Comments on Readings

Ethnography

The paper echoes many of the concerns in the Artifact [Bødker] paper (and the Scandinavian school in general):

• User involvement in all stages of design.
• Observing users in their natural habitat.
• Understanding existing practices and how they fit together.

The big lesson that designers need to understand from ethnography is that in order to do something for someone, you must understand them. Understanding comes from doing things with the community and from observation.

As a Zen-and-motorcycle-maintenance fan, I find little to fault in this paper. True understanding comes from doing, not from contemplation.

The paper is very good at describing ethnographic praxis. The vital concerns of ethnography are covered succinctly. Designers get practical hints on successful observation and analysis. The final exercises are terrific encouragements to go out and do something.

Problems with the paper:

• The important difference between observing and acting as an agent of change is not emphasized strongly enough. i.e: a designer has a double obligation to the user community, as observer and as outside agent with power to influence their environment.
• Similarly, the importance of gaining the cooperation of a user community is not raised. The Xerox PD project was working with 'enthusiastic' groups.
• 'How to convince skeptical co-workers and bosses' would have been a very useful section, given the problems that fuzzy methods have had in engineering. Nor does it discuss the high costs associated with field studies.
• Some important ideas get short shrift ('shared understanding' for example), but that is to be expected in this sort of paper.

Twinkling Loops

One key point raised here is that users are usually not interested in dealing with programming. The less hassle a spreadsheet is, the better. This has obvious consequences for the design of the user-interface. The basic and advanced levels of functions can be separated for ease of learning.

The authors point out that lab studies will give you results unrelated to the users normal praxis. In other words, usability tests are of limited use, as stated in the Ethnographic Methods paper.

I wonder how other CSCW applications would similarly be illuminated by ethnographic analysis. Do people really want shared editors?

Problems:

• The lack of non-interview observations. If the heart of ethnographic research is observing the user in its natural habitat, this study seems to fall short.
• The structure of the interviews is not covered. It makes a big difference if the cooperating team members were interviewed together or singly.
• It is very big on qualitative data - there is little quantitative data. Ethnography is not a quantitative science, but a bit more detail in the results section would be nice.