a taxonomy and survey of experience bases and their use
DESCRIPTION
A Taxonomy and Survey of Experience Bases and their Use. Torgeir Dingsøyr Department of Computer and Information Science Norwegian University of Science and Technology (NTNU) E-mail: [email protected] - PowerPoint PPT PresentationTRANSCRIPT
1
A Taxonomy and Survey of Experience Bases and their
UseTorgeir Dingsøyr
Department of Computer and Information Science
Norwegian University of Science and Technology (NTNU)
E-mail: [email protected]
Presentation at the ISERN 2001 meeting, 20-22 August, University of Strathclyde, Glasgow.
2
3
Why is this interesting to us?
• Software engineering is knowledge-intensive work!
• Many companies have invested in knowledge management tools, or developed their own.
• Making use of past experience is a natural task in quality improvement.
Plan
D o
C heck
Act
4
What is the purpose of the study?
Common problems with computer systems for knowledge
management: Usage
Time
Why is this?• No culture or environment for sharing
knowledge?• Tools are
• Not useful• Not taylored
5
Knowledge Management
• What is Knowledge?• Tacit: hard to express.• Explicit: can be expressed in words or
writing.
• What is knowledge management?• Different strategies for knowledge
management• An example
6
Tacit and Explicit Knowledge
Externalization
Internalization Combination
Socialization
Tac
itE
xpli
cit
Tacit Explicit
Nonaka & Takeuchi: The Knowledge Creating Company, Oxford University Press, 1995
7
What is Knowledge Management?
”to describe, organise, share and develop knowledge”
”systematic management of knowledge-related processes to create, build, collect, organise, package, use and protect knowledge”
”a systematic process to achieve organisational goals by collecting, creating and synthesising and share information, insights and experience”
Bjørn Emil Madsen in Monica Rolfsen (ed.) Trendens Tyranni, Fagbokforlaget, 2000. (From researchers on the KUNNE project (www.kunne.no) at Sintef Industrial
Management)
8
Knowledge Management
S tr a te g y
Goals and a way toachieve them
P r o c e s s e s
M ethods to m anagetacit and explicit
knowledge
T o o ls
Infrastructure forexplicit knowledge
+ +
9
10
An Example Intranet of a medium-sized Company
• From a medium-sized company that develops software.
• Approximately 150 employees.• Consulting work is mostly done in-house.• Internally developed Intranet-based
”corporate memory” from 1997.
11
Overview of Intranet
• Knowledge Marked - Overview of knowledge resources.• Project Guide - Process guide.• Competence Blocks - Overview and adm. of internal courses.• Skills Manager - Profiles of employee skills.• WoX - Knowledge repository• Product Idea Café - Discussion forum to promote innovation.• Software Tools - Lists of available tools.• Employees - List of employees with location, phone #...• Processes - List of processes in the company.• Projects - Profiles of ongoing and finished projects.• Finance - Financial reports• Fun and reference - Web museum, quiz...
12
13
Case StudiesWhat was done?• Found 8 articles on knowledge management
initiatives, mostly ”lessons learned”• Literature study!
What the companies did:• Strategy
– Codification? (increase externalization)• Quantitative?• Qualitative?
– Personalization? (increase socialization)• Processes
– Reorganise?• Tools
14
Case Studies
What the claimed benefits were:• Lower development costs?• Higher software quality?• Better work environment?
15
Strategy
Set up a separate organisation which collected anddistributed experience.
NASA SEL
What did they do?CompanyPersonalization? Quantitative? Qualitative?
Yes Yes
Daim ler ChryslerCreated three experience factories in three differentcom pany departm ents.
YesYesYes
Ericsson Software Technology
Set up new organisational roles to increase oralcom m unication of experience.
Yes Yes
Australian Telecom Com pany
Collected ex isting explicit inform ation regarding softwaredevelopm ent and m ade it searchable.
Yes
ICL High Perform ance System s
Introduced an Intranet-based system with an "engineeringknowledge database"
Yes
ICL F inland Made an Intranet-based system with three structural layers. Yes Yes
sd&mSet up a knowledge m anagem ent group and Intranetsystem .
Yes Yes
Knowledge M anagement Approach
Telenor Telecom Software
Made an expert system based on own em pirical data foreffort estim ation and risk m anagem ent, and m odified roles.
YesYes Yes
16
Processes
• Developers should actively participate in collecting, refining and distributing knowledge.
• Separate roles for knowledge management in five companies.
• New activities like kick-off and touch-down meetings in one company.
17
Tools
• Intranet systems for knowledge repositories:– Daimler Chrysler, both ICL departments, sd&m
and also at the Australian Telecom company.
• Profiles of skills, projects and customers:– sd&m and ICL Finland
• Expert Systems:– Telenor Telecom Software
• Guides to company work processes:– ICL and Telenor Telecom Softwvare
18
19
Reported Benefits
Reduced num ber of defects, reduced software productioncosts, increased use.
NASA SEL
What was the effect?CompanyDeveloper
satisfaction?Lower cost? Higher quality?
Yes Yes
Daim ler ChryslerThe case gives no inform ation on the effect for thecom pany.
Ericsson Software Technology
The com pany claim s that the initiative was "m ore v aluable"than a database and m easurem ent-approach.
Australian Telecom Com pany
Good acceptance of product am ongst users. Yes
ICL High Perform ance System s
A perception that it has facilitated a "new m ode of working" Yes
ICL F inlandSav ed tim e, because it is easier to f ind docum ents. Easierto learn new project m em bers about project work.
Yes Yes
sd&m Prev ious problem s due to rapid growth hav e dim inished. Yes
Reported benefit
Telenor Telecom Software
The com pany indicates that estim ation accuracy hasim proved, and focus on risk m anagem ent has increased.
Yes
20
Conclusions
• Theory– Most companies have a codification
strategy, but many also focus on personalization.
– Companies focus more on qualitative knowledge than quantitative knowledge.
• Technology– Low-tech Intranets are most widely used.– High-tech tools not much in use.
• Practice– Difficult to determine benefits!
21
Appendix: How sure can we be?
• NASA SEL Lessons learned• Daimler Chrysler Lessons learned• Telenor Telecom Software Lessons learned• Ericsson Lessons learned• An Australian Telecom Company Assertion• ICL High Performance Systems Lessons
learned• ICL Finland Lessons learned• Sd&m Lessons learned
Most of the studies are done just after introducing a new ”program”
22
What is Explicit Knowledge?
Information use type
Example/ explanation
Enlightment Company-internal news.
Problem understanding
To develop a better comprehension of a particular problem.
Instrumental Guidelines, templates
Factual Billing rate, travel exprense rates
Confirmational To verify other information
Projective Forecasts, scenarios, estimates
Motivational To sustain personal involvement.
Personal or Political Develop relationships