You are on page 1of 12

Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.

Innovation through Knowledge Codification
Carsten Sørensen
Department for Information Systems
The London School of Economics and Political Science, London, United Kingdom
Ulrika Snis
Laboratorium for Interaction Technology
Trollhättan Uddevalla University, Uddevalla, Sweden
Academics and business professionals are currently showing a significant interest in understanding the
management of knowledge and the roles to be played herein by information and communication technology
(ICT). In this paper we take a closer look at one of the primary issues raised when supporting the management
of knowledge how to understand the role of knowledge classification and codification as means for further
organisational learning and innovation. Two manufacturing cases are analysed using particular perspectives
from current theories on classification, the management of knowledge and organisational innovation. It is
concluded that a more complex understanding of the interplay between cognitive and community models for
knowledge management as informed by research on social processes of classification can inform our
understanding of both the role of classification of knowledge for organisational innovation and on the viability
of providing ICT support based on codified knowledge.
To date, the current knowledge management debate
has shown many interesting facets. It represents a
multitude of perspectives each with its own distinct
agenda and assumptions. As a research field it is grounded
in a broad and loosely defined agenda allowing researchers
and practitioners with differing perspectives to engage in a
rich but also often somewhat confusing debate given there
are several perspectives on what might be considered to be
knowledge and the reasoning about the role of knowledge
in organisations (Kakihara and Sørensen, Forthcoming).
The debate has been concerned with issues such as 1)
knowledge work as a reflection of structural changes in
society (Drucker, 1993), 2) knowledge as the most
important resource in the firm (Grant, 1996; Prusak,
1997), 3) the distributed nature of organisational
knowledge (Tsoukas, 1996), 4) the creation of knowledge
involving translation back and forth between tacit and
explicit (Nonaka, 1994; Nonaka and Takeuchi, 1995), 5)
the individual and social aspects of knowledge creation
(Spender, 1996; Spender, 1998). Generally, critical
assumptions regarding the nature of organisational
knowledge are made. For example in terms of the
entitative perspective on "knowledge" as opposed to the
performative action-orientated perspective of "knowing"
(Blackler, 1995; Boland J r. and Tenkasi, 1995).
Integrative or pluralist frameworks relating the
management of knowledge to learning have been suggested
to accommodate these competing assumptions, for
example, by Spender (1998), Cook and Brown (1999), and
Ciborra and Andreau (2000). The viability of categorising
types of knowledge, such as the tacit/explicit distinction,
has, however, been questioned by some (Blackler, 1995),
and forwarded as a necessity by others (Cook and Brown,
1999). Thus far the debate has been both principled and
pragmatic at the same time.
The exploration of Information and
Communication Technology (ICT) as a means of
supporting the management of knowledge has in particular
proven an interesting and potentially contentious research
topic (Boland J r. and Tenkasi, 1995; Alavi, 1999;
Scarbrough et al., 1999; Swan et al., 1999a; Robertson et
al., 2000). Research informed by social science theories
argues for increased focus on the human and social issues,
such as organisational learning and the management of
human resources. At the same time, there is significant
interest within computing, engineering and industrial
communities in developing and using increasingly
complex knowledge management systems (Swan et al.,
1999a; Swan et al., 1999b). Comparing the technological
scepticism within the social science community with the
technological optimism within the engineering and
business community, there seems to be a need for a more
principled debate of the possibilities for and limitations of
ICT to support the management knowledge in
organisations. Such a debate should be informed by
organisational theories, theories on the use of ICT and
specific knowledge on the contemporary technological
options that are available. We further suggest that
investigating how to support the generic management of
knowledge is somewhat futile. We cannot assume that
knowledge is managed in a similar fashion in all
organisational settings, across sectors and company sizes.
Importantly, neither do we perceive ICT as one
technological “black box”. Furthermore, in supporting the
management of knowledge, we believe it to be highly
problematic, a priori, to assume certain allocations of
functionality between humans and technology as well as
between manual and computer-based information
technologies. Successful use of ICT necessarily relies on a
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
complex pattern of manual and technologically driven
activities. Aspects of an ICT system will be manual e.g.
forms, classification schemes, boards, notes, written
procedures etc, and inherently other aspects will be
Inspired by an earlier debate within the field of
Computer Supported Collaborative Work (CSCW) on
classification and ICT support (Suchman, 1994;
Winograd, 1994), this paper looks at the role of the
classification or codification of knowledge for
organisational innovation. This is particularly significant
for organisational innovation and ICT support for the
management of knowledge. The role of socially
constructing classifications cuts to the core of both
managing knowledge and providing ICT support in that
technological support for the management of knowledge
inevitably will rely on some form of classification. The
potential contentious nature of forming categories
(Suchman, 1994) and the need for categories as means of
managing organisational complexity (Winograd, 1994;
Carstensen and Sørensen, 1996; Schmidt and Simone,
1996) presents a challenge to our understanding of how to
provide ICT support for this activity. Furthermore, there is
a lack of research focusing on detailed processes of
technical innovation and learning through codification. As
argued by (Bowker and Star, 1999, p. 59): "Traditionally
much ethno- or folk-classification research has examined
tribal categories in nonindustrial societies. How people
in industrial societies categorise on an everyday basis is
less well known, especially in natural workaday settings".
The two case studies presented here specifically
focus on manufacturing domains as examples of
organisational innovation. In both cases the codification of
manufacturing knowledge played a central role. The
original fieldwork was conducted from the dual
perspectives of Expert Systems and Computer Supported
Collaborative Work (CSCW) respectively. However these
cases highlight the relationship between the creation of
knowledge and the process of embedding that knowledge
in systems that are subsequently deployed in the
organisation. The aim of the re-analysis of these two cases
is to bring some of the concerns from these fields into the
debate regarding ICT support for the management of
knowledge. We provide a rich and contextual
understanding of the relationships between processes of
codification of manufacturing knowledge and the
potentials for ICT support. The analysis is based on
theories of the interrelationships between classification
systems and the social processes of constructing and using
classification systems (Bowker and Star, 1999), which
also relates to theories of knowledge and knowing
(Blackler, 1995; Spender, 1996; Spender, 1998; Cook and
Brown, 1999). In order to situate the discourse of
codification of knowledge within the context of
organisational innovation, we draw upon models linking
the management of knowledge with organisational
innovation (Brown and Duguid, 1991; Swan et al., 1999a;
Swan and Newell, 2000).
The paper highlights the importance of
distinguishing between innovation process and product as
well as between types of ICT products when analysing
models of knowledge management applied for the
classification of knowledge.
The next section highlights the main theoretical
perspectives used in this paper for analysing classification,
the management of knowledge and organisational
innovation. Section 3 discusses methodology, and Section
4 presents and analyses the two cases. Section 5 discusses
the findings and relates them to contemporary knowledge
management literature regarding codification,
organisational innovation and ICT support. Section 6
concludes the paper.
Classification, Knowledge Management and
This section briefly outlines relevant theories
concerning the social aspects of classification, knowledge
management and the relationships between knowledge
management and organisational innovation.
In this paper we aim to analyse the role of
classification and classification systems for process
innovation in manufacturing. Therefore the understanding
of both social and technological aspects of classification
and codification is at the centre of the analysis.
Classification is defined as spatial, temporal, or spatio-
temporal segmentation of the world exhibiting the
following abstract characteristics: consistent and unique
classification principles in operation, mutually exclusive
categories, a complete system classifying everything under
consideration (Bowker and Star, 1999). If the classification
structure has emerged as the result of multiple, and
possibly conflicting classification principles, it is called a
nomenclature. Codification is related to classification, in
that it means reducing to code and is synonymous with
arranging, cataloguing, classifying, condensing and
organising (Infopedia, 1996, Roget's Thesaurus and
Webster's Dictionary).
Classification is, according to Bowker and Star
(1999), often associated with ubiquity in the sense that
classification structures and classification work often goes
on unnoticed with everyday working life. Classification
activities depend on the use of a multitude of standards
and classification structures. Classification schemes are
both material and symbolic often implementing a mixture
of physical entities, e.g., paper forms and software
instructions, and conventional arrangements such as speed
and dimension. Classifications suffer from the
indeterminacy of the past in that we constantly revise our
knowledge of the past in light of new developments. In
addition, classifications are constructed and reproduced
through detailed negotiations, organisational processes and
conflict according to the practical politics of classifying
and standardising. Finally, the dialectics between formal
classification and the practice of interpreting the formal
structures is one of the key issues in providing ICT
support for collaborative work activities (Winograd, 1994;
Bowker and Star, 1999; Schmidt, 1999). This also
implies that studies of classification have tended to move
away from dichotomising the formal and the informal
aspects of classification (Bowker and Star, 1999, p.54,
Knowledge Creation and Knowledge Management
For the purpose of the analysis presented here we
assume that organisational knowledge is highly distributed
(Tsoukas, 1996). Therefore, when studying classification
and codification of technical knowledge several
perspectives can be of importance, such as the role of the
implicit and the tacit in knowledge creation (Nonaka,
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
1994; Nonaka and Takeuchi, 1995), the individual and
social aspect of knowledge creation (Spender, 1996;
Spender, 1998), and the role of geographical and
conceptual situatedness (Nonaka and Konno, 1998). All
the major frameworks including Cook & Brown's (1999),
Spender's (1996; 1998), and Nonaka's (1995) distinguish
types of knowledge, for example tacit/explicit,
individual/collective. Cook and Brown (1999) aims to
integrate the concern for the individual possessive
perspective with a collective, in terms of communities of
practice (Brown and Duguid, 1991), and an action-
orientated perspective, for example as expressed by
(Blackler, 1995). They argue that practice is distinct from
both action and behaviour and define practice as "action
informed by meaning drawn from a particular group
context" (Cook and Brown, 1999, p. 387). Practice is
what converts tacit knowledge into explicit knowledge.
This, they argue, will integrate knowledge and knowing,
and the two interactively constitute each other. Our interest
here is to investigate the practice of both classification and
codification of knowledge as an organisational innovation
process, as well as the system emerging as its product.
Knowledge Management and Organisational
Swan et al. (1999a) formulate two distinct
perspectives on knowledge management for innovation,
the cognitive and the community model. The community
model is formulated as a critique of the predominant
cognitive perspective within the technology driven research
field (Scarbrough et al., 1999; Swan et al., 2000). The
cognitive model denotes a perspective where valuable
knowledge is conceived as being captured and codified
from individuals, packaged, transmitted and processed
through the use of ICT, and hence disseminated and used
by other individuals in new contexts. In this perspective,
knowledge can also be exploited through the recycling of
existing knowledge “owned” and “experienced” by
individuals in cognitive networks. Here, ICTs are seen as
critical success factors. In contrast, the community model
portrays the management of knowledge as socially
constructed through interaction within communities of
practice. Communities of practice consist of collections of
individuals, between whom there is collaboration and
negotiation. Knowledge creation and learning are processes
making sense of knowledge in social activities deeply
rooted in daily work practice. Within the community
model, information and communication technology can
play a role, even though it is not seen as a critical success
factor. Table 1 summarises the main characteristics of the
cognitive and the community model.
Reviewing the literature, we find that the
distinction between a cognitive on the one hand and a
community model on the other, appropriately describes the
two main approaches both to knowledge management, but
in particular to a debate of the role of ICT in the
management of knowledge. The cognitive model, focusing
on the crucial role of technology as the mediator of
codified knowledge, represents the technologist view and
promotes the view that knowledge can be managed by
codification. The community models, focusing on social
interaction and negotiation, promote the idea of supporting
interaction and collaboration in order to manage
knowledge. When considering issues related to the creation
and sharing of technical knowledge, and to the role of
technical and organisational measures in doing so, we are
left with the question of who to trust? The technologist
touting the splendour of technology, or the social
scientists questioning this view? This is particular
interesting when considering the need to codify knowledge
into classification structures as a means for organisational
learning, innovation and supporting the management of
complexity (Winograd, 1994; Schmidt, 1999)
We will in the following seek to critically
reappraise the relative roles of the community and
cognitive knowledge management models for
understanding organisational innovation in manufacturing
through the codification of manufacturing design
Re-analysing Two Cases from
The cases are both drawn from studies where
qualitative interviewing, participant observation and
document inspection were the three primary data collection
methods used (Patton, 1980; McCracken, 1988; Cash and
Lawrence, 1989; Carstensen and Sørensen, 1994). Both
studies, but in particular Case A, contain elements of
action based research (Braa and Vidgen, 1999). The data
was not intentionally collected for this particular analysis.
However, the focus of both studies was to consider the
relationships between everyday manufacturing processes
and the development and use of ICT.
Cognitive Model Community Model
1. Knowledge for innovation is equal to objectively
defined concepts and facts.
2. Knowledge can be codified and transferred through
text: information systems have a crucial role.
3. Gains from KM include exploitation through the
recycling of existing knowledge.
4. The primary function of KM is to codify and capture
5. The critical success factor is technology.
6. The dominant metaphors are the human memory and
the jigsaw (fitting pieces of knowledge together to
produce a bigger picture in predictable ways)
7. Knowledge for innovation is socially constructed and
based on experience
8. Knowledge can be tacit and is transferred through
participation in social networks including
occupational groups and teams.
9. Gains from KM include exploration through the
sharing and synthesis of knowledge among different
social groups and communities
10. The primary function of KM is to encourage
knowledge-sharing through networking
11. The critical success factor is trust and collaboration.
12. The dominant metaphors are the human community
and the kaleidoscope (creative interactions producing
new knowledge in sometimes unpredictable ways)
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
Tabl e 1: Two contrasting views of the knowledge management process from (Swan et al., 1999a).
The intention is not to communicate the full
contextual characteristics of the case studies. Instead the
purpose is to focus on a few aspects pertinent for the
discussion of organisational innovation processes, the
codification of knowledge and issues concerning ICT
support. Both cases focus on process innovation - Case A
within a particular, complex manufacturing process (Snis,
1997), and Case B in the engineering design and
documentation specifications department (Sørensen, 1994;
Carstensen and Sørensen, 1996). In the following we
briefly outline the original methodologies applied for data
collection, and the following two sections then present and
discuss the cases in detail.
Case A: Volvo Aero, Sweden
Case A represents a particular aspect of field work
conducted in the Thermal Spraying Department at Volvo
Aero Corporation in Trollhättan, Sweden. The company
develop, produce and maintain jet engines for military as
well as civil use (Snis, 1997). Initially, there were vague
understandings about the problem domain. Thus an
explorative investigation was initiated before outlining the
issues of concern. During 16 months the empirical
material was collected from semi-structured interviews
with 19 employees - three material engineers, two
laboratory engineers, two managers, five production
engineers and seven operators. The study also involved
participation in a number of meetings concerning the
Thermal Spraying process, participant observation with the
people interviewed and analysis of internal documentation
from the Thermal Spraying department. A generic
interview protocol was used in order to identify relevant
questions and problems for further exploration. As the
study progressed, common themes began to emerge and a
reference group of major stakeholders in the problem
domain was formed in order to discuss the relevance of the
empirical findings. The purpose of the study was to
establish the requirements of an expert system to support
the improvement of the quality of the Thermal Spraying
Case B: Foss Electric, Denmark
Foss Electric is a Danish manufacturing
organisation that develops, manufactures and markets
instruments for automatically measuring quality
parameters of agricultural products, such as measuring the
compositional quality of milk, composition and
microbiological quality of food products and the quality of
grain. Designing these instruments involves a range of
expertise, for example from the disciplines of mechanical,
chemical, electrical and software engineering. The
organisation had implemented concurrent engineering
methods in a matrix organisation in order to shorten lead-
time. The aim of the empirical effort at Foss Electric was
to analyse co-operative work in a manufacturing setting
where the participants deal with the complexity and
uncertainty of going from a design idea to determining
how to manufacture the product. Two computer scientists
and one manufacturing engineer studied the engineering
design, process planning and software design at Foss
Electric. Approximately 20 open-ended interviews were
conducted (Patton, 1980). More than 10 project meetings
were observed, and just over 100 person-hours of project-
observation was conducted over a period of approximately
4 months. This project observation was followed up by
several meetings with the project members. At these
meetings we presented and had feedback on our
observations and interpretations of their work. The research
approach used in collecting data at Foss Electric can be
characterised as qualitative research inspired by
ethnographic approaches to studying collaborative work
(Heath et al., 2000) and engineering work (Bucciarelli,
Two Cases of Knowledge Codification
Case A: Codifying Thermal Spraying to Improve
Process Quality
The manufacturing process of thermal spraying
played an important role in ensuring the highest attainable
quality in new jet engine designs. Consequently the
organisation had over the years developed considerable
high-tech competence in this manufacturing process and 40
people was employed in the thermal spraying department:
two managers, three material engineers, two laboratory
engineers, five production engineers and seven operators
responsible for masking, cleaning and spraying
components. Thermal spraying involves partly melting
and throwing a material, usually a powder, onto a
substrate where a coating is built up by the condensing
particles (See Figure 1). The application and the desired
properties of the coating determine possible spraying
methods and materials. Metals, alloys, carbides, plastics
and composites can be applied by thermal spraying.
Thermal spraying had a troubled history concerning
quality assurance. The multitude of factors affecting the
quality and the difficulties involved in measuring,
estimating and testing process quality resulted in problems
of consistency and reproducability. The quality of the
coating was influenced by a complex set of parameters.
For example, in plasma spraying about 50 macroscopic
parameters needed to be adjusted. The set point
determination of the process parameters was often a matter
of trial and error and was time consuming. Moreover, the
stress build-up in the coating, which was determined by
the cooling conditions of the droplets on the surface, and
of the successive deposited layers, varied as the coating
grew due to changes in the local thermal field. In state-of-
the-art thermal spraying, the set point parameters were
determined for the entire duration of the praying process
with no consideration to the changing conditions at the
coating surface during spraying. These factors led to
coatings with variable quality due to the lack of control of
defects such as microscopic cracks, porosity.
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
- Current
- Gasflows
- Powderfeed
- Etc.
- Temperature
- Velocity
- Size
- Position
- Temperature
- Microstructure
- Hardness
- Strength
- Stresses
- Rel. properties
- Life
- Performance
powder particles
nozzle (anod)
Fi gure 1: A schematic drawing of the thermal spraying process indicating the four main categories of variables
affecting the process and product quality.
The lack of quality and reproducibility increased
the manufacturing costs and limited the expansion of areas
for applying thermal spraying. At the time of study, rapid
changes were being introduced across the world, applying
advanced continuous measurement and data processing
systems with the purpose of ensuring a higher quality
thermal spraying process. The types of problems which
thermal spraying employees faced can be summarised as
incomplete information about the state of the process and
inconsistency and uncertainty of what knowledge to
collect, codify and apply to the process.
In order to improve the quality of the thermal
spraying process, the organisation initiated a research
project with the purpose of investigating the main factors
affecting process quality. Part of this project involved the
critical appraisal of the management and co-ordination of
primarily tacit knowledge with the purpose of codifying
this knowledge into an expert system (Snis, 1997). It was
intended that the expert system would intentionally allow
dissemination and distributed application of the codified
process knowledge through a technology driven interaction
between a human operator (allegedly) without deep domain
knowledge and a computer-based expert system on thermal
spraying processes (Russel and Norwig, 1995; Turban,
1995). The project did not involve the task of automating
existing explicit rules and principles previously codified.
Rather, the task of the knowledge engineer was to carefully
elicit complex and tacit process knowledge situated
throughout the various professional groups involved, thus
making it explicitly formulated as rules that could be
codified into a new expert system. As a result of
preliminary interviews, more questions were raised than
answered and the extremely complex nature of the tacit
process knowledge was identified. One such example was
the crucial local adjustments requiring both experience and
skills made by the operators who monitored and controlled
the process. These were made as a direct result of observed
process anomalies, such as a 'strangely' shaped flame that
was caused by too low power feed in the robot gun. The
study demonstrated that it was not only technical factors
that affected the quality of the process. It highlighted the
crucial role of human judgement exemplified in the
experience of the operators, personnel training and
education, and collaborative efforts among workgroup
members, etc. Internal inconsistency within the rule-base
for quality highlighted another example of the problems of
codifying the tacit process knowledge. Pre-established
different and conflicting rules promoted that this
inconsistency was premised on different sub-activities and
roles. As a direct result of these significant problems, the
expert system project was ultimately abandoned.
This case exemplifies an organisation looking for a
technical fix. The aim was to resolve a very complex issue
of manufacturing process improvement through a fairly
straightforward, although mentally complex, individual
“knowledge elicitation” process. The process of codifying
the complex, distributed and partly conflicting process
knowledge was stifled because the knowledge elicitation
process was conducted by one knowledge engineer who
was not an expert on thermal spraying but an expert
systems expert. A hypothetical collaborative negotiation
process would have had to take into consideration the
different perspectives of people involved in various aspects
of thermal spraying, such as robot operators, robot
programmers, engineering designers, process planners,
quality assurance experts, and project managers. The
people in thermal spraying did not work as a community
and they did not share a common view. The fact that the
knowledge engineer interviewed people individually and
discussed with them the possible codification of the rules
governing the thermal spraying process did not amount to
a proper unfolding of collective knowing - the participants
did not engage in a generative dance involving individual,
collective, tacit and explicit knowledge on the one hand
with the deeply rooted practice of thermal spraying
operational work on the other (Cook and Brown, 1999).
Constructing an expert system would have been an effort
associated with great risk, given that the high degree of
complexity and the unfavourable environment (Mockler,
1992). There were far too high expectations around a
technological solution. The knowledge which was to be
codified was deeply rooted in human expertise, an
expertise that would require significant collaboration and
negotiation to even begin to formulate. Had the firm
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
somehow succeeded in the endeavour, the solution would
have constituted a fairly complex knowledge management
technology in combination with a relatively simple
organisational solution - in terms of individual members
of the process being able to consult the expert system. As
far as the participants were concerned, the problem was
purely a technical one. This implied that the solution
should be found in the technical domain. Attempts to
persuade them that thermal spraying is a socio-technical
process proved quite difficult. The attention of the
organisation was closely focused on the technical
parameters, and a closer inspection of the espoused
theories on factors affecting the quality of the
manufacturing process demonstrated a rich socially
constructed picture of both conflicting and non-technical
explanations. This case thus exemplifies codifying and
'packaging' individual, conscious knowledge with the
purpose of transforming it via an expert system into
objectified knowledge using an objectification strategy for
knowledge communication (Scarbrough, 1995; Spender,
Case B: Collaborating to Codify Engineering Design
At the time of study, Foss Electric had just
adopted Computer-Aided Design (CAD) workstations as a
replacement for conventional paper-based design tools.
This could facilitate faster development processes, less
uncertainty regarding sub-assemblies and a more rapid
transfer from engineering design via process planning into
production. The introduction of CAD workstations also
created the need for reuse of specifications. One of the
advantages of using CAD instead of a traditional paper-
based system was the improved opportunity of reusing old
components in new products or at least reusing the
specification of standard components. This could both
save time and support standardisation. Standardisation of
components and reuse of component specifications across
projects was, however, crucially dependent upon
distributed storage and retrieval of CAD specifications
across temporal and spatial barriers. Categorising a CAD
specification would enable one person to store it and
another to subsequently find and reuse it. Hence, in
relation to introducing CAD, the company developed the
classification scheme illustrated in Figure 2 capturing
components and units for all instruments produced.
It was ordered in a tree-structure with classes,
categories, sub-categories, and sub-sub-categories. There
were 16 classes, and approximately 357 different sub-
categories categories 1 to 4 levels down from the class. As
an example, class number 5 was hydraulic and pneumatic
components, which had 11 categories. One of these was
“valves”, which had 6 sub-categories and no sub-sub-
categories (see Figure 2). At the most detailed level, each
sub- or sub-sub category is specified in a data management
system according to the predefined database records.
7. Spring/Damper
9. Distancing
Rod (3 sub)
12. Media-handling
(10 more)
13. Other mechanical
Other component
Other units
16. Paper
14. Instruments
Specialist Tool
1. Electrical connection
Plug accessory
Elect. connector
Cable bunch
Cable with plug
Wire with plug
4. Optics
(7 more)
15. Signs
5. Hydraul i cs/
Reduction valve
Safety valve
Tube valve
2/2 valve
3/2 valve
Magnet system
Pressure cylinder
Piston cylinder
(2 more)
6. Seal s / Protecti on
Seal component
Protection component
11. Casing
Plate casing
Cabinet accessories
2. El ectroni cs
(11 more)
(3 more)
(3 more)
10. Fixing components
Sixedge bolt
(6 more)
3. Measuring comp./
Signal source
8. Transmission comp.
Transmission mechanical
Transmission optical
Transmission thermal
Fi gure 2: The product classification scheme, as printed on an A3 sheet of paper, outlining the 16 main
groups, some of the 79 categories, examples of the 177 sub-categories and none of the 99 sub-sub-
categories or 2 sub-sub-sub categories.
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
This will in greater detail specify the type of data
saved about a specification, such as physical
characteristics, the name, what project it was specified in,
who stored the specification, the inventory part number,
the material and form. The product classification scheme
provided a conceptual structure making it possible for
draught-persons and engineering designers to perform
distributed storing and retrieval of CAD models. It
reflected a common standard for categorisation of the
components and units in any instrument produced by the
company and can be viewed as the negotiated order of how
an instrument could formally be described. The
classification scheme was partly paper-based represented on
A3 size prints and partly computer supported by an
alphabetically ordered list of categories integrated into the
CAD system. The classification process was not stipulated
in any explicit organisational procedure. The draught-
person performing the classification chose the appropriate
category from the printed scheme and subsequently entered
it into the data management system by selecting the
appropriate category from the list of categories in the CAD
The classification scheme appears to be more a
nomenclature than a "proper" classification structure in that
it is not constructed from one governing principle. Instead
it can be viewed as the pragmatic result of negotiation and
the need for a working model of the products designed at
Foss Electric. The structure displays several "other"
categories at several levels enabling the everyday
accomplishment of classification work to go smoothly.
Category 13 is a whole class intended to capture any
mechanical component not captured any other place in the
structure, and there are "other" categories at lover level
scattered throughout the scheme. In this respect the scheme
is a good example of the day-to-day balancing of formal
and informal aspects of classifying. The classification
system is thus prototypical and is based on the emerging
shared understanding of how to classify components. The
structure represents a boundary object in that it spans
several professions within Foss Electric from the drafts
person conducting the initial classification, the engineering
designer re-using the specification in a subsequent design,
to the process planner who uses the detailed specification.
The textuality of the classification scheme is manifested in
the physical layout and numbering of the general
categories. The sequencing of categories interestingly is
made according to the inner logic of balancing importance
of class with the sequencing of activities, hence 1)
electrical and 2) electronics before 5) hydraulics and 11)
casing. The sequencing of categories produces an uneven
textuality when printed on an A3 sheet of paper but makes
perfect sense when each category is picked from a
sequential list within the CAD system.
Significantly, the initial scheme was designed by
an ad hoc group of six people who spent a number of
weeks designing it in their spare time outside of normal
working hours. The work of determining the categories
promoted major discussions and even fierce arguments
amongst the members of the group. As in the case of the
International Classification of Diseases (Bowker and Star,
1999), the use and management of the product
classification scheme can be characterised as a struggle of
carefully finding “the appropriate level of ambiguity”. It
was not a primary requirement that any CAD specification
stored by one person should be easily retrievable by
another, rather the idea was to support retrieval of the most
commonly used components and sub-assemblies. These
components were most often characterised by a very well
defined functionality, and therefore the classification
scheme was primarily based on functionality. New
components and units were constantly designed as a result
of technological innovation in measurement technologies
and manufacturing processes, at times making it difficult
to perform the classification. This did not however, result
in constant changes to the classification scheme. Because
of the highly distributed use of the scheme, changes had to
be negotiated. The categories were from time to time
modified, and new categories were added in order for the
scheme to represent the type and function of components
specified. Changes to the scheme were results of
negotiations between designers and draught-persons at
designated meetings. The classes and categories in the
scheme were based on a mix of functional and geometrical
properties of components and units. Some of the
categories reflected the practical problems of classifying
components. There was, for example, Class 13 named
“other mechanical components and units” containing
categories such as: “console”, “plate”, “cylindrical
component”, “tube component”. This gave the draught-
person a means of classifying non-standard irregular
components, which otherwise would have been impossible
to fit into the scheme. It was however reported by several
interviewees, that the people who originally had designed
the scheme found it much easier to use than people who
had not been directly involved.
This case demonstrates some radically different
characteristics to Case A. The organisational problem was
relatively simple, i.e., co-ordination support for
distributed storage and retrieval of CAD specifications
with the purpose of reuse. The problem was addressed
however by a relatively complex process of interaction
between a group of six domain experts with deep
knowledge. These individuals spent their spare time over
several weeks arguing and negotiating until, in functional
terms, they were able to present a theory of what the
company was all about. This represented amongst the
subgroup a legitimate example of collective classification
work where the deep prior experiences from practice was
allowed to engage in interaction with espoused and
possessed theories of the domain of work considered
(Bowker and Star, 1991; Cook and Brown, 1999). The
solution was in technological terms stunningly simple. It
consisted of a tree-structure to be used in association with
implicit conventions for classifying, rendering the use of
the technology reliant on tacit knowledge. Knowledge was
communicated through a combination of objectification
(i.e., the classification structure) and professionalism (via
the tacit conventions that were unconsciously applied
when using the scheme) (Scarbrough, 1995). However, the
process that brought about the classification scheme was
mainly driven by professionalism in the sense that
constant negotiation and discussion were the primary
means for bringing about or creating the “packaged”
knowledge. In Spender’s (1996) terms, the individuals in
the ad hoc group combined their conscious and automatic
knowledge resulting in the development of collective
knowledge necessary for producing objectified knowledge
residing in the classification structure. This case is a good
example of the community model of knowledge
management in the sense that the important aspects of
knowledge creation or innovation were the establishment
of a community of workers negotiating the categories and
the structure of the classification scheme. Information
technology was of course a crucial element, but only in
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
terms of recording the results of the group interaction and
as an externally visible representation of the codification of
tacit knowledge.
The two cases discussed above can be
accommodated largely within the classification scheme
developed by Swan et al. (1999a) which highlights the
cognitive and the community approach to the management
of knowledge. Case A displayed a particular configuration
of knowledge management models through the process.
There was little reliance on the community model since
the expert systems expert drove the interview and
observation process with some debate in a steering
committee of stakeholders. The innovation process
attempted to go directly from agenda formation to the
codification of manufacturing process knowledge. From
the beginning the project stipulated the selection of expert
systems technology with the purpose of codifying
organisational knowledge. Case A renders itself relatively
easy to be classified as an example of the cognitive model
(Swan et al., 1999a), as can be seen in Table 2, left
column. In Case B the organisational innovation process
was based entirely on internal knowledge with no need for
boundary spanning activities. The main aspect of this Case
was a complex process of negotiating the design of a large
classification structure representing a mixture of tacit,
explicit, individual and collective knowledge (Cook and
Brown, 1999). The right column of Table 1 illustrates
Case B according to the community model. We conclude
that the successful codification of technical knowledge
within manufacturing for organisational innovation must
be based on a collaboratively established consensus
reflected in the community model. The interdependent
actors bringing different skills to highly complex technical
work processes will therefore have different perspectives
(Schmidt and Simone, 1996). Classification as a social
process is therefore subject to continuous renegotiations
(Bowker and Star, 1999).
Analysing the two cases, we gain more substantial
insights in the complex interrelationships between models
for knowledge management, classification work,
organisational innovation processes, and ICT support. To
frame this debate we initially highlight the distinction
between the innovation process and the innovation
product, proposed as two essential aspects of systems
development (Bannon, 1993; Mathiassen, 1998). Crucially
a distinction between the process of shaping ICT and the
type of technological product produced will further support
discussion of types of ICT support for knowledge work.
Secondly, we introduce the basic distinction
between technological support for the individual and for
automation as opposed to support for collaboration
(Schmidt and Bannon, 1992; Bannon, 1993). The
introduction of this distinction is motivated by Swan et al
(1999a) "black-box" view of technology in the sense that
ICT is promoted as a critical success factor within the
cognitive model and a barrier within the community
model. Although their arguments concerning the
challenges of codifying and technologically embedding
tacit knowledge are indeed valid, a simple distinction
between two views on information technology can
facilitate a deeper analysis of the two cases. The
distinction between an individual and a collective
perspective on ICT relates to the changing views of
computer applications in terms of three eras: 1) As separate
entities increasing productivity and efficiency through the
automation of existing manual processes; 2) as networked
entities supporting collaboration between professional
groups; and 3) as integral entities of business strategies
and of global networks facilitating collaboration across
organisational and national boundaries (Mathiassen, 1998).
The distinction has also been cultivated from a computer
science perspective, distinguishing between algorithms and
interaction (Wegner, 1997).
Case A: Codifying Thermal Spraying Case B: Classifying CAD Models
Thermal Spraying Quality Assurance Project.
Project aims to overcome problems of quality control
within the thermal spraying process
Collaborative development of classification scheme for
CAD models.
Scheme aims to support distributed storage, retrieval and
reuse of CAD specifications.
In the main, the organisation viewed the effort as one of
objectively defining and codifying the rules governing
the process qualities.
The ad-hoc group emerged informally as a response to
the need for reuse of Computer-Aided Design (CAD)
specifications when the organisation migrated from
paper-based to computer-aided design.
The rule-base specifying the qualitative rules affecting
process quality was elicited by the expert interviewing
different people in the organisation, reflecting the
various roles involved in thermal spraying.
The classification scheme was constructed as
collaborative after-hours “skunk works” over a period of
several weeks characterised by many discussions and
Expert system at the centre of the development effort. System of experts responsible for critical aspect of
designing the codification of the manufacturing domain
in the organisation.
The development of a computer-based expert system
technology was a critical success factor
The CAD system support for the classification system
was a minor but helpful feature. However, the printed
classification scheme was considered crucial.
Project unsuccessful in providing a viable technological
solution. It was concluded that due to the individual
participants’ different and conflicting opinions
regarding the rules and the ranking of the rules affecting
the quality of the thermal spraying process, an expert
system would not solve the problem at that stage.
Project resulted in a classification scheme that was
subject to constant debate in the organisation.
Characteristically, most critique was voiced by people
who had not participated in the design of the scheme and
therefore had not had their perspective sufficiently
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
Tabl e 2: Summarising and categorising Case A and B according to the cognitive and the community model
(Swan et al., 1999a).
Figure 3 illustrates the positioning of the
innovation processes and innovation products in Case A
and B with respect to the cognitive and community
models. Regarding the innovation process, the weakness
of Case A, the lack of negotiation and socially situated
interaction, was the strength of Case B. Had the latter
process entailed one person without deep domain
knowledge defining a classification scheme, chances are
that the scheme would never have been adopted. Case A
demonstrated the problems associated with the lack of
community (Swan et al., 1999a) and practice (Cook and
Brown, 1999) in the innovation process. This resulted in
the process being abandoned altogether. Case B,
demonstrated the strength of a community for codifying
organisational knowledge. In terms of innovation product,
the aim of Case A was to encode the rules governing the
thermal spraying process and embed them into a tool
partly automating the process. Since the process
essentially was viewed exclusively as a technical process,
human interference was considered negligible. In Case B
the aim was to produce a partly manual and partly
computer-based tool supporting collaboration.
Carstensen and Sørensen (1996) argue that a need
for systematic support for co-ordination of distributed
activities can emerge from the increased complexity of
work processes. In Case B, the classification system
reduces the complexity of distributed storage and retrieval
activities. The formalisations of agreed categories and sub-
categories of components in instruments together with
clearly defined data-sets that must be entered for each
stored CAD specification made it possible for people
within or across projects to store and retrieve specifications
without any significant need for “expensive” person-to-
person interaction.
It could be argued that without interaction and
community building no system will successfully address
the thermal spraying problem in Case A. It was not a
matter of finding the “technical fix” but rather of creating
new situational and generic knowledge on the optimisation
of multiple parameters. Eventually a system may be able
to support the storage and use of the principles governing
the quality of this manufacturing process. Initially,
however, the most effective way of addressing the problem
would be community building and negotiation. We
therefore argue that the expert system should be replaced
by a system of experts. These experts could, in terms of
knowledge exploration, exchange explicit, tacit, individual
and collective knowledge regarding the dynamics of the
rule base for optimising quality.
The organisational “blindness” inhibiting the view
of thermal spraying as a social process could not be
explained by a predominant engineering or technical
culture, given Case B also was from the manufacturing
domain. However, the organisation in Case A
conceptualised the problem as related to improvements of
the processes itself - an aspect that directly added value.
Case B, on the other hand, demonstrated improvements in
the co-ordination of the process rather than adding value
directly (Carstensen and Sørensen, 1996; Schmidt and
Simone, 1996). Here, relatively little direct support was
provided for the design of components that were re-usable
in other contexts (a value adding process). Instead, the
support was focused on facilitating an effective distributed
'negotiation" of where to place and where to look for
potentially useful CAD specifications.
The fact that the two cases each fall within one
model for knowledge management does not imply that
different configurations cannot be encountered in other
circumstances. It is, for example, not obvious that the
application of a community model for exploring
organisational knowledge necessarily will lead to
technological support for collaboration. For example in
Case A, an innovation process initiating within the
community model (C in Figure 3) as a precursor for the
codification process could perhaps have established a more
solid basis for developing a complex expert system.
Cogni tive Communi ty
Case A
Case B
I: Knowledge
III: Collaborative
Knowledge Exploration
II: Tool for Individual
or for Automation
IV: Collaboration
Knowledge Management Model
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
Fi gure 3: Within Swan et al's (1999a) framework, distinguishing between the innovation process and the
innovation product, cases A and B each match a model for knowledge management. However, two other
configurations of innovation process and product can be envisioned (C and D)
Also, Scenario D (See Figure 3) could, for
example, be the stipulation of a simple collaboration tool,
for example supporting communication using simple
categorisations as discussed by Robertson et al (2000), or
providing a simple common information space for shared
work activities (Schmidt and Bannon, 1992; Bannon,
1993). A more comprehensive support for collaborative
activities in the form of co-ordination mechanisms
stipulating the co-ordination of distributed activities
would to a large extent rely on classification structures and
computer-based protocols (Carstensen and Sørensen, 1996;
Schmidt and Simone, 1996). Such support would most
likely not be a successful outcome of knowledge
exploitation, but would rely on collaborative processes of
knowledge exploration in order for the participants to
negotiate the appropriate classifications and algorithmic
properties of the support. A central aspect of this
negotiation concerns the allocation of functionality
between humans and formalised systems as well as
between computer-based and non-computer-based artefacts.
In Case A, these decisions were to a large extent made at
the outset. The expert system would by definition embed
essential rules modelling the thermal spraying process, and
as such automate and "hide" the complexity. The purpose
of the technological support was to automate the
standardisation of process quality. In Case B the purpose
of the knowledge exploration process was to uncover the
assumptions, opinions and experiences of people with
different professional backgrounds about products and
components, and express the negotiated order as a
formalised classification system - to discuss, express and
formalise the complexity. The knowledge possessed by the
group was drawn out through the action of debating (Cook
and Brown, 1999) or socialisation (Nonaka and Takeuchi,
1995). In Case B the allocation of functionality between
humans and different types of manual and computer-based
information technology was not determined from the start.
The result could potentially have been a simple paper-
based classification scheme that mapped onto a set of
directories for storage of CAD specifications. However, the
end result was a complex system with an essential paper-
based part - the printed classification scheme - as well as
the complex computer-based part in the form of a database
system integrated with both the CAD and material
planning systems. An essential aspect of the classification
system in Case B is that although it supported
collaboration and learning within and across projects, it
did not stipulate this collaboration. The individual person
using the system could choose when and how to classify a
CAD specification. In this respect major aspects of the
functionality resides outside the formalised system.
Swan et al (1999a) highlight the problems of
innovation supply and demand. Case B illustrates a
demand approach, especially in terms of the negotiation
process, but also to some extent in terms of the use of the
classification scheme. The ad hoc group codifying the
theory of products produced by the company did so out of
a perceived need. No manager told them to participate in
these activities. However, it could be argued that the
implicit pressure on members of the organisation to use
the scheme is slightly biased towards being supply-driven.
For example, the additional time spent classifying a CAD
model would perhaps benefit someone else and not
necessarily the person classifying the components. The
egalitarian culture in the organisation meant that this was
not perceived as a problem. It is, however, important to
clearly distinguish between innovation process and
product. The classification scheme was a result of a
negotiation process amongst a small group of people
exploring new knowledge. This knowledge was codified
into the classification scheme, which with its tacit
conventions for use represented exploitation of the
knowledge. The product of the innovation - the scheme
with associated heuristics for the use of it, was
subsequently the subject of a diffusion process within the
organisation. Over time the classification scheme gradually
became out of “sync” with the organisational
manufacturing reality that it was created from, it was thus
subject to re-coding through further negotiation, critique
and exploration. The initial collective theory building
exercise was constantly challenged and informed by the
practice of everyday use of the classification scheme, thus
illustrating the duality between knowing and doing or
classifications and classifying (Bowker and Star, 1999)
Given the increasing geographical and temporal
distribution of work activities, and the emergence of
information and communication technologies in most
walks of organisational life, there is and will be an
increasing need for computer supported codified
knowledge. There will also increasingly be a need for
applying ICT in the process of codification itself. The
process of co-ordinating, negotiating and planning work
applies recursively to the co-ordination of work (Schmidt
and Simone, 1996). Therefore, to the extent that we can
reduce the complexity of negotiating where CAD models
are located on a computer network, we can perhaps also
through ICT support the negotiation of the categories we
use when classifying. Communities that only exist
virtually and interact through computer-based networks
have little other choice. They establish principles and
systems for categorising and codifying the world they
inhabit (Sørensen, 1999). For many years, a process of on-
line negotiation has governed the emergence of new Usenet
News Groups. By focusing exclusively on the community
model, we may overlook opportunities for technology
supported implicit codification by observed behavioural
patterns. In Case B, categories rarely used could be drawn
to the background with occasional votes as to which of the
infrequent ones should remain, what new categories could
be suggested and which ones no longer were relevant. We
could argue that the notion of classification structures as
rigidly codified aspects of the world could be de-
emphasised partly through the use of ICTs.
People and ICT are increasingly interwoven.
Addressing the relationships between people creating and
managing knowledge, and systems supporting, facilitating
and enabling them to do so, involves complex
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
considerations and difficult design choices. This paper has
discussed organisational innovation through the
classification and codification of manufacturing
knowledge. Such classification and codification processes
will, if successful, lead to the development of information
artefacts. It is a complex process to negotiate what aspects
of the information systems should be formalised and
embedded into information artefacts and what aspects
human actors should carry out. Furthermore, the
information artefacts can be entirely physical, for example,
as paper-forms, they can be digitally embedded into ICT,
or they can be a complex mixture of both (Carstensen and
Sørensen, 1996). This paper has analysed two cases of
knowledge classification and codification both aiming at
innovating manufacturing processes through ICT. The
analysis demonstrated strength of knowledge exploration
situated in a social context of shared practice as a means
for providing ICTs codifying knowledge. The paper also
demonstrated that in the exploration of the conditions for
providing ICT support for knowledge work, the cognitive
and the community model for knowledge management
based on organisational theories (Swan et al., 1999a) can
greatly inform a discourse. However, the lack of proper
attention to aspects of ICTs and to ICT development in
the framework can be alleviated by explicitly
distinguishing between innovation process and product. In
this paper we further suggested that tools for individuals
or for automation was the innovation product best
matching the cognitive model. Innovation products within
the community model can best be characterised as ICTs for
collaboration. Separating the innovation process and
innovation product, we argued that the choice of
knowledge management model for the innovation process
need not determine the type of innovation product
resulting from this process, in terms of individual or
collective support. Engaging in a collaborative knowledge
exploration process may indeed be an essential precursor
for a cognitive knowledge exploitation process if the
purpose is to provide ICTs automating aspects of the work
Faced with difficult design decisions regarding the
support for managing knowledge, simple contingencies
and classifications of possible solutions will not suffice.
The decisions regarding development of ICT support for
knowledge management based on the classification and
codification of knowledge cannot be exclusively analysed
from one perspective on knowledge management.
Competing perspectives can inform us on different aspects
of a complex phenomenon. In an attempt to balance
concerns for social and technical issues, ICT can serve a
role as support technology for the management of
knowledge, but firm conclusions must be based on careful
consideration of issues of organising and issues related to:
(1) the emergent properties of different technologies; (2)
the allocation of functionality between humans and
systems and; (3) the design of both manual and computer-
based technologies. The aim of this paper has been to
further some of the work on the management of knowledge
within organisational theory concerning the classification
of knowledge and analyse the findings based on view
towards developing ICTs. However, more research is
needed on the interrelations between discourses on
knowledge management and ICT development.
Thanks to all the people at Volvo Aero and Foss
Electric, as well as to Peter Carstensen and Henrik
Borstrøm for participating in the original research project
for Case B. Also thanks to Maxine Robertson for
comments and for proof-reading the manuscript. This
research was partly funded by the European Union Area 2
Project Laboratorium for Interaction Technology at
Trollhättan Uddevalla University, Sweden.
Alavi, M. (1999): Knowledge Management and
Knowledge Management Systems. Journal of AIS, vol.
1, no. 1.
Bannon, L. (1993): CSCW: An Initial Exploration.
Scandinavian Journal of Information Systems, vol. 5,
pp. 3-24.
Blackler, F. (1995): Knowledge, knowledge work and
organizations: An overview and interpretation.
Organization Studies, vol. 16, pp. 1021-1046.
Boland J r., R. J . & R. V. Tenkasi (1995): Perspective
Making and Perspective Taking in Communities of
Knowing. Organization Science, vol. 6, no. 4, pp. 350-
Bowker, G. & S. L. Star (1991): Situations vs. Standards
in Long-Term, Wide-Scale Decision-Making: The Case
of the International Classification of Diseases. In
Proceedings of the Twenty-Fourth Annual Hawaii
International Conference on System Sciences, Kauai,
Hawaii, January 7-11, 1991, ed. J . F. Nunamaker, J r.
and R. H. Sprague, J r. IEEE Computer Society Press,
vol. IV.
Bowker, G. & S. L. Star (1999): Sorting Things Out:
Classification and Its Consequences. Cambridge,
Massachusetts: MIT Press.
Braa, K. & R. Vidgen (1999): Interpretation, intervention
and reduction in the organizational laboratory: a
framework for in-context information systems research.
Accounting, Management and Information
Technologies, no. 9, pp. 25-47.
Brown, J . S. & P. Duguid (1991): Organizational learning
and communities-of-practice: Towards a unified view of
working, learning and innovation. Organization
Science, vol. 2, pp. 40-57.
Bucciarelli, L. L. (1984): Reflective practice in
engineering design. Design Studies, vol. 5, no. 3, pp.
Carstensen, P. & C. Sørensen (1994): The Foss Electric
Study — Some Methodological Issues. In CSCW '94
Workshop on Etchnographic Research and Design of
CSCW Systems, Chappel Hill, North Carolina, ed. J .
Hughes and K. Schmidt.
Carstensen, P. & C. Sørensen (1996): From the Social to
the Systematic: Mechanisms Supporting Coordination
in Design. Computer Supported Cooperative Work:
Journal of Collaborative Computing, vol. 5, no. 4,
December, pp. 387-413.
Cash, J . I. & P. R. Lawrence, ed. (1989): The Information
Systems research Challenge: Qualitative Research
Methods, vol. 1. Boston Massachusetts: Harward
Business School Research Colloquium Harward
Business School.
Ciborra, C. U. & R. Andreu (2000): Knowledge Across
Boundaries: Managing Knowledge in Distributed
Special issue on Knowledge Management of Journal of Information Technology, 16(2):83-97, 2001.
Organizations. Technical Report. Department of
Information Systems, The London School of Economics
and Political Science. November. ISSN: 1472-9601,
issue 93.
Cook, S. D. N. & J . S. Brown (1999): Bridging
Epistemologies: The Generative Dance Between
Organizational Knowledge and Organizational Knowing.
Organization Science, vol. 10, no. 4, pp. 381-400.
Drucker, P. (1993): Post-capitalist society. Oxford:
Butterworth-Heinemann Ltd.
Grant, R. (1996): Towards a knowledge based theory of
the firm. Strategic Management Journal, vol. 17, no.
Winter, pp. 109-122.
Heath, C., H. Knoblauch, & P. Luff (2000): Technology
and social interaction: the emergence of 'workplace
studies'. British Journal of Sociology, vol. 51, no. 2,
pp. 299-320.
Infopedia (1996): The ultimate multimedia encyclopedia
and reference library.
Kakihara, M. & C. Sørensen (Forthcoming): Exploring
Knowledge Emergence. Department of Information
Systems, The London School of Economics and
Political Science.
Mathiassen, L. (1998): Reflective Systems Development.
Scandinavian Journal of Information Systems, vol. 10,
no. 1+2.
McCracken, G. (1988): The Long Interview. London,
United Kingdom: SAGE Publications.
Mockler, R. J . (1992): Developing knowledge-based
systems using an expert system shell. New York:
Macmillan Publishing Company.
Nonaka, I. (1994): A dynamic theory of organizational
knowledge creation. Organization Science, vol. 5, no.
1, pp. 14-37.
Nonaka, I. & N. Konno (1998): The Concept of "Ba":
Building a foundation for Knowledge Creation.
California Management Review, vol. 40, no. 3, pp. 40-
Nonaka, I. & H. Takeuchi (1995): The knowledge-creating
company. How Japanese companies create the dynamics
of innovation. New York: Oxford University Press.
Patton, M. Q. (1980): Qualitative Evaluation Methods.
USA: Sage Publications.
Prusak, L. (1997): Knowledge in Organizations. Oxford:
Robertson, M., C. Sørensen, & J . Swan (2000):
Managing Knowledge With Groupware: A Case Study
of a Knowledge-Intensive Firm. In Proceedings of the
33rd Hawaii International Conference on System
Sciences (HICSS-33), Maui, Hawaii, ed. R. H. Sprague
J r.
Russel, S. & P. Norwig (1995): Artificial Intelligence - A
Moderne Approach. Prentice Hall.
Scarbrough, H. (1995): Blackboxes, Hostages and
Prisoners. Organization Studies, vol. 16, no. 6, pp.
Scarbrough, H., J . Swan, & J . Preston (1999): Knowledge
Management: A Literature Review. Issues in People
Management. London: Institute of Personnel and
Schmidt, K. (1999): Of maps and scripts: The status of
formal constructs in cooperative work. Journal of
Information and Software Technology, vol. 41, pp.
Schmidt, K. & L. Bannon (1992): Taking CSCW
Seriously: Supporting Articulation Work. CSCW, vol.
1, no. 1-2, pp. 7-40.
Schmidt, K. & C. Simone (1996): Coordination
mechanisms: An approach to CSCW systems design.
Computer Supported Cooperative Work: An
International journal, vol. 5, no. 2-3, pp. 155-200.
Snis, U. (1997): Kundskabsutveckling med stöd av
expertsystem (Expert systems support for knowledge
creation). MPhil Thesis, Göteborg University.
Sørensen, C. (1994): The Product Classification Scheme.
In Social Mechanisms of Interaction, ed. K. Schmidt.
Lancaster, England: Esprit BRA 6225 COMIC, pp.
247–256. (COMIC Deliverable 3.2).
Sørensen, C. (1999): Interaction in Action: Learning from
studying the use of technology. In Informatics in the
Next Millenium, ed. F. Ljungberg. Lund:
Studentliteratur, pp. 117-135.
Spender, J . (1996): Organizational knowledge, learning
and memory: three concepts in search of a theory.
Journal of Organizational Change, vol. 9, no. 1, pp.
Spender, J . (1998): Pluralist epistmology and the
knowledge-based theory of the firm. Organization, vol.
5, no. 2, pp. 233-256.
Suchman, L. (1994): Do categories have politics? The
language/action perspective reconsidered. Computer
Supported Cooperative Work. An international journal,
vol. 2, no. 3, pp. 177-191.
Swan, J . & S. Newell (2000): Linking Knowledge
Management and Innovation. In Proceedings of the 8th
European Conference on Information Systems, Vienna,
Austria, ed. H. R. Hansen, M. Bichler, and H. Mahrer.
Wirtschaftsuniversität Wien, vol. 1, pp. 591-598.
Swan, J ., S. Newell, & M. Robertson (2000): Limits of
IT-Driven Knowledge Management Initiatives for
Interactive Innovation Processes: Towards a
Community-Based Approach. In Proceedings of the
33rd Hawaii International Conference on System
Sciences (HICSS-33), Maui, Hawaii, ed. R. H. Sprague
J r. IEEE.
Swan, J ., S. Newell, H. Scarbrough, & D. Hislop
(1999a): Knowledge management and innovation:
networks and networking. Journal of Knowledge
Management, vol. 3, no. 3, pp. 262-275.
Swan, J ., H. Scarbrough, & J . Preston (1999b):
Knowledge Management- The Next Fad to Forget
People? In 7th European Conference on Information
Systems, Copenhagen, Denmark, ed. J . Pries-Heje, C.
Ciborra, K. Kautz, J . Valor, E. Kristiansee, D. Avison,
and C. Heje. Copenhagen Business School, vol. II, pp.
Tsoukas, H. (1996): The Firm as a Distributed Knowledge
System: A Constructionist Approach. Strategic
Management Journal, vol. 17, no. Winter, pp. 11-25.
Turban, E. (1995): Decision Support Systems and Expert
Systems. Prentice Hall.
Wegner, P. (1997): Why Interaction is More Powerful
Than Algorithms. Communications of the ACM, vol.
40, no. 5, pp. 80-91.
Winograd, T. (1994): Categories, disciplines, and social
coordination. Computer-supported cooperative work: An
international journal, vol. 2, no. 3, pp. 191-197.