QUALITY MANAGEMENT, FRAMEWORK OF KNOWLEDGE CAPITALIZATION AT RESEARCH ORGANIZATIONS Astrid JAIME*, Mickaël GARDONI*, Joël MOSCA*, Dominique VINCK** * Laboratory GILCO, National Polytechnic Institute of Grenoble ** Laboratory CRISTO, University Pierre Mendès France Abstract: Innovation and thus the production of knowledge becomes a factor of competitiveness. The research activities aim, in particular, the production of new knowledge. Researchers could profit from approches and methods such as quality management. In this communication, we study the approaches used while implementing a quality management system within various research organizations. Our assumption is that quality management could be complemented by knowledge management improve the process of the research activities. For this reason, we propose an approach integrating quality management and taking into account elements of knowledge management to support the scientific activity. Key words: Quality management, knowledge management, research activities, knowledge capitalization.

Introduction The research activity implies managing information and knowledge. From these resources, new knowledge is produced, to become, itself, the resource of new researches. In France, reflections on the possibility of applying the quality concepts and methodologies to the research process were carried out and document ed in "Experimental Guide for Quality in Research" written by the French Work Group "Quality in Research", (1997). This document was taken by the AFNOR1 (2001) as the base of the Documentation Booklet FD X 50 - 550 “Quality Management in Research - General principles and recommendations", in which quality management is proposed as a possibility to face the multiple issues of the research activities. Moreover, during the last years, some research organizations have been interested in quality management as a mean to improve their activities. Indeed, the research activity requires rigour in the knowledge production process. However, it presents specificities in terms of goals, resources, practices and organization which make this activity very different from the industrial activity, where quality management has been traditionally used. Thus, the introduction of quality management, step by step, as used by the industry is not possible in the scientific environment. One AFNOR document recommends the utilisation of quality management by the scientific actors. In fact, research organizations do set up quality management systems within their organizations. Therefore, we have started a research process that aims at clarifying the problems that the research organizations face when implementing a quality management system. We wonder in particular about the role that quality management 1

French Standardization Association

could play in the transmission of knowledge. Our objective is concretely to check the hypothesis according to which quality management can be used to support the knowledge production process, by providing methods that improve the total performances of the research processes. In the first part, we describe the observed functioning of a few research units trying to implement a quality management system. In the second part, we will look into some knowledge management aspects at these research units. Finally, we will present some proposals outlining certain elements aiming at facilitating the introduction of the knowledge management at research organizations, while implanting a quality management system.

Quality management in the Context of Research The AFNOR (2001) propose, to the research organizations wishing to engage in quality management, to establish practices to maintain the quality of the research activity during all the scientific production process. It leaves the precise definition of these practices to each research organization. For that reason, we have studied the various approaches used by some research entities setting up a quality management system. We paid special attention to the management of information (data, documents, etc), considering the support it offers to the capitalization of knowledge. This work has been completed in several phases: - Collection of information (direct observations and interviews) on the current operation of a research laboratory, over a period of four months. - Interviews with the people responsible for quality management in seven research laboratories where formal efforts of introduction of this methodology are being done. - Follow- up of the implementation process of quality management at a research laboratory. This work is undertaken via the participation in work meetings 2. Hereafter, we will present the results of this field work. The Reality Observed at a Laboratory The observation has allowed us to note that several characteristics of the research activity make difficult its manageme nt: the diversity of activities within a laboratory, the great quantity of records (digital reports and files in particular) to manage, the freedom granted to the researchers for the register or the traceability of their production, the multiplicity of working methods, the great turn-over of researchers, the multiplicity of activities that must be developed in parallel, with various time delays, and that should be coordinated to lead to valid results, the difficulty to establish, from the beginning of a project, the precise characteristics of the product of research (which could be a physical product or a conceptual product), etc. which explains the interest of having support practices during the research process, of capitalizing the history of a project, of setting up procedures for the validation of results, etc.

2

This work is still in process. Consequently, the information collected until now is not included in this communication.

All these characteristics make difficult the knowledge management and the definition of a standardized instrumentation. Therefore, the question that arises is how to capitalize the knowledge produced, and how to rationalize and instrument the activity. Some Experiences of Implementation of Quality Management We have carried out several interviews at seven research entities attached to the CNRS 3 , already engaged in quality management. These organizations combine activities of industrial research and basic research, except a unit that works for the research laboratories as supplier of special equipments required in research projects. However, we noted that nowadays, quality management is primarily applied to the administrative and/or the technical activities and very little to the basic research activities. The implementations observed are inspired on the ISO 9001 standard (AFNOR, 2000) and the result is often the establishment of information systems, which aim at facilitating the realization of the repetitive processes. We have observed that the type of activity carried out causes divergences in the way quality systems are established. Thus, two of the organizations followed a traditional process fo r the establishment of a quality system according to the standard ISO 9001. Though, those working mainly in basic research were challenged by the way in which quality management could be applied to research. The field work carried out has been of key importance to support our research. For this reason, we will continue this observation and interviewing work all along our project. We will now relate our observations to some major elements of the theory of knowledge management.

Knowledge management at Research organizations Our interest is centered on quality management at research organizations as establishments devoted to the production of knowledge. This activity, according to the results of researches in sociology of sciences and our own observations, is usually developed in the form of more or less structured research projects 4 . Moreover, knowing that the constitution of a phenomenon, the establishment of a fact or of a statement is closely associated to a history of contingencies of the research process, the documentation issue becomes a concern for the researchers themselves when they try to reconstruct a former stage or to take in hand a project engaged by a colleague. That is why it is important to study the issues of knowledge management while being based on the management of information and to try to find ways of improvement. Within this framework, we will present, hereafter, some elements that show the way in which knowledge management is present at the surveyed research organizations. 3

National Centre of Scientific Research of France (Vinck, 1995) wrote: "the activities at the laboratory are structured in projects. The project is a sequential unit of which the completion is the writing of a research report or of a publication. The project sees to be the unity of organisation that allows affecting tasks to the members of the laboratory, to order supplies, to prepare equipment, to propose phenomena to be studied and to orient the library searches… There is a discrepancy between formal descriptions of the procedures, methods and work schemes and the effective realization of the activities. […] The methods and research protocols do not account for the effective sequence of the activities." (p. 154). 4

The current knowledge management practices at research organizations Grundstein (1995) says that there is a “logic of capitalization that proceeds according to two axes: - an axis oriented to the management of knowledge (management of technical data, document management, management of configurations); - an axis oriented to the formalization of know-how (acquisition/representation of the fields of knowledge and of the reasoning relating to this knowledge).” The quality management implementation at the research organizations observed until this day is oriented towards the first axis (management of knowledge). It has started with an objective of taking into account the organisational aspects, mainly through the writing out of documents (operational procedures and documents). For the management of these documents, this methodology has been translated into information processing systems, often an Intranet that sometimes manages other documents of the organism. This verifies the situation described by Gandon (Gandon et al., 2002) about the use of Intranets and the Web as means to manage documentation. However, the information processing systems that we have observed do not seek “the management and the circulation of distributed knowledge” as projects like CoMMA (Gandon et al., 2002) aim. For the surveyed laboratories, the approach selected is to facilitate the realization of the activities by providing a tool that makes it possible to find documents or information and to organize those produced. In general, the documents resulting from the research process are not managed by these systems. Moreover, the second axis (the formalization of know- how) has not been yet addressed for the research activity. Within this framework, we believe that there is an important place for the utilisation of knowledge capitalization methodologies. Why quality management without knowledge management? We have seen that the quality management implementation at the surveyed organizations does not directly address the management of the knowledge resulting from the research activity. The origin of this situation seems to be double: On one hand, there exists a known methodology to address documentary organizational management. In other words, traditionally, quality management has been used for this kind of management, which is easily adaptable to the organizational aspects of laboratories. Conversely, a defined methodology to make knowledge management in scientific organizations does not exist to apply quality management to scientific research activities. On the other hand, the results obtained by document management are easily perceived, in the short run, by the personnel, who justify the utilization of the developed systems. Nevertheless, it is observed that the lack of real experiences of implementation of quality management complemented by knowledge management is a factor that affects the eventual implementations. There is an effort of formalization and capitalization of the organizational memory of the support activities, which is not accompanied by a similar effort for the scientific activities. However, we believe that in research activities, there are elements that could be formalized and handled, keeping the flexibility of the research

process. The objective would be that the generated knowledge could be located, preserved, developed and brought up to date (Grundstein, 2000) for the benefit of the research activity. The difficulties of knowledge capitalization Mahé (2001) shows the "barriers that oppose that a knowledge created on a precedent project be re-used on a project in progress". These barriers are mainly the personnel turnover (either because people leave, or because the person who holds knowledge is not the one who is responsible for the project) and the lack of information (either because this one has not been produced or formalized, or because it has not been capitalized). At research organisms, these same difficulties exist and are increased partly because they relate to knowledge within the organism and also to knowledge held by the scientific network of the field to which the organism belongs. For that reason, we intend to use methodologies that would make possible to partially overcome these difficulties, by taking into account the characteristics of research.

Proposal of implementation of quality management at research organizations Our proposal is to use quality management to introduce the principles of knowledge management, making it possible to capitalize the knowledge produced when realizing research projects. Our objective is to seek the improvement of the knowledge production process. Thus, by inspiring ourselves on the recommendations given by the AFNOR (2001), we propose a representation of the method as follows (Figure 1): Pragmatism Pragmatism 1. Definition of the Objectif

Documentation of results

3. Valorization of Results

Documentation of objectives

2. Realization of the Research

Integration

Pedagogy Documentation of the evolution

Process Formalization

Knowledge

Knowledge Management

Figure 1. Implementation of quality management at research organizations

This diagram emphasizes the importance of documentation throughout research process and thus of its management to support the process of knowledge creation. The subjacent idea is that there is knowledge produced throughout the research process, so it could be profitable to exploit this richness. However, Wunram (Wunram et al., 2002) indicates that "the approaches that start with the goal of capturing all the knowledge of the employees are predetermined to fail" . It is thus necessary to define the knowledge that can be more beneficial to research activities. Which knowledge to capitalize? Given our intention to improve the knowledge production process, we carried out an analysis of the research projects activities. Moreover, since “knowledge is based on the data and the information” (Wunram et al., 2002), we analyzed the information used and generated during the realization of this kind of projects. Thus, we have been able to identify the most critical aspects in terms of unexploited possibilities of knowledge capitalization. With this aim, we have represented the research activities in the form of a matrix which crosses these two points of view: activities carried out, and information used and generated 5 . From this matrix, we extracted the aspects where the principles of knowledge management could be used to ease the realization of the research activity by improving the management of information. These aspects were selected by taking into account the practices currently used in research, where the activity leans very strongly on the capitalization of the final results. Though, at least for the observed laboratories, it is common to grant great freedom to the researchers for the management of the recordings resulting from the realization of the research projects. Thus, the result of this study is that there is a very important potential of capitalization of the knowledge produced during the realization of research projects. By retaking the representation of quality management (Figure 1), we note that knowledge, resulting from research projects, is already capitalized thanks to the existing valorization mechanisms existing in scientific research. However, a great amount of the knowledge produced during the research process remains barely capitalized. In this context, the concept of artefact seems useful to us. Michaux and Rowe (2003) present the position of Groleau6 related to this concept, by saying that he “defines the artefacts as elements having a durable material form containing knowledge. Two elements seem important to retain in the design of artefacts. On one hand, distributed cognition considers that artefacts contain a part of the knowledge necessary to conclude a daily action with effectiveness: the other part being held in a complementary way by men... On the other hand, the intervention mode of these artefacts is the representation that they are able to convey. Indeed, the artefacts are often similar to objects (speed chart, paper-board, indicator on a data-processing screen or a measuring apparatus…).”

5

The matrix mentioned is not included in this paper, because its size makes it impossible to show it in a legible way. 6

Groleau C. (2002) “Structuration, Situated Action and Distributed Cognition: Rethinking the Computerization of Organizations”, Systèmes d’Information et Management, Vol. 7, n° 2, pp. 13-36. In Michaux V., Rowe F. (2003) Grille d’analyse des situations de coordination médiatisée par un système d’information informatisé In Proceedings of the Conference CITE’2003 Troyes, France 3 – 4 December, 2003.

In the research context, the daily action, which we are interested in is related to research projects. Thus, we observed that within the realization of research projects there is a great quantity of artefacts produced. Given that those convey knowledge, we will focus on artefacts capitalization, as a mean of capitalizing at least part of the knowledge resulting from the realization of research projects. We were thus interested in the methods of capitalization of project memory present in the literature. We have identified many works regarding the knowledge related to decision- making, in particular, Bekhti and Matta (2003). However, for the capitalization of the artefacts produced when realizing research projects, we have noted that in spite of the existence of several methods for the capitalization of artefacts, those are not adapted to the characteristics of research projects, specially because of the dynamic environment and the non repetitiveness of the project. Thus, we focus our efforts on the capitalization of artefacts, while following the principle of pragmatism recommended by the (AFNOR, 2001) and by Weber (Weber et al., 2002). The latter indicates us that "a strong argument for a pragmatic knowledge management is the fact that anyway we cannot control all knowledge". Then, we want to find ways to capitalize these artefacts as a mean to facilitate the realization of other research projects. The question is how to do it. Within this framework, it is appropriate to remember what Davenport and Prusak (1998) 7 say about information technology (IT). For them, the main role of IT in knowledge management is “to accelerate the speed of knowledge transfer”. That is why, after studying the possibilities offered by informatics, we have defined the main functions we would need to support the management of artefacts resulting from the projects: - Constitution of families of artefacts, according to the work in progress. - Possibility of making annotations on the artefacts or on certain aspects of the artefacts. - Automatic classification of artefacts (with the option of doing it manually) according to several criteria: fields and sub- fields, application field, type of artefact, etc. - Dynamic construction of shared ontologies for each organism - Search engine of the existing artefacts. - Automatic suggestion of the existence of artefacts having a relationship with the work in progress. - Graphic visualization of the projects according to produced artefacts. - Progressive enrichment of a library of concepts, intermediate results and of other resources. - Traceability of the decisions taken. For that reason, we were interested in the existing knowledge management software tools to study their adequacy to the identified needs. The existing software tools for knowledge management Baroni de Carvalho R and Araújo Tavares M. (2002) have defined knowledge management (KM) software as “a kind of software that supports any of the three basic 7

In: Baroni de Carvalho R and Araújo Tavares M. (2002) Using information technology to support knowledge conversion processes. Information Research, Vol. 7 No. 2, January 2002.

KM processes (Davenport & Prusak, 1998): generation, codification and transfer.” That is why we have tried to identify the most important knowledge management software tools currently available on the market in order to analyze the ir functionalities and their capacity to facilitate the capitalization of the artefacts or intermediate results issued from research projects. 8 For doing so, we used an automated search and web intelligence solution9 during a two month period (July – August, 2003). This allowed us to identify 53 enterprises, offering 224 KM tools, which we classified according to their main functionalities. The result of this classification is presented as follows: first the type of functionality regrouping some tools, then the number of tools found under this functionality and finally, some examples of tools available on the market. - Document Management : (38) iManage WorkDocs™, Hummingbird DM - Document Routing: (4) iManage WorkRoute, ISYS:rdu - Text Base: (2) Inmagic’s DB/TextWorks - Linkage of documents by hypertext: (2) Tikit Document Link - Content Management : (17) IBM® DB2® Content - Collaboration, Groupware: (29) Hummingbird CollaborationTM - Business Process Management, Workflow: (12) IXOS-eCONprocess, Staffware Process Suite, AppianWorkflow - Graphical Visualization, Knowledge Maps Systems: (5) IBM’s DB2 Intelligent Miner Visualization, Lotus Discovery Server, Inxight’s Star Tree™ SDK - Content Aggregator, Portals: (27) IBM’s DB2® Information Integrator for Content, Computer Associates’ CleverPath Portal - Project Management: (6) Accelrys’ DS ProjectKM, Kinematik’s eNovator. - Knowledge Agents: (8) Autonomy’s Active Knowledge™ - Data Integration tools: (5) Hummingbird ETL, Newgen’s OmniExtract - E-Learning: (6) LongView’s LRALTM, Eedo’s ForceTen. - Search Engines: (28) Nexidia’s NEXminer, Open Text’s BRS/Search™ - Others: (35) Open Text’s Livelink Review Manager for Acrobat, Serviceware’s Cognitive Processor This classification shows that the offer is very rich and varied. We where even forced to allow a category for all the tools that did not fit into the others or that responded to several functionalities. Therefore, our conclusion is that the data-processing developments offer many possibilities for knowledge management at research organizations. Nevertheless, we are inclined to think that it still lacks a tool adapted to the basic research activity, focused on the capitalization of the intermediate results (the artefacts). In the future, we want to define the specifications of such a tool, by re-using the possibilities of the existing tools, to provide to researchers a support for their activities.

Conclusions and prospects Our research starts with the reflections carried out in France by organisms that recommend quality management as a mean to support the research process. For that reason, we wanted 8

This information is not included in this paper, because the size of the table that contains it makes it impossible to show it in a legible way. 9 The solution used was Google A lert.

to study the situation of the research organizations when they implement a quality management system. To this end, we completed a field work in order to know their reality and to collect information on the difficulties encountered when implementing this kind of system. The first step of this work allowed us to observe several characteristics of the research activities. They imply major differences between this type of activities and the industrial activities where the quality management has been traditionally used. Moreover, we have noted that the fundamental problems of the surveyed organizations are the improvement of the process of their activities. The quality management of the organizations we observed do not directly address the research activities. Indeed, the field work allowed us to note that even if the main activity of the analyzed organizations is the production of knowledge, the systems are centered on the formalization of certain activities, mainly the activities that support the research activity, and on the management of part of the documents, primarily the final documents (and not the intermediate results or the artefacts). On the other hand, we have exemplified that the quality management systems observed do not really address the knowledge management aspect in research, and that document management does not address the artefacts of research. The cause seems to be the lack of methodologies and of real experiences that would formulate a way to set up a quality management system focused on knowledge management. To address this last aspect, we carried out an analysis of the activities realized during the research process and of the information used and generated by these activities. This analysis enabled us to define that a very important aspect for the capitalization of knowledge resulting from research projects is the capitalization of intermediate results. Then, to allow this capitalization, we analyzed the existing knowledge management tools and we assessed the current richness of the offer: In spite of the great possibilities for knowledge management, there is a lack of tools that could facilitate the capitalization of intermediate results issued from research projects. In fact, we noted that the tools address mainly the management of the container and not the management of the contents. For that reason, we consider the use of shared ontologies to represent the concepts and to support the capitalization of knowledge, particularly those created within the interior of the organism. Once the development of such a tool is finished, we will have to test it to analyze its capacity to effectively improve the knowledge capitalization. Moreover, it is necessary to study the way in which quality management could be used to introduce the knowledge management principles and how the tool could support these approaches. We think that quality management could be established at research organizations in a more beneficial way by contributing to the improvement of their knowledge management. The next phase of our research is thus the definition of the specifications of such a tool.

Bibliographical references AFNOR (2000) NF EN ISO 9001 – 2000 « Système de management de la qualité – Exigences ». AFNOR. Paris. AFNOR (2001) Fascicule de Documentation FD X 50 – 550 « Démarche qualité en recherche – Principes généraux et recommandations ». AFNOR, Paris.

Bekhti S., Matta N. (2003) Project Memory: An approach of modelling and reusing the context and the design rationale. In Proceedings of the Knowledge Management and Organizational Memories Workshop. Mexico, August 11, 2003. Eighteenth International Joint Conference on Artificial Intelligence. Baroni de Carvalho R, Araújo Tavares M. (2002) Using information technology to support knowledge conversion processes. Information Research, Vol. 7 No. 2, January 2002. Gandon F., Dieng R., Corby O., Giboin A. (2002) Web Sémantique et Approche MultiAgents pour la Gestion d’une Mémoire Organisationnelle Distribuée. In Proceedings of the IC’2002 Conference, Rouen, 28 – 30 mai 2002. 13ème journées francophones d’Ingénierie des Connaissances. p. 15 – 26. Rouen Groleau C. (2002) “Structuration, Situated Action and Distributed Cognition: Rethinking the Computerization of Organizations”, Systèmes d’Information et Management, Vol. 7, n° 2, pp. 13-36. In Michaux V., Rowe F. (2003) Grille d’analyse des situations de coordination médiatisée par un système d’information informatisé In Proceedings of the Conference CITE’2003 Troyes, France 3 – 4 December, 2003. Groupe de Travail Français « Qualité en Recherche » (1997) Guide expérimental pour la qualité en recherche. http://www-dsm.cea.fr/Qualite/MENRT/page.html. France. Grundstein M. (2000) Le Management des Connaissances dans l’Entreprise : Problématique, Axe de progrès, Orientations, Research report #050010, , MG Conseil. Jaime A., Gardoni M., Vinck D. (2003) Capitalizing Knowledge at Research Organizations through Quality Management, Proceedings of the Third European Knowledge Management Summer School: Knowledge Management in Action. San Sebastian, Spain, September 7 – 12. Jaime A., Gardoni M., Vinck D. (2004) La Démarche Qualité, Cadre de la Capitalisation des Connaissances dans les Organismes de Recherche. Proceedings of the IPI Conference. Autrans, France. 22 – 23 Janvier. Mahé S. (2001) PUMEO : un modèle actif pour la gestion des connaissances tacites et explicites dans l’entreprise. Identification et intégration des connaissances dans les tâches de tous les jours. In J. CHARLET Proceedings of the Conference IC 2001. 25-27 juin 2001, p. 21 – 39. Leibniz – IMAG. Paris : PUG. Michaux V., Rowe F. (2003) Grille d’analyse des situations de coordination médiatisée par un système d’information informatisé In Proceedings of the CITE’2003 Conference Troyes, 3 – 4 décembre, 2003. Vinck D (1995), Sociologie des Sciences, p. 154 – 155. Paris: Armand Colin Editeur. Weber F., Thoben K-D, Wunram M., Müller D. (2002) Practising KM in an RTD Organisation - Lessons from a Pragmatic and Successful Case. In: Successful Knowledge Driven Organisations. Proceedings of UNICOM Seminar, London, 28-29 May 2002. Weber, F.; Thoben, K.-D (2002). (Eds.). Proceedings of the 8th International Conference on Concurrent Enterprising - Ubiquitous Engineering in the Collaborative Economy, pp. 151 – 158, Rome, Italy, 17th - 19th June 2002. Wunram M., Weber F., Pawar K. S., Gupta A. (2002) Proposition of a Human-centred Solution Framework for KM in the Concurrent Enterprise. Published in: Pawar, K.