Opening quote on the screen:
“Knowledge management was a theory or rather a Weltanschauung supported by dysfunctional technology, while social computing represents an increasingly functional technology utilizing dysfunctional and outmoded theory.“
Focus: to put some things together … to build basic theory, and then at new and emergent methods. How do you effectively do system design and project management.
Taxonomy based design and other hierarchical approaches are bound to fail … there has to be room for human intervention. Eg, a new computer system that co-evolved with human intelligence, showing people what the system thought, and what the people were doing.
System development is not a linear progression.
– doing system requirements, then getting sign-off, then building systems … bound to fail from the very beginning.
– complete distributed computing challenges many ideas about system design.
– co-evolution is a key concept … eg, the brain and language co-evolve. We don’t remember the same thing in the same way twice.
Dave’s key concepts:
– Everything is fragmented. Found that summarization of documents is perhaps the wrong thing to do, because it’s access to the broad underlying fragments that are really helpful. People reasoning from the fragments can make much more effective decisions, than those who work from summarized information.
– … if this is the case, why is KM about putting summarization on the portal? Is it any wonder that people are leaping outside of the current systems.
– People make decisions based on patterns. The only people who make rational decisions are autistic … everyone else knows there is too much information for making true rational decisions. We don’t process information to make decisions. See The Curious Incident of the Dog in the Night-Time. The two departments at University who are partially autistic are economics and computer science.
– People by nature are messy. We create structures that work for a while, but then it breaks down and becomes messy, and then we re-structure again. This is made possible through social computing.
– Complexity and constraint … there are three types of systems … an audit system (the system constrains all behavior; works okay with a stable and closed system), a chaotic system, and a complex-adaptive system (light constraints enforced by the system, but people can make changes). The latter one can’t be managed or predicted in the future. Story: planning a party for 12 year old boys … on three different ways.
– The more you put structures in place, the more you can’t see the real patterns. Have to shift from a fail-safe design to safe-fail experimentation.
– Natural numbers … 5 plus or minus 2, 15 is the maximum number of people you can trust, 150 is the maximum number of people we can have some acquaintance with. What works with small groups DOES NOT scale to much higher numbers. Fail to notice the human dynamics of collaboration.
– narrative based requirements … after 2-3 interviews, you will see what you have already seen. Therefore huge cognitive bias if you do interviews. One of Dave’s approaches is to collect 20000-30000 stories from the field, and then ask people to tag these into a central system. The metadata adds meaning to the content. Main metadata elicitation question: “What are the keywords about your story?”, not “What are the keywords in your story.”
– cross silo self-forming teams … get people to self-form into teams that have constraints. Allow people to self-form and to structure the problem in the way that works best.
– manage emergence
– “crews” … eg, you go on tour, and you go on watch. You become someone else during being on tour … you assume a role. Rituals will activate certain behaviors and patterns. Identify key roles and in expectation of roles. You don’t mind if people switch out of the team, because you can find another person who can do the role. A crew also gives a way of distributing authority without challenging status.
– “coherence mapping”
(photo to come)