Source-linked AI summary
Knowledge Management in Software Engineering: A Systematic Review of Studied Concepts, Findings and Research Methods Used
Finn Olav Bjørnson, Torgeir Dingsøyr
TL;DR
Software engineering has an increased demand for knowledge management as the field matures. This systematic review uses a comprehensive, unbiased search to examine the evidence and finds that explicit knowledge should not be considered exclusively, while lessons-learned reports predominate in industry studies.
Problem
As software engineering matures, increased demand for knowledge management motivates examining how the field manages knowledge.
Method
The paper conducts a systematic review using a comprehensive, unbiased search and classifies studies according to Earl’s schools.
Results
The major finding repeated across several papers and schools is that knowledge management should consider tacit knowledge rather than focusing exclusively on explicit knowledge; industry studies are predominantly lessons-learned reports.
Takeaways & Limitations
The main implication for research is to focus more on the organizational school.
Takeaways & Limitations
The review’s main threats to validity concern study selection and correct classification, alongside a heavy publication bias toward success stories.
Abstract
from arXiv · showhide
Software engineering is knowledge-intensive work, and how to manage software engineering knowledge has received much attention. This systematic review identifies empirical studies of knowledge management initiatives in software engineering, and discusses the concepts studied, the major findings, and the research methods used. Seven hundred and sixty-two articles were identified, of which 68 were studies in an industry context. Of these, 29 were empirical studies and 39 reports of lessons learned. More than half of the empirical studies were case studies. The majority of empirical studies relate to technocratic and behavioural aspects of knowledge management, while there are few studies relating to economic, spatial and cartographic approaches. A finding reported across multiple papers was the need to not focus exclusively on explicit knowledge, but also consider tacit knowledge. We also describe implications for research and for practice.
1. Introduction
Software engineering is knowledge-intensive, creating a need to understand how knowledge-management initiatives are studied and what they find. This review maps empirical studies, concepts, findings, and methods in the field.
- The review responds to growing difficulty in critically evaluating and synthesising an expanding software-engineering research literature.
- The review examines empirical knowledge-management studies in software engineering, including their concepts, findings, and research methods.
- The authors ask which knowledge-management concepts, findings, and research methods have been investigated in software engineering.
- The intended audience includes software-engineering researchers, knowledge-management researchers, and reflective software-engineering practitioners.
2. Background
Knowledge management spans technological, organizational, economic, spatial, and strategic approaches to sharing and applying knowledge. Prior literature emphasizes knowledge creation and transfer, while warning that knowledge is socially embedded and cannot be handled through one optimal solution.
- Knowledge-management strategies include codification of organizational information and personalization through knowledge-source directories.
- Earl’s framework groups knowledge-management work into technocratic, economic, and behavioural schools with distinct emphases.Technocratic work includes systems, cartographic, and engineering approaches; behavioural work includes organizational, spatial, and strategic approaches.
- Knowledge management is described as socially embedded and dynamic, requiring varied approaches for diverse knowledge types rather than a single optimal solution.
- Prior reviews found emphasis on knowledge creation and transfer, with inadequate coverage of integration, production, measurement, retention, and reflection.
- Critiques caution that information technology may not create knowledge where people lack shared understanding, contact, or awareness of useful information.
- The review identifies a lack of broad systematic reviews of knowledge-management research in software engineering.
3. Method
The authors conducted a protocol-driven systematic review using broad searches, staged screening, quality assessment, and framework-based synthesis. The process reduced the literature through successive selection stages and separated empirical studies from lessons-learned reports.
- The review protocol specified research questions, search strategy, inclusion and exclusion criteria, and synthesis methods in advance.
- The search used broad software-engineering and knowledge-management keywords across Web of Science, Compendex, IEEE Xplore, and ACM Digital Library.
- 2102 articles were initially identified, 762 remained after title screening, 133 after abstract screening, and 68 after full-text review.
- Inter-rater agreement was Kappa 0,655 after initial screening and Kappa 0,523 after full-text selection.
- The 68 papers were classified using Earl’s framework and assessed as empirical studies or lessons-learned reports.
- Synthesis used empirical studies to avoid problems associated with the limited scientific rigor of lessons-learned reports.
4. Results
The reviewed literature concentrates on technocratic and behavioural knowledge-management approaches, with systems, engineering, organizational, and strategic topics receiving the most attention. Empirical studies increasingly replaced lessons-learned reports, but case studies remained the dominant method.
- The literature heavily concentrated on technocratic schools, with substantial behavioural coverage and no papers in the economic school.
- Systems and engineering received the most attention among technocratic topics, while organizational and strategic topics led the behavioural schools.
- Four empirical studies did not fit Earl’s framework.
- Interest in the area increased from 1999 onward, with a shift from lessons-learned reports toward empirical papers from 2003.
4.1 Technocratic schools
Technocratic knowledge-management research in software engineering emphasizes systems for storing and retrieving knowledge and engineering processes for transferring experience. Findings also show that formal routines and codification must be complemented by social interaction and tacit knowledge.
- Systems school: Systems-school studies focused on developing and using knowledge repositories and knowledge bases.
- Systems school: Knowledge repositories produced distinct usage patterns, and their informal organization and integration into daily work were important.
- Systems school: Experience repositories can provide tangible benefits quickly while remaining useful and accumulating further benefits over time.
- Cartographic school: The cartographic school had only one empirical study, examining a skills-management tool and identifying four major uses.
- Engineering school: Engineering-school studies covered formal routines, knowledge-flow mapping, project reviews, and social interaction, with formal processes requiring complementary collaboration.
- Engineering school: Across approaches, both explicit and tacit knowledge were required, while pure codification was associated with ineffective knowledge sharing.
4.2 Behavioural schools
Behavioural-school research examines organizational networks, networking, and strategic knowledge-management practices, but remains limited compared with technocratic research. Findings emphasize informal networks, interaction, and the joint importance of tacit and explicit knowledge.
- The review found three organizational empirical studies, no spatial empirical studies, and three strategic empirical studies.
- Organizational school: Organizational studies examined networks and networking, combining technological knowledge sharing with trust and practitioner collaboration.
- Organizational school: Knowledge management was relevant to introducing new software-engineering methods, with companies needing a suitable balance between strategies.
- Organizational school: Building a knowledge network on existing informal networks provided the highest organizational value in one medium-sized software company.
- Strategic school: Strategic-school studies examined processes supporting learning and factors associated with successful knowledge management.
- Strategic school: Improving software practice required both tacit and explicit knowledge and ongoing interaction among different participants.
4.3 Knowledge management in general
Studies outside Earl’s schools examined knowledge-management impacts, knowledge delivery, organizational factors, and explicit-knowledge use. Their findings highlight internalization, user-driven retrieval, leadership, and organizational conditions.
- Four empirical studies formed a broad category addressing knowledge-management initiative impacts and knowledge management itself.
- Survey evidence from 41 North American software companies indicated that user-driven “pull” approaches led to more effective software development.
- Knowledge creation affected process capability when knowledge was embedded after creation, requiring internalization before process improvement.
- Explicit-knowledge use related to perceived complexity, perceived relative advantage, and perceived risk.
- Leadership was the most significant positive factor for knowledge management, while existing tools, techniques, and methodologies were inadequate.
5. Discussion
The review finds that software-engineering knowledge-management research is broadening but remains concentrated on technocratic and engineering concerns. Across the literature, effective knowledge management requires attention to tacit knowledge, social interaction, and organizational context, alongside technical infrastructure.
- The review identified more studies than earlier assessments, but software engineering still emphasizes knowledge storage and retrieval over creation, transfer, and application.
- Research covered all knowledge-management schools except the economic school, although empirical studies were concentrated in the systems and engineering schools.
- Agile development emphasizes tacit-knowledge activities, whereas traditional development processes require activities related to explicit knowledge.
- The literature was divided between technocratic and behavioural schools, with emphasis on the technocratic side and especially tools and processes.
- Technical infrastructure facilitated explicit-knowledge sharing, but studies also found it difficult to achieve actual technology use.
- Only one empirical study addressed the cartographic school, whose tools could stimulate individual and organizational learning without requiring knowledge codification.
- Engineering studies covered formal routines, knowledge-flow mapping, project reviews, and social interaction, while established processes could improve communication and learning.
- Sharing tacit knowledge was important for changing development practices, and formal routines alone did not ensure effective dissemination or organizational learning.
5.2 Research methods
The reviewed studies were dominated by case studies, with few experiments, while empirical studies represented a notable share of the software-engineering knowledge-management literature. The authors argue that maturing the field requires more in-depth studies, including ethnographic work.
- 29 studies were empirical and 39 reported lessons learned among the 68 industry-context studies identified.
- Case studies constituted the largest number of empirical studies, followed by field studies and action research.
- Only one experiment was found, reflecting the difficulty of isolating knowledge-management factors in a broad field.
- The authors support many case studies but call for more in-depth studies of knowledge-management effects as the field matures.
- They specifically recommend a larger use of ethnography in companies.
- The final selection constituted about 3% of the initially selected papers, while empirical studies appeared to form a larger share of this area than of software engineering overall.
5.3 Implications for research and practice
The review’s implications concern aligning knowledge-management strategies with development context and broadening research beyond dominant technocratic approaches. It recommends more attention to organizational, spatial, cartographic, behavioural, and in-depth research perspectives while recognizing review-validity constraints.
- Implications for research and practice: Knowledge-management activities should be determined by how a company develops software, distinguishing traditional and agile development.
- Implications for research and practice: Traditional development has received the most attention through systems and engineering schools, whereas agile development motivates further attention to agile-relevant schools.
- Implications for research: The authors argue that organizational, cartographic, and spatial schools deserve further research, with the organizational school relevant to both agile and traditional development and inexpensive.
- Implications for practice: Traditional teams will probably benefit more from technocratic schools, while agile teams would benefit more from behavioural schools.
- Implications for practice: Several papers underline the importance of not focusing exclusively on codification, whose heavily technical strategies can be costly to implement.
- Implications for practice: Behavioural approaches cost less than application-heavy technocratic solutions, but fewer publications are available from which to learn.
- Limitations: The review’s main validity threats involve study selection, classification according to Earl’s framework, and potential author bias.
- Limitations: Selection used one researcher for title-based discarding, while later abstract and full-paper stages involved both researchers and showed a good degree of consensus.
6. Conclusion
The review found that software-engineering knowledge-management research emphasizes technocratic and behavioural approaches, with limited coverage of economic, spatial, and cartographic approaches. Across several papers, findings stressed addressing tacit as well as explicit knowledge, while the evidence base was dominated by lessons-learned reports and case studies.
- Most studies examined technocratic and behavioural aspects of knowledge management.
- Few studies addressed economic, spatial, or cartographic approaches to knowledge management.
- Findings across several papers emphasized not focusing exclusively on explicit knowledge, but also considering tacit knowledge.
- Most industry applications were reported as lessons learned rather than scientific empirical studies.
- More than half of the empirical reports were case studies, alongside field studies, action research, ethnographic studies, and one laboratory experiment.
- The review recommends more research on the organizational school and practitioner attention to activities managing tacit knowledge.
Appendix
The appendix contains extended tables categorizing articles and surveying the research methods used.
- Table 11 presents categorized articles in an extended Systems Cartographic table.
- Table 12 provides an extended overview of research methods.