What is the difference between discourse analysis and ethnography?

The conversations both in class and on the blog have made me realize that I'm not quite clear about the distinction between discourse analysis and ethnography. I feel I could identify each by example--Epstein's book is discourse analysis, while Orr's book is an ethnography--but it's not exactly clear to me which makes which which. (Two asides: Firstly, I think Wittgenstein's response to my observation here would be, "Duh." Secondly, to jump ahead a bit, this problem of example versus universal defining feature is discussed in an interesting way in pp 67-69 of the Flyvbjerg book. I'm Plato in that example, and that's not a good thing according to Flyvbjerg. But anyway, leaving aside the question of whether I'm asking the right question here....)

On the topic of the boundaries of discourse that Caroline raised and that we discussed briefly in class, I find myself wondering not only where a given discourse (for example, discourse about whaling) begins and ends, but rather how the parameters of what constitutes discourse are established. For example, Epstein describes discourse analysis as a focus "not primarily on words as such but rather on the production of meaning. Thus , it is concerned with any type of signifying practice, that is, any practice that functions as a site for the production of meaning" (186). So we include not only talk, but consumption habits, etc. She goes on to note that this meaning emerges "in the exchange between social actors" and thus the context for such an exchange must be taken into consideration.

But wasn't this what we were doing in ethnography? Trying to observe, understand, and analyze practices that function as sites for the production of meaning? Weren't we documenting exchanges between social actors in distinct contexts to try to grasp at local meanings? Is the difference in discourse analysis that these meanings are broader reaching? That they are public? Is this what distinguishes Orr's analysis of dialogue between technicians from Eptein's analysis of rhetoric among anti-whaling activists? Is it the historical dimension of discourse analysis that makes it distinct? That we can trace the evolution of a discourse over time, whereas an ethnography captures something that is happening in the moment? But then, in our ethnography, would it not be appropriate to include analysis of historical documents? For example, Orr looks at the way that the technician's manual has evolved over time and reads from that a larger story about the systematization or institutionalization of work.

Perhaps this will become clearer as we undertake our own discourse analysis in the in-class workshop. But if anyone has any thoughts on this, I'd love some help in trying to understand!