Skip to main content
Wikispaces Classroom is now free, social, and easier than ever.
Try it today.
Pages and Files
What is KM
Denham's posts on the net
03/26/06 = 6080
04/29/06 = 7200
05/25/06 = 9411
11/22/06 = 19870.5
12/17/06 = 25597.5
11/22/06 = 2490
12/17/06 = 1980
Pointers to communities of practice posts
Evolution & trajectory:
Innovation in cops:
Generating new knowledge:
knowledge across communities:
knowledge in community:
Talk is work:
Why talk cops:
CoP course key posts:
Model for a Research CoP
Bounding the discourse, segmenting the domain, establishing an ontology, surfacing & negotiating issues or distinctions, conducting a relationship audit, constructing issue profiles, supporting collaboration, networking & marketing to the outside. Need to explore the synergies between the MIT SOL learning histories and knowledge mapping. A place to start: * Field Manual for a Learning Historian (available by as part of courses or sending an Email to
Helping me to remember
My introduction to expert systems (hence KM) came during a visit to Saaveld in January 1985 by Tony Starfield
He talked about modeling decisions and capturing expert heuristics - a transformation on 01/13/1985 that changed my career forever.
Prior to expert systems, I was involved with statistical deterministic modeling using regression and multivariate techniques including cluster, PCA and some gradient analysis. Much of this was incorporated in my MSc and PhD work on site and soil matching.
After being introduced to expert systems and attending workshops by Roger Layton,
I became interested in induction and fuzzy logic systems. Did a little work with Quinlan's C4, but lacked the 'C' coding experience to really make things work.
While working for HLH research 1988 - 1994, I teamed with Margaret Brinck and developed the
and a few smaller advisors using DmX. The inputs were all done by hand, as we never completed the induction module, that would generate rules and bounds from the clonal trial data. By 1991 this was the largest commercial expert system in Africa I believe.
Social knowledge construction
Upon moving to the USA in 1994, I became involved with KM in helpdesks. Travelled around the US and Canada helping companies build knowledge bases for their call centers while working for Serviceware - a small startup in those days.
- I remember thinking that URL was way too long to be useful - my how things have changed.
There are a number of people who believe strongly in the value of personal knowledge management (PKM). Strictly speaking a person does not have access to their tacit knowledge, it is what you know and do without being aware that you know or do.
PKM seems to be mainly about information organization and contact management which is not the same as working working with 'knowledge' that has been socially learned, constructed and validated. An interesting emergent mix is klogs or kblogging where the focus is on publishing and 'pushing' personal view points, while inviting comments and delivering commentary on another's texts.
The key to knowledge work IMO is community, where you share, create, critique, validate new connections with others. Arranging reflecting and organizing your personal beliefs, perceptions and values without sharing is NOT KM as I see things.
I'm skeptical about the PKM movement as a viable route to knowledge sharing, learning and understanding here is why:
Self-elicitation is an endless circle to nowhere
Real Key discoveries mostly require external triggers
Reflection is a poor substitute for testing in dialog
Emergence and sense-making are tied to insightful questions
Knowledge needs (social) relationships to surface, flow and be tested Knowledge work is not about PKM - it is all about communcation, community, knowledge creation, conversation. Getting into a personal huddle and organising your thoughts can easily take you away from the real flow and emergence - where it all happens.
Or you could say these are just the confessions of an evangelist for the sociology of knowledge!
An understanding for the dynamics between individual 'knowing' and knowledge which is socially mediated. Context is peripheral to the distinction here. I'm saying you need to have social mediation (validation, testing, acceptance) before individual 'knowing' becomes knowledge. There are many people who experience strong individual 'knowings', but society does regards their 'knowledge' as ravings and places little value on any personal 'knowing' which is not socially mediated. My fundamental is not around the context but the necessisity and role of social mediation.
A pharma bio-processing plant
A 37% difference between leaders and laggards and a maximum shift difference of 27% within the same plant, 5 plants spread around the world all with versions of the same basic equipment and process but diffrent due to sub-component replacements and planned upgrades at different times.
Sharing tacit knowledge: We set up virtual forums in Lotus Notes for technical operators to share their tips, tricks, heuristics. We ran this for one year and reduced the differences to 9%. The final product sells for 3.4 billion USD. There are multiple other processes after making the raw stck (where we were involved).
How & Why?: There were people and cultural issues that had accumulated over the years preventing sincere knowledge sharing, strong interplant and shift competition and issues with the transfer of subtle tacit knowledge. Our forums started by defining key issues, we helped the top technicians share their 'feelings', their fine sense of what was good, what was OK and what needed intervention. One example was surges through the systems leading to poor quality stock. We helped a lead tech. explain in fine detail how to recognize the symptoms, - machine huming, and how to check go to the 3rd motor and feel if the tingling is fingers, lower or upper arm, - the solution turn 3 valves in sequence.
We improved sharing between shifts by giving folks laptops that they could use from home, we elicited a whole new set of rules for working backwards from machine readings to causal analysis.
NYC PR agency;
Contex: Partners employ female account representatives and improve staff retention by allowing work from home. Loss of 'coporate memory', informal talk and personal contact slows time to deliver new bids and there were possible savings to be had in up-front research which partners felt was being duplicated.
Solution: Work with company to develop an ontology (taxonomy / indexing system) for files and electronic documents. We reduced the time spend on searching for stuff on their LAN by a very conservative estimate of 56 hours X 27 people X $180/hour (billing rate) in one year. We facilitated many to many conversations (like here) allowing the bussy partners to pop in an and get upto speed fast. Gave the clients 'direct' access to the partners, greatly increasedn the customer and account satisfaction ratings.
We created yellowpages and repositories so new bids could rapidly access past research (found 5 copies of the same very expensive marketing report had been purchased, poor conrolm and communication between accounts did not pick this up)
Context: Very rigid highly documented process in place to track mainenance on fighter engines. Issues with sharing insights between bases.
Solution: Installed a parallel web-conferencing environment with links to the document management system. This allowed folks doing the work to ask for help and advice and if any colleagues had suggestions. We minned the repair logs, for cascading failures and for 'signs' of future failures. We were able to extract correlations and heuristics that proved very useful. One I remember was a paricular pattern of Newton rings on the turbine blades that indicated early metal fatigue (before the X-rays could detect it). Besides saving on individual repair time, we were able to keep the planes in the air for two additional years. I never was allowed to see the exact ROI.
My expertise is around helping virtual teams surface, structure and implement their tacit insights. I have worked on the people side and have a strong background in datamining.
- tracking rankings
- give it a try - Thanks to the Internet Archive
help on how to format text
Turn off "Getting Started"