March 2007: Managing and representing knowledge


Knowledge and definitions of what knowledge is has been the domain of philosophers and leading thinkers for millennia. The philosophical approach, while useful, is difficult to implement from an engineering point of view. To define and use knowledge effectively requires more explicit definitions and the scientific literature over the past decade has provided the basic tools to do this.

... knowledge defines ... the ability of a company to achieve its business goals

Knowledge is the most important aspect of an engineering company - it is a primary asset. It defines the ability of that company to achieve certain tasks, which define its business goals. For example this may be to build: bridges, ships or aeroplanes. The importance of this knowledge cannot be overestimated, indeed there are many anecdotal examples of companies, no doubt driven by financial issues, driving out valuable experience in the form of redundancies and thus inflicting a form of capability-amnesia upon themselves.

Retention of knowledge by companies is often an issue where long lead times are involved. Certain LMTO projects can involve lead times of many decades. In cases such as these retention of company knowledge is vital.

A view of information and knowledge

To apply knowledge concepts within an engineering environment, there is a need to define reliable representations of data, information and knowledge. This should be in a form which can be readily adapted and used within the ordinary everyday computational systems we all use. This requires the answers to basic questions such as:

  • what is knowledge? and
  • what is information?

These questions when asked at the level of human knowledge and information are philosophical in nature. So while of interest to intellectual thinkers they are of little use in a computer information domain. This is because application of philosophical approaches are beyond current computational domain capabilities. In short a generalised definition which covers everything is presently impossible (Aamodt 1995, Stenmark 2001).

If we cannot describe a general definition we can consider a limited subset of one. One potential perspective is that of a computational information processing system. This covers the domain we are primarily interested in and focuses on a representation of data, information and knowledge appropriate to the types of modelled system information engineers use (Brézillon 2001).

A consequence of this approach is that we have to operate within a specialised context, which allows a limited range of cover, but remains coherent within the more general concepts of knowledge and information described previously.

This view of a framework of such concepts (Aamodt 1995) is inspired by influential work in cognitive sciences. Although no formal agreement has been reached in the literature, the distinction between data and information is well established within the fields of database and information systems.

Polymorphic concepts

It is likely that no formal agreement can ever be reached since these concepts are polymorphic in nature (Aamodt 1995). A polymorphic concept is one which cannot be defined by a classical definition. For example, terms such as behaviour and functionality can mean entirely different things depending upon which group of people are viewing them. Alternatively, mathematical concepts are non-polymorphic, since they are normally very well defined and require a strictly formal definition. The terms we are interested in therefore have very different definitions depending on the context of interpretation.

The only way to get a meaningful definition of a polymorphic concept is to understand it within a particular context. This relates directly to the purpose, functionality or some intended use as seen from a particular users perspective. This highlights the reasons why the definitions of information say, varies quite considerably depending on who is using it. The definitions described in this domain for use with knowledge management are therefore highly context driven.

Data, information and knowledge

Although there are many variations in the definitions of data, information and knowledge, the following are offered as reasonably acceptable to the literature at large. These are that:

Data are syntactic entities
i.e. patterns with inherent meaning (but without context). These are inputs to an interpretation process (the initial step of decision making).
Information is interpreted data
i.e. information is data with meaning or context. This is the input to (and the output from) the knowledge based process of decision making.
Knowledge is learned information
i.e. knowledge is information incorporated into an agent’s reasoning resources, and made ready for use within a decision making process. This is the output of a learning process..

The transformation processes between these states provides the most interesting aspects of this, since it involves interpretation and elaboration. The only reliable agent capable of this is knowledge itself.

This leads us to a key view that knowledge is central to these ideas and involves a bootstrapping operation where knowledge is used to create further knowledge. Figure 1 shows an extended view (click to enlarge) of the role of knowledge highlighting its central role.

Image of: Fig 1. Transformation of data into knowledge

Fig 1. Transformation of data into knowledge

Therefore the role of knowledge is to play the active part in the process of transforming data into information, deriving other information and acquiring new knowledge (i.e. through learning). The role of knowledge in these processes may be summarised as the:

  • transformation of data into information (interpretation),
  • derivation of new information from existing information, i.e. elaboration, and
  • acquisition of new knowledge, also known as learning.

Conceptualised representation

If knowledge management is to be formally represented it has to be based on a conceptualisation (Gruber 1993). This conceptualisation takes the form of an abstract view of the objects, concepts and any other entities that exist within the domain that is being designed or represented. This consists of a formalised description (an ontology) and an agent capable of reasoning and performing tasks based on the ontology (a Problem Solving Method).


An ontology associated with knowledge management (as opposed to its normal philosophical connotations) is an explicit specification of this conceptualisation. The purpose of an ontology is to capture domain knowledge in a generic way and provide a commonly agreed understanding of a domain (Perez 1999), which may be reused and shared across a number of applications and groups (Chandrasekaran 1999). In practice this generally means that an ontology is a hierarchically structured set of terms for describing a domain that can in turn be used as a skeletal foundation for a knowledge base. (Swartout 1996)

Problem Solving Methods

The term Problems Solving Methods (PSM’s) describe the reasoning process of a knowledge based system (KBS) (Perez 1999). In general a PSM defines and encapsulates a way of how to achieve the goal of a task. The actual form of a PSM may differ quite considerably between applications, however in general a PSM describes which reasoning steps have to be performed, and which type of domain knowledge is needed to perform a task (Fensel 1996).

PSM’s are normally in the form of a domain independent implementation, which specifies how data flows between its subtasks. PSM’s are also described at the knowledge level (Fensel 1996), where a knowledge level implementation shows reasoning in terms of goals to be achieved, elementary reasoning operators and knowledge needed to achieve objectives. This description is used (Benjamins 1996) to show that PSM’s can be used to efficiently achieve goals of tasks through the application of domain knowledge.

Essential elements

Ontologies and PSM’s are therefore required elements in the implementation of knowledge management in computer applications. There are a large number of ideas about ontologies and PSM’s in the literature. These also include the issues surrounding PSM’s and include proposed architectures for the implementation of these concepts. Since PSM’s are intimately involved in reuse of ideas, the issues about libraries of PSM's, their structure, organisation and indexing this will form a future topic for this research highlight.


The importance of knowledge management cannot be overestimated and should be of fundamental importance to all industries. This brief treatment of the basics of knowledge management serves well as a basis for further investigation. The ability to store knowledge in a computational form is becoming a reality and is swiftly becoming a essential means of storing business ability in a rapidly changing and complex engineering environment.

Author: John Dalton



  1. Aamodt, A.; Nygard, M., "Different roles and mutual dependencies of data, information, and knowledge - an AI perspective on their integration", Data and Knowledge Engineering, vol 16, 1995, pp 191-222
  2. Benjamins, V.R.; Fensel, D.; Straatman, R., "Assumptions of problem-solving methods and their role in knowledge engineering", ECAI 96. 12th European Conference on Artificial Intelligence. 1996.
  3. Brézillon, P.; Pomerol, J-Ch., "Is context a kind of collective tacit knowledge", European CSCW 2001 Workshop on managing tacit knowledge, Bonn, Germany, 2001.
  4. Chandrasekaran, B.; Josephson, J.R.; Benjamins, V.R., "What are ontologies, and why do we need them?", IEEE Intelligent Systems, (Vol 14, No.1) pp 20-26, January/February 1999.
  5. Fensel, D.; Straatman, R.; Harmelen, F., "The mincer metaphor for problem-solving methods: making assumptions for reason of efficiency", in Proceedings of the Knowledge Engineering Methods and Languages Workshop (KEML 96), Paris – Orsay (January 15-16 1996).
  6. Gruber, R.G., "A translation approach to portable ontology specifications", Knowledge Systems Laboratory, Technical Report, KSL 92-71, 1993.
  7. Pérez, A.G.; Benjamins, V.R., "Overview of knowledge sharing and reuse components: ontologies and problem-solving methods", Proceedings of the IJCAI-99 workshop on ontologies and problem-solving methods (KRR5) Stockholm, Sweden, August 2,1999.
  8. Stenmark. D., "The relationship between information and knowledge", Proceedings of IRIS 24, Ulvik, Norway, August 11-14, 2001
  9. Swartout, B.; Patil, R.; Knight, K.; Russ, T., "Toward distributed use of large-scale ontologies", Proceedings of the 10th Knowledge Acquisition for Knowledge Based Systems Workshop (KAW 96), Gaines B and Musen M, Eds, (Nov 9-14, Banff, Alberta, Canada) 1996.