned in my own contribution to this discussion, indexing one's corpus--however minimally--must occur. This does remain a task that must, theoretically, be done pre-theoretically.
Two indexes are minimally necessary:
An example from my fieldwork in Ireland.
An exercise from my fieldwork in Ireland: how would you index these pictures?