Friday, February 20, 2004

Posting at Brint.com

I responded to a post over at Brint.com. For those of you who don't frequent Brint, here's a peak:

Posted by Jim Nash on February 17, 2004 at 02:15:44:

We are in the process of designing knowledge management systems that have self-adaptive capacities for adapting to the changing business environment. We are facing a tension of extremes between our technical KMS architects and behavioral specialists. While the technical experts are pushing the artificial intelligence and expert systems agenda, the behavioral specialists are pitching Senge's organizational learning and soft systems thinking. Has anyone faced a similar situation in any other organization. How did you resolve this deadlock. All suggestions welcome. Thank you in advance. - Jim


---- Response ----

Jim:

There is no resolution. You’re going to need both the techies and the behavioral specialists. It’s important, however, to realize why you need them. Let’s start with the techies.

Expert systems and case based reasoning tools are great for what they do: provide answers to known questions. These tools were developed for rote problem solving environments such as call centers and service departments. They are effective due to the nature of the information they contain and the types of answers they provide. The questions are all related to a canonical literature without discursive interpretations. It’s difficult, for example, to argue with information like the service recommendations for a 1987 Nissan Stanza. Although the users often have to seek hard-to-find information they are generally searching for known information. Expert systems and CBR tools merely afford a better way of accessing the information.

While service writers are rarely called on to create novel new information or synthesize disparate sources of information, a number of other professions are. Engineers, for example, have a wide variety of tools for creating physical artifacts such as machines, buildings, and micro-chips. While information management systems may make it easier for engineers to find documents that have already been created, the core of their work exists in the act of actually creating those documents. The same logic applies to most professions. While expert systems may not radically improve this creative process, organizational learning principles may.

The real trick is to create systems that enable these two epistemic communities—known item recallers and professional creators—to work together. They work in different ways; they have radically different vocabularies; And they use different internal power and discourse structures, etc. Indeed, it is almost impossible for a system to bridge this gap due to the polysemity and heterogeneity of their documentary practices. Without common vocabulary and usage any information retrieval system (such as CBR) is shot regardless of how well it indexes or cross references information. It’s all apples and oranges.

I realize this advice doesn’t really move you toward your goal of resolving the deadlock between techies and behaviorists. My only recommendation is to look at the most fundamental aspects of knowledge organization and ensure that they’re baked into your initiative. I suggest looking beyond the recent decade of KM-mania to the deep dark historiography of knowledge practices, particularly to the earliest common reference tools that were developed: directories, encyclopedias, thesauri, and dictionaries.

Regardless of the whether you adopt a behavioral or technical solution, be sure you can span your epistemic communities and attain the goals of the most primitive information systems. Can your solution allow people to find the people they need (like directories)? Get condensed repackaged information about aspects of their world of which they are ignorant (like encyclopedias)? Decipher the polysemous and heterogeneous language of their peers (like dictionaries)? Understand the dendritic relationships between words and concepts (like thesauri)? If you’re really keen, extend your thinking to those other dusty ancestors of KM like card catalogues and archives! Just stick with the basics.

I hope this helps. Feel free to drop me an email if you have questions or ideas.

Cheers,
George

Wednesday, February 18, 2004

Interesting Quotation re. Systems and Corporations

For some reason this passage has me thinking. I'm not yet ready to provide commentary but I figured that I would post it anyways.

“The practices found in drifting also tell us something about knowledge in organizations. Drifting stems from those mundane, invisible practices that, compared to the crisp world of procedure and method, in a way represent the dark, nocturnal side of organizational work. They are intelligent practices: the expression of practical intelligence… Far from the by now conventional distinction between tacit and explicit knowledge, practical knowledge is the metìs of the Greek—the intelligence of the octopus: flexible, polymorphic, ambiguous, oblique, twisted, circular. In one word; it is the opposite of the straight, direct, rigid, and univocal knowledge embedded in method. To orient oneself in the complex and changing world, when dealing with forces that are too strong to be fully controlled, one needs to leverage the situation at hand, by détournement, false moves, wavering behaviour, never facing such forces up front, but accomplishing with a sudden move (improvisation) the project at hand.” Pg. 94

References

Ciborra, C. (2002). The labyrinths of Information : challenging the wisdom of systems. Oxford ; New York, NY: Oxford University Press.

Sunday, February 15, 2004

The Social Network of Science Studies/ Sociology of Science

Inspired by the work of Valdis Krebs's work on the political book buying habits of the American populace, I decided to create my own map. I have a general interest in the the sociology of science so I decided to take a random walk through Amazon.com to explore the books that readers "have also bought". I used Kuhn's Structure of Scientific Revolutions as a starting point.

The first figure shows the relationships of the book titles as indicated by Amazon buying behaviour.



The second figure represents the net relationships of authors.



I noticed that Pajek provides a VRML export option. The 3D world of science studies is pretty interesting... although awfully disorienting.