TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRYSven-Eric SchapkeInstitute of Applied Computer Science in Civil Engineering, TU Dresden, Germany Sven.Schapke@cib
.bau.tu-dresden.deKarsten MenzelInstitute of Applied Computer Science in Civil Engineering, TU Dresden, Germany Karsten.Menzel@cib.bau.tu-dresden.deRaimar J. SchererInstitute of Applied Computer Science in Civil Engineering, TU Dresden, Germany Raimar.Scherer@cib.bau.tu-dresden.deSven-Eric Schapke, Karsten Menzel, Raimar J. Scherer ABSTRACT: The predominant challenge of knowledge management in A/E/C is to overcome the inter- and intra-organisational barriers of knowledge exchange within project-centred and fragmented organisations. This paper discusses the necessary knowledge transformation processes in corporate and project organisations as well as the potentials for its support by information and communication technology. Critical aspects for supporting the identification and externalisation of knowledge in A/E/C are discussed. A framework for combining several knowledge management services towards an organisational memory system is proposed. A first prototype of a document modelling (DocMo) service integrated with operational document management and e-learning systems is presented. The service allows for evaluating various methodologies within the context of an organisational memory system implementing various methods for manual as well as for (semi-)automatic knowledge externalisation from text documents.1.INTRODUCTIONIt is widely accepted that the current market dynamics and the trends towards specialised and customer-oriented services in the AEC-industry demand a more efficient and effective application of knowledge within corporate as well as project organisations. In practice, the exchange of knowledge is first of all realised on a social level. Today, the processes of organisational learning in construction are mainly fostered by (a) ICT-supported collaboration, (b) modified project organisations and delivery systems, as well as (c) reorganised work processes. Especially, new concepts of work organisation such as competence centres, lean construction or agile production are expected to provide for continuous dissemination of knowledge and learning within work teams. However, an effective use of knowledge within a particular work process does not guarantee its exchange beyond project and company borders. Moreover, increasing organisational flexibility and concentration on core competencies can hinder the development of common values and a shared understanding throughout an overall (virtual) organisation.eSM@RT and CISEMIC 20021Sven-Eric Schapke, Karsten Menzel, Raimar J. SchererIn the AEC-industry, a generally project-centred and fragmented industry, one of the predominant challenges of KM is overcoming these inter- and intra-organisational barriers. This paper discusses various aspects of identifying, retrieving and disseminating knowledge throughout corporate and project organisations in A/E/C as well as its ICT-support. Several methods and technologies from ontologies to yellow pages and knowledge maps have been explored to distribute organisational knowledge among business participants. However, the straight adoption of these technologies in today’s AEC-industry remains a difficult task. So far, research projects have often neglected practical problems such as the up-to-date instantiation of knowledge models. Before KM-systems can be deployed in the industry, special restrictions and distinctive features of the AEC sector, such as its fragmentation, must be considered. These additional aspects obviously result in the need for developing new business structures and integrating sophisticated ICT tools in order to align social and organisational measures with the various technological approaches. The next chapter of the paper revisits the functional requirements of KM tasks and processes based on common KM models from the perspective of the AEC-industry. Critical factors for an effective deployment of KM are identified and alternative technological solutions are discussed. A framework for successively extending and combining basic technologies towards an integrated organisational memory system (OMS) is presented in the third chapter. A first prototype for retrieving and extracting knowledge from project corpora is described in the forth chapter. The prototype provides for comparing different methods for semi-automatic indexing and structuring of documents as well as reconfiguring documents for web-based information dissemination using e-learning systems. Future extensions for ontology learning and a strategy for future integration with workflow management systems and personalised configuration are presented.2.PROCESSES OF KNOWLEDGE EXCHANGE IN A/E/C-ORGANISATIONSWhat are the tasks and processes which need to be addressed for managing knowledge distribution in the AEC-industry and to what extend can these be supported by adequate ICT? During the last years several models, concepts and frameworks for knowledge management were proposed [e.g. Lehner 2000]. It is not the intention of this paper to present yet another concept. However, it is worthwhile to revisit some common KM concepts from the A/E/C perspective. Figure 1 illustrates the four processes of knowledge transformation (Socialisation, Externalisation, Internalisation, Combination) proposed by Nonaka and Takeuchi [Nonaka and Takeuchi 1995] among knowledge workers in a corporate setting. In regard to other managerial frameworks [e.g. Probst et al. 1999] the processes of knowledge identification and controlling were added to account for the aspects of directing and controlling the development of knowledge within the organisation. These processes are discussed in the next paragraphs. 2.1. IdentificationConcepts for identifying important knowledge and competencies such as Knowledge Asset Road Maps [Macintosh et al. 1998] often fail when analysing the individual knowledge worker’s perspective. However, particularly the day-to-day technical and procedural knowledge contributes to the effectiveness of project-centred organisations. Diverse decisions have to be supported in interdisciplinary design and construction teams. The potentials of top-down knowledge identification and dissemination are limited due to (a) disperse management and project teams, (b) dramatically changing product/service configurations, as well as (c) limited possibilities for divisional reorganisation. Due to these facts, knowledge in A/E/C needs to be identified on the ongoing building and construction projects. However, in practice it most often remains difficult for the individual to make out the essential findings of the latest project. Errors and failures are seldom reported because of internal and external competition. Consequently, knowledge boards and yellow pages often remaineSM@RT and CISEMIC 20022TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRYsparse, outdated repositories, and turning the knowledge transfer into a pull-process providing sophisticated search methods does not necessarily lead to acceptable improvements.Figure 1: Processes of Knowledge TransferIn our opinion, the prerequisite for effectively identifying relevant knowledge within remote work processes is to continuously track project activities and to document the relevant incidents. This can be achieved at minimum cost for all participants by integrating knowledge management systems into the common project ICT-infrastructure. This approach also provides for gathering additional context information that enables retrieval applications as well as knowledge engineers to employ more sophisticated and thorough analysis methods for externalising and combining the collected information. 2.2. SocialisationAs already discussed in the introduction, knowledge dissemination through socialisation is most commonly fostered by different reorganisation approaches. While ICT has an influence on the design of organisations and work processes, its potential of directly enabling socialisation is generally regarded as very limited. However, technologies, such as yellow pages or knowledge boards, can at least allow for initiating socialisation. Furthermore, it is argued that advanced communication technologies, such as video conferencing, can partly support socialisation processes and should thus be considered in business plans. 2.3. ExternalisationExternalisation of knowledge is regarded as the most challenging task of knowledge management. Making the knowledge of a person accessible to the overall organisation requires its codification using common models, metaphors and analogies. In this context, explicit knowledge can be referred to as well structured and contextualised information. It is classified as formalised, semi-formalised or unformalised. While formalised knowledge in the form of procedural or declarative knowledge bases can provide for automatic processing and reasoning, it is in general not understandable for users and its generation is very costly. Thus, it is assumed that, despite the emerging standards for organising design and construction data/information, the majority of A/E/C knowledge will remain documented in little-structured or unstructured forms such as graphics or text paragraphs. Thus, remaining on a3Sven-Eric Schapke, Karsten Menzel, Raimar J. Scherercontent level, the predominant task of externalisation is the transformation of these informal sources into well organised or, at least, semi-formalised repositories. During the last years, various European approaches to knowledge structuring, such as On-ToKnowledge [Staab et al. 2001] or Knowledge Desktop Environment ‘KDE’ [Winkels 2000], have explored the possibilities of further formalising the content of un-formalised or semi-formalised documents. Similar efforts for the A/E/C domain are currently undertaken by the e-COGNOS [Lefrancois et al. 2001] and, partly, by the eConstruct [Van Rees et al. 2002] projects. Generally, some form of metadata schemata is used to annotate the text of documents. Most recent approaches employ formal ontologies that provide domain-specific vocabularies and conceptstructures according to explicitly specified conceptualisations. This approach extends traditional metadata technologies, such as relational databases, taxonomies and thesauri, by additional relationships, axioms and general logical constraints to allow for reasoning by respective inference engines. However, formalising and annotating related information sources remains a difficult and time-consuming task and ontology-based retrieval remains limited to narrowly defined domains. Furthermore, although directly reflecting common concepts of a certain discipline or problem domain, ontologies remain hardly humanly readable. To increase the flexibility of systems some projects such as FRODO [van Elst and Abdecker 2000] and CoMMA [Gandon 2001], structure information according to several ontologies in parallel. While this approach seems feasible for fairly standardised business processes and organisations, its deployment in A/E/C remains a challenging task. With the goal of transferring knowledge among several different projects and disciplines, a large variety of knowledge models have to be employed, and the costs for ontology generation, management and annotation will rise accordingly. It is claimed that a variety of methods and technologies has to be further explored and integrated to allow for an efficient structuring of A/E/C knowledge. First, methods for automatic generation and adaptation of concept hierarchies and ontologies need to be explored to support knowledge engineers and system designers in deploying an adequate semantic network. Second, methods for automatic classification and annotation of information need to be refined thus supporting efficient information organisation. Third, common concepts for A/E/C knowledge models must be identified and standardised to allow for mapping between different semantic networks or metadata schemata. 2.4. Storage, Distribution & RetrievalProject portals or platforms and corporate intranets are becoming state of the art in A/E/C, providing an infrastructure for integrated knowledge management. Currently, the spectrum of information management services ranges from web-based content management systems employing sophisticated search methods to multi-agent environments. Especially two technologies are available for information distribution and knowledge management: (multi-) agent technology and data warehouse technology. Respective multi-agent systems have been proposed for general business applications e.g. CoMMA [Gandon 2001] as well as for the construction domain e.g. AMECS [Scherer et al. 2002]. While agent systems provide for most flexible retrieval, combination and reasoning, its performance, as well as the cost for preparing the information, have not been evaluated so far. The data warehouse (DW) can be optimised for handling complex information analyses. However, the DW-dimensions are commonly generated from particular database schemata so that queries and aggregations have to be explicitly specified. So far, applications of DWs have focused on managerial domains. New approaches for managing more complex information, e.g. product data models, and for dynamically adapting DWdimensions were proposed only recently [e.g. Menzel et al. 2002]. We argue that DW-technology is worthwhile considering for knowledge dissemination due to its good performance and systematic data structures, which can easily be complemented with knowledge management services such as data mining services.eSM@RT and CISEMIC 20024TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRY2.5.CombinationMore sophisticated information structures support the identification of new concepts, business relations, or procedural advice. On the one hand, this can be achieved by human knowledge engineers who analyse information sources and benchmark various project practices. Quality circles, expert groups, and problem-based learning scenarios are examples for initiating this knowledge combination process. On the other hand, well-formalised operational and contextual information can be used for automatic information analysis and knowledge discovery. While ontology-based reasoning provides for identifying corresponding information and complementing data sets, it is necessary to integrate robust data mining services that allow for recognising new patterns and concepts in less structured documents and analytical records such as cost estimates. Various retrieval and decision support services should provide knowledge combinations to the individual end user and knowledge engineers, as well as to other operational applications. 2.6. InternalisationThe processes of internalisation are often neglected by the different research approaches to KM systems. However, the questions of when and how knowledge is most effectively presented to result in improved business performance needs to be further explored. Most current KM systems require the user to initiate a search into the knowledge repository. The presentation formats are restrictive, and additional information on the context in which the knowledge can be applied is often missing. New efforts in the field of e-learning are a good catalyst for bridging the gap between the users’ knowledge requirements and the technical tasks of externalisation and combination. To provide for a more effective internalisation of knowledge throughout the organisation we see the need for: (a) contextualising information to enable users to integrate new knowledge with their individual mental model [Schapke et al. 2002], (b) providing for flexible adaptive presentation to prevent tiring descriptions or inappropriate formats, (c) enhancing information presentation and structuring in regard to didactic aspects e.g. by providing topics, sequencing, and background information, (d) personalising knowledge repositories to support archiving of personal retrieval paths, and organising information in regard to individual mental models, (e) sensing user needs to allow for proactively providing appropriate information in regard to the users current task and situation . 2.7. ControllingConcepts for controlling knowledge management activities and evaluating their effectiveness are seldom discussed. An estimate of the economical impact for the organisation remains questionable. However, the required user profiles for enhancing internalisation processes provide for measuring dissemination and acceptance of the systems and should be utilized for the development of future knowledge management plans.3.A FRAMEWORK TOWARDS ORGANISATIONAL MEMORY SYSTEMSToday, it is hardly feasible to develop such comprehensive knowledge management systems from scratch. In this chapter, a framework for combining several knowledge management services towards an organizational memory system (OMS) is presented. As discussed in chapter two, the various operational systems, such as electronic document management (EDM), workflow management (WFM), and product data management (PDM) systems, can deliver information elements required by the knowledge management services. Although often lacking formalisation, EDM-systems offer a large amount of rich and reusable content-based knowledge. More formalised information can be retrieved from WFM, PDM, CAE, or PM (project management), and ERP (enterprise resource5Sven-Eric Schapke, Karsten Menzel, Raimar J. Schererplanning) systems. However, in most of the cases the data/information models of these services are optimised for single problem domains and certain project phases. Information from WFM, PDM and especially CAE, PM and ERP systems requires extensive analysis and re-contextualisation before it can be presented to the knowledge worker e.g. by a Decision Support System. Nevertheless, we think that especially WFM and PDM are most beneficial systems for KM because they provide extensive context information that allows for evaluating and contextualising content-based knowledge elements. Some approaches to OMS focus on WFM systems providing an adequate operational platform as well as a large amount of business-related (context-)information such as time, organisation, function, task etc. [van Elst and Abecker 2002]. This process-oriented approach is especially promising in industries where divisions, functions and, respectively, information management are already organised along the organisations’ value chains. However, in A/E/C especially the technical knowledge is generally organised according to discipline-specific knowledge schemata, and workflows are less strictly defined. However, with an increasing utilisation of WFM systems throughout the AEC-industry and corresponding formalisations of common tasks and work descriptions, it will be feasible to successfully employ process-oriented approaches to knowledge management in A/E/C as well. In addition to process-oriented OMS, we consider the contextualisation of information in regard to product models as a very promising approach. We claim that despite the differences in the individuals’ mental models most of the domain-specific knowledge models are closely related to geometric models and can be derived from corresponding product data models. Several technologies that allow for translating product data into task or discipline-specific knowledge models, such as thesauri [Kosovac et al. 2000], taxonomies [Van Rees et al. 2002] and ontologies [Katranuschkov et al. 2002], have already been explored. In practice, it will be necessary to combine different approaches to KM and to utilise a variety of knowledge sources in order to provide for most effective and sustainable KM. Thus, we propose building organisational memory systems to support an overall knowledge transformation process of identification, externalisation, combination, internalisation, and controlling, which can be integrated successively with different information and communication systems of an organisation. In the organisational memory, operational information is captured in small atomic information elements and meta-data is as much as possible added to allow for generating a large variety of ‘knowledge perspectives.’ One important advantage of an OMS can be the ability of gathering specific context information from the users of the knowledge management as well as the operational information services for guiding and controlling the knowledge transformation processes and further supporting (semi-)automatic analysis and transformation methods. Figure 2 represents a general framework which illustrates different approaches to disseminating A/E/C knowledge from various sources by an OMS. The different components supporting the five processes of knowledge transformation are addressed on three layers, namely the data acquisition layer (operational information systems), the information management layer (the kernel of the organisation memory systems), and the interface layer (personalized information space). Meta-models that guide information structuring and management on the information management layer are depicted on the right. The instantiated knowledge models are presented on the left. On the data acquisition layer, relevant information needs to be selected from the operational data and organised into self-contained, reusable information elements. This first process of transformation and contextualisation should be carried out automatically. A variety of necessary meta-data is already provided by the operational systems. Synchronising information from different operational information systems and external (context-)services, e.g. a weather service, allows for capturing the particular context in which the information was created. In order to further support knowledge externalisation and combination processes on the information management layer, it is necessary to provide flexible memory-models in which the information elements can be pre-configured and efficiently organised in correspondence to typical disciplineeSM@RT and CISEMIC 2002 6TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRYperspectives and analyses. As discussed, different technologies, such as concept hierarchies, can be used for multi-dimensional data management. Context models should support the generation and modification of the memory models based on common meta models and metadata acquisition rules. In combination with the memory models they improve the flexible information representation and performance of complex information requests by defining information combination in different granularities.Figure 2: Proposed Framework for A/E/C-specific OMS Mental models can be interpreted as the personalised “counterpart” of context and memory models. They are to simplify the data internalisation process by reducing the complexity, decreasing the input efforts as well as personalizing the data input process. They shall be generated in regard to the discipline-specific memory models. The second important component of the interface layer are the adaptable presentation models. These adaptable presentation models should be developed in correspondence to standard educational metadata schemata such as SCORM [Schapke et al. 2002] and must be linked with the actually implemented memory models.4.RETRIEVAL AND STRUCTURING METHODS FOR BUILDING OMSAfter having discussed the various tasks and processes of knowledge distribution as well as general technological solutions, our first steps of developing knowledge management services that can be combined towards an OMS are presented. In the big picture of the overall framework the Document Modelling (DocMo) service presented focuses on knowledge externalisation using content-based information/knowledge elements (further called knowledge elements). The DocMo service explores various methods for (semi-)automatically structuring document content in different memory models. In a second step these models shall be tested in building-model-based as well as traditional e-learning scenarios. It will be evaluated to which extend the different - partially automatically generated structures - support directed knowledge dissemination as well as information retrieval by knowledge engineers and individual knowledge workers. Currently, corporate and project document repositories provide the most comprehensive collections of business and engineering knowledge. There is a need for making this experience accessible to the engineer. Since we expect the majority of such contents to remain documented in little-structured7Sven-Eric Schapke, Karsten Menzel, Raimar J. Schererdocuments, the analyses implemented in the DocMo service mainly focuses on text information. Data of related management and engineering applications as well as external context information are currently not considered. Focusing on content-based knowledge exchange, the DocMo service uses a flexible document repository that can be connected to operational document and content management services. The repository is based on a relational database that implements the IFC 2x IfcExternalReference schema. It provides standardised metadata descriptions for documents, libraries and classification schemata. For the DocMo service the IFC schema was extended to also handle keywords as well as single fragments of an original document. Thus, different text paragraphs, pictures or data models of a document can be treated as its self-contained knowledge elements. In conjunction with the database of the ‘ISTforCE Personal Planning System’ [Keller et al. 2002] that implements the IFC 2x process model, the repository can also capture document-related management information such as actors, organisations, addresses, etc. The compliance with the IFC 2x data model and the Personal Planning System shall allow for a later integration with workflow-management and product data services in the form of context as well as content-providing services.Figure 3: GUI of the first prototype The two different corpora of original documents and corresponding knowledge elements are managed separately by the DocMo service. A module for regularly importing new documents from different operational sources is provided. The original documents can then be fragmented automatically by a tokenscanner or using a manual HTML-based annotator. The fragments are classified as either text, graphic or model information and stored in the ‘knowledge management corpora.’ All transformation processes are implemented using the Oracle InterMedia package that already provides a variety of filters for extracting the text from common word processors, spreadsheets, and presentation file formats. Figure 3 depicts a first GUI of the DocMo service, illustrating readily classified document fragments as well as the content and the token-vector of a selected knowledge element.eSM@RT and CISEMIC 2002 8TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRYTwo modules for initially identifying significant information within the document repository are implemented. A search engine based on the Oracle InterMedia package offers sophisticated full-text search features for queering the text as well as the metadata in both corpora. Furthermore, the ‘KEA’ keyphrase-extractor based on a Naive Bayes algorithm by Frank and Paynter [Frank et al. 1999] is integrated to supplement metadata of unlabelled documents as well as newly generated subdocuments. As a first approach to knowledge structuring, the DocMo service allows for classifying the text fragments into simple concept hierarchies. A ‘Classification-Module’ provides for editing classification trees and manually assigning knowledge elements to categories. Furthermore, a Naive Bayes classifiers trained on the manually selected a-priori configuration can be used to automatically classification new knowledge elements. An additional ‘Clustering-Module’ providing for automatically generating concept hierarchies is currently under development. Both modules employ basic text mining methods based on a ‘bag of word’ description of the text paragraphs. Furthermore, two additional modules are under development. With an ‘Information Extraction Module’, methods for automatically identifying the essential attributes within work descriptions are explored. A ‘Modelling Module’, analysing the similarities among different documents fragments from a project repository as well as the possible correlations with the corresponding product data model shall provide first experience of an adaptive structuring of knowledge in regard to actual building models. For a later validation of the analysis methods the repository can also easily be extended with educational metadata. Integrating the information specified in the educational section of the SCORM Metadata Information Model, the textual and graphical knowledge elements can be re-configured for the use in standard e-learning environments.REFERENCESAbecker A., Mentzas G. (2001): Active Knowledge Delivery in Semi-Structured Administrative Processes. In: Proc. of 2nd Int. Workshop on Knowledge management in Electronic Government, 2001, Italy. Frank E. et al. (1999): Domain-specific keyphrase extraction. Proc. of 16th Int. Joint Conference on Artificial Intelligence, Morgan Kaufmann Publ., San Francisco, pp. 668-673. Gandon F. (2001): Engineering an Ontology for a Mutli-Agents Corporate Memory System. ACACIA, INRIA Sophia Antipolis, 2001. Katranuschkov P. et al. (2002): An Engineering Ontology Framework As Advanced User Gateway to IFC Model Data. In: Proc. of 5th European Conference on Product and Process Modelling, 2002, Slovenia. Keller M., Scherer R., Menzel K. (2002): A Personal Planning Approach for the Integration and Coordination of Multi Project Process Information. In: Proc. of 5th European Conference on Product and Process Modelling, 2002, Slovenia. Kosovac B., et al. (2000): Integrating Heterogeneous Data Representations in Model-Based AEC/FM Systems. In: Proc. of Int. Conference on Construction Information Technology, pp. 556-567, Iceland, 2000. Lefrancois G., Lima C., Fiès B., Wetherill M., Zarli A., Rezgui Y. (2001): e-COGNOS Base Technology Selection. Deliverable D2.1, e-COGNOS: IST-, December 2001 Lehner F. (2000): Organisational Memory C Konzepte Systeme für das organisatorische Lernen und das Wissensmanagement. Hanser, Munich, Germany.9Sven-Eric Schapke, Karsten Menzel, Raimar J. SchererMacintosh A., Filby I., Tate A. (1998): Knowledge Asset Road Maps. In: Proceedings of the 2nd Int. Conference on Practical Aspects of Knowledge Management C PAKM 98, Basel, Switzerland, 1998. Menzel K., et al. (2002): Potentials of Data Warehouse Technology to Support Case Sensitive Information Representation on Mobile Devices. In: Proc. of the 9th ISPE Int. Conference on Concurrent Engineering, Cranfield, U.K., 2002. Nonaka I., Takeuchi H. (1995): The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation, Oxford University Press, N.Y., 1995. Probst G., Rohhardt K., Raub S. (1997): Wissensmanagement C Wie Unternehmen ihre wertvollste Ressource optimal nutzen. Frankfurt, 1997 Schapke S., Menzel K., Eisenreich T., Otto C. (2002): Virtual Environments for Content Presentation and Knowledge Management in Civil Engineering Education. In: Proc. of 1st International Workshop on Construction Information Technology in Education, Portoroz, Slovenia, September, 2002. Scherer R.J., Katranuschkov P., Gehre A. (2002): Towards an Agent Enabled Environment for Mobile e-Work on the Construction Site. In: Proc. of European Conference on Information and Communication Techn. of Advances and Innovation in the Knowledge Society. 19-21 November, 2002, Salford, UK. Staab S., Schnurr H.-P., Studer R., Sure Y.(2001): Knowledge Processes and Ontologies. In: IEEE Intelligent Systems. 16(1), Special Issue on Knowledge Management, January/February 2001 Van Rees R., Tolman F., Beheshti R. (2002): How BcXML Handles Construction Semantics. In: Defining the matrix of Communication processes in the AEC/FM Industry: Current Developments and Gap Analysis. Proc. of Internat. Conf. on Construction Information Technology C CIB W78, IABSE, Vol. 1, Denmark. Winkels R.G.F et al. (2000): Extended Conceptual Retrieval. In: Breuker et al. (Eds): Legal Knowledge and Information Systems 2000. IOS Press, Amsterdam, pp. 85-98. Van Elst L., Abecker A. (2002): Domain Ontology Agents for Distributed Organizational Memories. In: Rose Dieng-Kuntz et al.: Knowledge Management & Organizational Memories. Kluwer Acad. Publ., 2002.eSM@RT and CISEMIC 200210
TOWARDS ORGANISATIONAL MEMORY SYSTEMS IN THE CONSTRUCTION INDUSTRY―汇集和整理大量word文档,专业文献,应用文书,考试资料,教学教材,办公文档,教程攻略,文档搜索下载下载,拥有海量中文文档库,关注高价值的实用信息,我们一直在努力,争取提供更多下载资源。
[] [] [] [] [] [] [] [] [] [] []}