opening it up with Common Lisp
Book review: Darwinia
Summer reading: Spin
the Omnivoire's Delimma
the Golem's Eye
Information Ecology: Open System Environment for Data, Memories and Knowing
Bowker and Baker examine in the interconnections between memory, data, information and knowledge in the context of a study within the Long Term Ecological Research (LTER) community. In our time, databases have become fundamental but the "data [within them] never stands alone." Indeed, the memory of an organization exists in multiple interacting forms and includes both data and procedures. As organizations grow, the data and the interacting web of procedures grow with them. This becomes especially problematic when data must be shared across space and time; standards must be created and agreed upon, units must be unified, formats must be formed and all must be maintained. The obvious answer is to make heavy use of metadata. Here, however, the problem recurses -- how are to set standards for the metadata? Indeed, "the proliferation of metadata standards within environmental science [is] as significant as the proliferation of data standards themselves." "This suggests the need to accept that there are very real social, organizational and cognitive machineries of difference which continually fracture standards into local versions." The solution for this requires not more standards but a "careful analysis of the political and organizational economy of memory practices..."
The authors posit two dimensions: data and knowledge. Data can be local or global; knowledge can be tacit or explicit. This creates four quadrants:
Information flows from quadrant to quadrant with feedback across boundaries and with change within each quadrant. The whole forms an ecology of information. If the dynamic flux of this ecology is ignored, we "risk putting in place systems that create barriers to inquiry." The examples the authors present are suggestive and show that more than technology is required for database management.
Another insight of this paper is that the process of infrastructure building is far more complex than it first appears. It is hard to define what Information Managers do.
Information management is a vital part of any system but we "don't have good ways of talking about [it]..." This process oriented work is "frequently invisible and rarely supported." In attempting to bring this work to the forefront, Bowker and Baker are performing a valuable service. As they put it, we live in a tension between homogenization and diversification. "The question then is not only 'with what epistemological and ontological frameworks shall we work?' but also 'how can we work at the intersection between different frameworks?" The goal of long term studies and real world data mining is not 'how can we capture the data?' but rather 'how can we build an open information ecology in which the changing data can live and prosper today and tomorrow?'
This is challenging paper from well outside the standard views of Computer Science. Perhaps that is why it feels so refreshing and correct to me. It is easy for technologists to view problems in simple terms but real problems and answers must live in contradiction with one another in a world open to change. I believe that the broader analysis of systems in political, organizational and essentially human terms will help us do better science and produce better answers for a complex world.
Copyright -- Gary Warren King, 2004 - 2006