Corporate Knowledge Repository LO431

KAI A. SIMON (kai@adb.gu.se)
Thu, 16 Mar 1995 11:18:14 +0000

I received several replies to my message "Let's get practical" which were
adressing the Corporate Knowledge Repository (CKR) I mentioned. I am
therefore sending my answer to the list.

This message contains:
- General commants about CPRs
- Reply to Michael Stein
- Reply to Steven S. Taylor

A CKR is basically a useful way of storing the knowledge gained in an
organization, and providing it to all people who might need it. The idea is
behind this is a couple of years old, and builds on the belief, that
reality can entirely be mirrored in information systems by using global
data models. Companies like IBM and Digital even developed tools for
developing CPRs. However, there are several problems related to the idea,
as it requires a standardization of the terminology within the organization
(e.g. what is a customer?), and that the information stored tends to become
complex. Additionally, there are technical and cognitive problems to be
solved as well.

Nevertheless, a less ambitious approach still appears to be useful to me.
E.g. Andersen Consulting has a global database containing 'generic
processes', that they use in their BPR-projects, and that can be accessed
by all employees, in order to use previously gained knowledge. In this
case, the intention is not to build a global model, but to support specific
purposes. Whenever new knowledge within the domain to be supported is
gained, the repository can be updated to include it fur reuse.

REPLY TO Michael Stein, Shared resources

><<"Corporate Knowledge Repository",>> Do you have a specific process or
>technique for this? It must be one hell of a flexible database. Can you say
>more? Thanks

Actually, from the beginning, the entire idea was to create global data
models, which then could be used for catalogs, aso. You probably read your
Date, so ...
However, due to several reasons most of these attempts failed, especially
in large, diversified organizations.
So, there are two optional ways. Either you try to use more unstructured
ways of representing knowledge (free-text, Hyper-text), that can be added.
This would probably result in a rather chaotic structure (Example: WWW).
Another way is to create local repositories, where only the terms that are
communicated with others are predefined. In this situation, you would
consider the organization as a network of bilateral relationships, which
occur on a basis of negotiations.

REPLY TO Steven S. Taylor

[...]
>To be a learning organization requires second
>order change, double loop feedback - whatever you call it - the ability to
>learn and change the models, frames, the culture that exists based upon
>experience and thoughtful analysis of that experience.
[...]

True. This is a reason why I am not very fond of the sequential
"input-throughput-output" model that is often used for describing
processes. IMHO, a process is a dialectical phenomenon, where the knowledge
gained during the performance of a process-instance has to be used as input
when it is performed the next time, thus allowing to improve performance in
several dimensions.
When looking at process-mapping techniques/tools, you will often find
required skills to be an input, but I am not aware of a technique/tool that
allows a definition of the desired knowledge improvement and the spin-offs
on a corporate level.

Kai

From: kai@adb.gu.se (KAI A. SIMON)