Journal of Software Engineering and Applications, 2011, 4, 335-344
doi:10.4236/jsea.2011.46038 Published Online June 2011 (http://www.SciRP.org/journal/jsea)
335
Research on Explicit and Tacit Knowledge
Interaction in Software Process Improvement
Project*
Jiangping Wan1,2, Dan Wan1, Weiping Luo1, Xiaoyao Wan1
1Faculty of Business Administration, South China University of Technology, Guangzhou, China; 2Institute of Emerging Industriali-
zation Development, South China University of Technology, Guangzhou, China.
Received May 19th, 2011; revised June 5th, 2011; accepted June 13th, 2011.
ABSTRACT
This research develops a knowledge model for Software Process Improvement (SPI) project based on knowledge crea-
tion theory and its twenty-four measurement items, and proposes two hypothesizes about the interaction of explicit
knowledge and tacit knowledge in SPI. Eleven factors are extracted through statistical analysis. Three knowledge-
creation practices for capturing tacit knowledge contribute greatly to SPI, which are communication among members,
crossover collaboration in practical work and pair programming. Two knowledge-creation practices for capturing ex-
plicit knowledge have significant positive impact on SPI, which are integrating project document and on-the-job train-
ing. Ultimately, suggestions for improvement are put forward, that is, encouraging communication among staff and
integrating documents in real time, and future research is also illustrated.
Keywords: Knowledge Creation, Software Process Improvement, Explicit Knowledge, Tacit Knowledge,
Communication, Document
1. Introduction
The theory of knowledge creation [1] is based primarily
on Polanyi’s [2] categorization of knowledge as explicit
and tacit. It prescribes the capture of both explicit and
tacit types of knowledge, making it available to the or-
ganization in order to generate competitive capabilities.
Explicit knowledge is codified knowledge articulated in
words, figures, and numbers. It is objective, and rela-
tively easy to share in the form of specifications, stan-
dard operating procedures, and data. Tacit knowledge is
knowledge that has not been codified and is relatively
difficult to codify. It is subjective and based on individ-
ual experiences.
The software process is the set of tools, method, and
practices we use to produce a software product. The ob-
jectives of software process improvement (SPI) are to
process produce products according to plan while simul-
taneously improving the organization’s capability to
produce better products [3]. The six basic principles of
SPI by Watts S. Humphrey are in the following: 1) Major
changes to the software process must start at the top. 2)
Ultimately, everyone must be involved. 3) Effective
change requires a goal and knowledge of the current
process. 4) Change is continuous. 5) Software process
changes will not be retained without conscious effort and
periodic reinforcement. 6) Software process improve-
ment requires investment [3]. Alfonso Fuggetta argues
that the scope of software process improvement methods
and models should be widened in order to consider all
the different factors affecting software development ac-
tivities. We should reuse the experiences gained in other
business domains and in organizational behavior research.
Statistics is not the only source of knowledge. We should
also appreciate the value of qualitative observations [4]. J.
P. Wan argues that managers should think deeply into
their think processes. The following issues in software
organization can be resolved with SPI: 1) The processes
and their principles for how to inherit and acquire others’
knowledge. 2) The processes and their principles for
conversion knowledge into their capability [5]. 400 pro-
cess improvement experiments and presents patterns are
*This research was supported by Key Project of Guangdong Province
Education Office (06JDXM63002).

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
336
in a repository to help organizations plan their improve-
ment initiatives [6].
We begin by relating knowledge creation to SPI pro-
jects in Section 2. In Section 3, we develop our concep-
tual arguments, and present hypotheses that relate prac-
tices for capturing explicit and tacit knowledge to SPI
project success. Section 4 describes the development of
our survey instrument and the empirical methodology
used to test our hypotheses. We present the results of our
analyses and discussions in Section 5, the conclusions are
in Section 6.
2. Literature Review
2.1. Software Process
Since the computer software is a kind of logical product,
its quality improvement is difficult and complex. It is
clear that a fully effective software process must consider
the relationships of all the required tasks, including the
tools and methods used, the skill, training, and motiva-
tion of the people involved, SPI is usually implemented
by project [3,7].
An economist, Howard Baetjer, commented on the
software process as following [8]: because software,
likes all capital, is concrete knowledge, and because that
knowledge is initially dispersed, tacit, latent, and incom-
plete in large measure, software development is a social
learning process. The process is a dialogue in which the
knowledge on the software is brought together and em-
bodied in the software. The process enables interaction
between users and designers, between users and evolving
tools, and between designers and evolving tool (technol-
ogy). It is an iterative process in which the evolving tool
itself serves as the medium for communication, with each
new round of the dialogue eliciting more useful knowl-
edge from the people involved. It is obvious that soft-
ware process is also an organizational knowledge-inten-
sive learning process and needed to be supported with
knowledge management.
Software organization is a highly knowledge-inten-
sive enterprise, knowledge transfer is critical for soft-
ware enterprise. It is obvious that software process is also
an organizational knowledge intensive learning process
and needed to be supported with knowledge management
[9].
2.2. Explicit and Tacit Knowledge Types
Nonaka’s framework [1] provides a rationale for the use
of knowledge-creation practices to generate group
knowledge by engaging individual team members in
process improvement projects. The framework depicts
the process of knowledge creation as cycles of conver-
sions between two types of knowledge—explicit, and
tacit (Figure 1). It is worthwhile to note that this classi-
Figure 1. Nonaka’s framework of knowledge-creation me-
chanisms [10].
fication of knowledge as either explicit or tacit is one of
two prominent classifications in the knowledge manage-
ment literature (Table 1 provides a brief overview of dif-
ferent classifications of knowledge creation efforts [10]).
Explicit knowledge is codified and documented, and
its transfer can take place in impersonal ways—for in-
stance, through written instructions and diagrams. Tacit
knowledge is knowledge that is difficult to articulate,
especially in terms of cause-effect relationships. It is
context-specific, and is transferred mainly through social
interactions [2]. Language is an excellent example of
tacit knowledge: native speakers of a language are often
unable to articulate the grammatical and syntactic rules
governing it. Tacit knowledge contributes to the “sticki-
ness” of information required for problem-solving, mak-
ing it difficult for others to gather, transfer, and utilize.
The difficult-to-codify nature of tacit knowledge con-
tributes to difficult-to-imitate capabilities that may pro-
vide competitive advantage to the organization. Success
of process improvement projects depends on the capture
of both explicit and tacit types of knowledge [10,11].
2.3. Knowledge Enabling Software Process
Improvement
Knowledge transfer model of SPI and the conceptual
Table 1. Selected classifications of knowledge-creation me-
chanisms [10].
Author(s) Year Knowledge-creation mechanisms
Argyris 1977 Single & double loop learning
Nonaka 1991
Combination, internalization,
socialization & externalization
Kogut and
Zander 1992 Operational & conceptual
Spender 1996
Capturing individual and
organizational knowledge
Nahapiet and
Ghoshal 1998 Acquiring intellectual & social capital

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
337
framework of influencing factors are established in [12].
The model includes five elements which are knowledge
of transfer, sources of knowledge, recipients of knowl-
edge, relationship of transfer parties, and the environ-
ment of transfer. The conceptual framework includes ten
key factors which are ambiguity, institutionalization,
transfer willingness, capacity of impartation, capacity of
absorption, incentive mechanism, culture, technical sup-
port, trust and knowledge distance.
The knowledge creation effective factors were found
in both necessary elements for stimulus of knowledge
creation and the key influencing factors of software pro-
ject success. The research was carried with the specific
successful practices of both Microsoft Corporation and
William Johnson’s analysis of R&D project knowledge
creation in [13]. The knowledge creation effective factors
in requirement development project are clarified through
deeply interviewing the software enterprises in Guang-
dong province as well as other corporate information
departments. After field survey and literature review, we
found that software requirement development (SRD) is a
knowledge creation process. Knowledge creation theory
of Nonaka is appropriate for analyzing knowledge creat-
ing of SRD. The issue of this research is exported: how
to improve SRD with knowledge creation theory? And it
includes three sub-issues: 1) What factors are impacting
SRD in the view of knowledge creation theory? 2) What
do enable knowledge flow during SRD? 3) How can we
guide SRD by using knowledge creation theory? Case
study findings include [14]: 1) It can facilitate the im-
plementation of the project to have the necessary diver-
sity of the project team. 2) The introduction of experts on
requirement can achieve the transformation of knowl-
edge effectively, thus helping to carry out the project. 3)
Methodology and related technologies are important ba-
sis for carrying out the project.
Knowledge creation theory is also applied to analyze
the open source software community with successful
application of the typical agile software methods, ten
principles of knowledge creation in open source software
community are put forward as follows: self-organizing,
code sharing, adaptation, usability, sustention, talent,
interaction, collaboration, happiness, and democracy
[15].
The problems studied are as follows: 1) explaining the
different success levels of SPI project; 2) potential bene-
fits of capturing tacit knowledge which is easier to be
ignored.
3. Conceptual Development
We focus on the explicit-tacit distinction, and develop
hypotheses related to the value of capturing the two types
of knowledge in SPI projects (Figure 2).
3.1. Capturing Explicit Knowledge
SPI project teams often deal with cross-functional and
cross-divisional issues that warrant the use of integrative
knowledge practices. In addition, the teams consist of
members from diverse backgrounds who come together
only for the duration of the project and, even while in-
volved in a project, work only part-time on the project.
Using combination (explicit-explicit) practices, project
team leaders can help their teams sift through explicit
data, drawing explicit insights about the targeted proc-
esses. In addition, internalization (explicit-tacit) practices
make it possible for the explicit knowledge that is har-
nessed to be comprehended and absorbed by team mem-
bers and people working on the processes. Such recom-
bination of explicit knowledge and its conversion into
tacit knowledge is critical for the creation of team
knowledge about the working of the processes being tar-
geted for improvement. Thus, our first set of hypotheses:
(H1) cover the importance of capturing explicit knowl-
edge for the success of SPI projects.
Figure 2. Proposed conceptual model and hypotheses: knowledge-creation practices as predictors of SPI project success.

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
338
H1. The following knowledge-creation practices for
capturing explicit knowledge contribute significantly and
positively to SPI project success:
1) combination (explicit-explicit knowledge), and
2) internalization (explicit-tacit knowledge).
3.2. Capturing Tacit Knowledge
In SPI projects, externalization (tacit-explicit) practices
aid the conversion of difficult-to-codify tacit knowledge
into explicit knowledge by providing templates [16],
such as cause-and-effect diagrams and failure modes and
effects analysis charts. Such templates serve as a common
and convenient language for team members, facilitating
communication and analysis, and resulting in knowledge
that helps to achieve project goals.
Socialization and externalization (tacit-tacit, and tacit-
explicit) practices are designed to capture the more-dif-
ficult-to capture tacit knowledge from team members that
may be crucial for the success of SPI projects. The cap-
ture of such knowledge in SPI projects can provide in-
sights that result in higher levels of process improve-
ments than could be achieved solely through explicit-
knowledge-capturing practices. Our second set of hy-
potheses (H2) state that the integration of tacit knowl-
edge through socialization (tacit-tacit knowledge) and
externalization (tacit-explicit knowledge) practices adds
value over and above that created by concentrating solely
on the utilization of explicit knowledge.
H2. The following knowledge-creation practices for
capturing tacit knowledge contribute significantly and
positively to SPI project success over and above the ef-
fects of practices that capture explicit knowledge:
1) socialization (tacit-tacit knowledge), and
2) externalization (tacit-explicit knowledge).
3.3. Measure Items of SPI with Knowledge
Creation Theory
The heterogeneity of tacit knowledge makes a significant
direct contribution to SPI success. Explicit knowledge
can be transferred and stored in the database with coding,
text and diagram and so on, so that it can contribute to
SPI project indirectly in the long run.
3.4. Knowledge Creation and Success of SPI
Project
This research proposes the following four items to meas-
ure the contribution of knowledge-creation mechanism to
SPI project (Figure 3): 1) Continuous communication and
discussion among related project members. 2) Document-
ing subjective ideas and requirements. 3) Project team
members absorbing combination knowledge through
training, and converting it to his operation knowledge. 4)
To integrate knowledge from existing objective data.
4. Methods
The questionnaire for this research consists of six parts:
basic information, socialization for capturing tacit know-
ledge, externalization for capturing tacit knowledge, in-
ternalization for capturing explicit knowledge, combi-
nation for capturing explicit knowledge, the relationship
between knowledge creation and SPI. A total of 200
questionnaires were distributed, with 104 responses
among which two are invalid, so there remained 102 va-
lid ones, resulting in a valid response rate of 51%. The
respondents are SPI personnel from domestic (Guangzhou,
Shenzhen and Hangzhou) software organizations (com-
panies).
4.1. Data Characteristics
The usable sample distribution characteristics were ana-
lyzed in Table 2. Some one has one more positions, es-
pecial in SME.
4.2. Data Validation
The Cronbach’s alpha coefficients were all over 0.6,
proving high validation (Table 3).
4.3. Factor Analysis
Three factors were extracted through factor analysis with
Varimax rotation. The three factors in Table 4 account
for 64.717% of the total variance. The first factor repre-
sented items S1, S2 and S3, so it was named communica-
tion factor (SF1). The second factor represented items S4,
S5 and S6 with a name crossover collaboration factor
(SF2). The last factor represented items S7 and S8, and
we call it work results factor (SF3).
The three factors in Table 5 explain 77.639% of the
total variance. The first factor represents items E1, E2
and E3, and we call it requirements documentation factor
(EF1). The second factor represents items E5 and E6
with a name pair programming factor (EF2). The third
factor represents item E4, called recording in real time
(EF3).
The three factors in Table 6 extract 72.806% of the
total variance. The first factor represents items I1and I2,
and we call it on-the-job training factor (IF1). The sec-
ond factor represents items I5 and I6 with a name in-
formation utilization factor (IF2). The third factor
represents items I3 and I4, called document discussion
(IF3).
The two factors in Table 7 extract 75.734% of the to-
tal variance. The first factor represents items C1, C2 and
C4, and we call it integrating documents factor (CF1).
The second factor represents items C3 with a name pro-
cedure standardization factor (CF2).

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project339
Socialization
S1. Continuous communication among
project team members
S2. Communication among project team
members and executives
S3. Frequent communication among project
team members and clients
S4. Helping and discussing with each other
on work problems and divergence
S5. Communicating frequently by face-to-
face conference, video and phone.
S6. Project team comprised of various
profession backgrounds and knowledge skills
S7. Apprenticeships, learning by doing
S8. Gains of personnel in the practical work
of this project
Externalization
E1. Formalizing implied project objectives by
documentation
E2. Formally and systematically listing
implied customer requirements
E3. Linking tacit customer requirements to
specified process characteristics
E4. Recording improvement ideas in a data
base in real time
E5. Converting subjective customer
requirements to objective requirements
E6. Checking the work reciprocally among
personnel and then inducing problems
Internalization
I1. Providing training to personnel according
to job and project requirements
I2. Using diagrams and models to initiate
discussions during the project
I3. Using documents to initiate discussions
about project performance
I4. Using documents to generate discussions
after implementation of results
I5. Communicating among team members
through Email, fax and BBS
I6. Applying information in the existing
organizational database into practical work
Combination
C1. Integrating documents from past projects
C2. Systematically recording objective
findings and results for future reference
C3. Codifying standard operating procedures
to direct the work
C4. Integrating knowledge in real time, and
compiling written documents
Tacit knowledge
Explicit knowledgeTacit knowledge
Explicit knowledge
Knowledge-creation practices
for capturing tacit knowledge
Knowledge-creation practices
for capturing Explicit
knowledge
Great direct contribution to
process improvement
Long-term indirect contribution
to other improvement
Figure 3. SPI measurement items based on knowledge creation theory.
Table 2. Sample characteristics.
Basic information of responds N %
Project manager 33 32.4
Requirement
personnel 15 14.7
Designer 19 18.6
Developer 36 35.3
Position
Test personnel 16 15.7
≤10 14 13.7
11 - 20 24 23.5
21 - 50 24 23.5
Membership of
development team
>50 40 39.2
Internal requirement 30 29.4
External client 39 38.2
System developed
Both 33 32.4
Yes 76 74.5
Whether do knowl-
edge management No 26 25.5
Yes 82 80.4
Whether implement
SPI No 20 19.6
Table 3. Basic information analysis.
Indicator Items α
Socializaiton for
capturing tacit
knowledge
S1、S2、S3、S4、
S5、S6、S7 0.729
Externalization for
capturing tacit
knowledge
E1、E2、E3、E4、
E5、E6 0.675
Internalization for
capturing explicit
knowledge
I1、I2、I3、I4、
I5、I6 0.698
Combination for
capturing explicit
knowledge
C1、C2、C3、C4 0.695
Knowledge creation Q1 、Q2 、Q3 、
Q4 0.614

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
340
Table 4. Rotated factor loading matrix of socialization
practices.
Factors Items Loadings
S3 0.784
S2 0.723
Communication
S1 0.646
--
S4 0.812
S6 0.756
Crossover
collaboration
S5
--
0.691
S7 0.874
Work results S8 0.761
Accumulated variance
explained 23.287% 46.471% 64.717%
Table 5. Rotated factor loading matrix of externalization
practices.
Factors Items Loading
E3 0.841
E2 0.785
Requirement
documentation
E1 0.766
E6 0.901
Pair programming E5 0.724
Recording in real
time E4
0.964
Accumulated variance
explained 36.576% 59.048% 77.639%
Table 6. Rotated factor loading matrix of internalization
practices.
Factors Items Loadings
I1 0.849
On-the-job training I2 0.849
I5 0.868
Information utilization I6 0.770
I3 0.781
Document discussion I4 0.780
Accumulated variance explained 25.584% 50.032% 72.806%
Table 7. Rotated factor loading matrix of combination
practices.
Factors Items Loadings
C1 0.872
C2 0.812
Integrating documents
C4 0.584
Procedure standardization C3 0.947
Accumulated variance explained 44.095% 75.734%
Table 8. Rotated factor loading matrix of combination prac-
tices.
Factors Items Loadings
Q3 0.916
Capturing explicit knowledge Q4 0.736
Q2 0.783 Capturing tacit knowledge
Q1 0.723
Accumulated variance explained 35.414% 67.862%
The two factors in Table 8 extract 67.862% of the total
variance. The first factor represents items Q3 and Q4,
and we call it capturing explicit knowledge factor (QF1).
The second factor represents items Q1 and Q2 with a
name capturing tacit knowledge factor (QF2). This is
completely corresponding to the hypothesis noted previ-
ously—knowledge creation practices contribute to SPI
success in both of capturing explicit knowledge and cap-
turing tacit knowledge.
4.4. Regression Analysis
We take socialization practices and externalization prac-
tices as the independent variables in discussing the cau-
sality of capturing tacit knowledge factor. And when
discussing the causality of capturing explicit knowledge
factor, the independent variables are internalization prac-
tices and combination practices. Stepwise regression is
used, including: 1) the regression analyses of practices
and capturing tacit knowledge (Tables 9-11 ); 2) the re-
gression analyses of practices and capturing explicit
knowledge (Tables 12-14).
As to the significance level mentioned in this paper: p*
> 0.05, p** > 0.01.
It can be illustrated in Table 9 that communication
(SF1) and capturing tacit knowledge (QF2) are signifi-
cantly positively correlated (p = 0.01). Crossover col-
laboration (SF2) is also significantly (p = 0.01) positively
correlated with capturing tacit knowledge, with the coef-
ficient being 0.433. Work results (SF3) is significantly (p
= 0.01) positively correlated with capturing tacit knowl-
edge, so is pair programming (EF2) and recording in real
time (EF3).
Thus, the regression is significant with F = 8.258 and
P = 0.005. Meanwhile, variance inflation factor (VIF) is
relatively small, indicating that the multicollinearity is
not significant. Adjusted R2 is 0.309, suggesting that the
3 indicators of knowledge transfer can explain 30.9% of
the total variance. All the variables are significant with P
< 0.05. According to the sequence of independent vari-
ables entering the regression equation, SF2 has the big-
gest influence on dependent variable, with SF3 and SF1
followed (Tables 10 and 11).

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
341
Table 9. Correlation matrix of socialization and externalization practices and capturing tacit knowledge.
Practices
Knowledge
creation
SF1
Communication
SF2
Crossover
collaboration
SF3
Work
results
EF1
Requirements
documentation
EF2
Pair programming
EF3
Recording in
real time
QF2 Capturing tacit
knowledge factor 0.238** 0.433** 0.292**0.178 0.386** 0.257**
Table 10. Effect parameters of stepwise regression.
Model R R2 Adjusted R2 Standard error F Sig.
1 0.423 0.188 0.180 0.906 23.103 0.000
2 0.523 0.273 0.258 0.861 11.629 0.001
3 0.574 0.330 0.309 0.831 8.258 0.005
Table 11. Regression and significance coefficients.
Factors Regression
coefficient
Standard
error
Standard regression
coefficient t Sig. VIF
SF2 0.433 0.083 0.433 5.238 0.000 1.000
SF3 0.292 0.083 0.292 3.533 0.001 1.000
SF1 0.238 0.083 0.238 2.874 0.005 1.000
Standard regression
equation V = 0.433 * SF2 + 0.292 * SF3 + 0.238 * SF1
Table 12. Correlation matrix of internalization and combination practices and capturing explicit knowledge.
Practices
Knowledge
creation
IF1 On-the-job
training
IF2
Information
utilization
IF3 Document
discussion
CF1
Integrating
documents
CF2 Procedure
standardization
QF1Capturing explicit
knowledge 0.338** 0.136 0.286** 0.272** 0.341**
Table 13. Effect parameters of stepwise regression.
Model R R2 Adjusted R2 Standard error F Sig.
1 0.341 0.116 0.107 0.944 13.133 0.000
2 0.436 0.190 0.174 0.909 9.051 0.003
3 0.476 0.226 0.203 0.893 4.604 0.034
Table 14. Regression and significance coefficients.
Factors Regression
coefficient Standard error
Standard
regression
coefficient
t Sig. VIF
CF2 0.289 0.092 0.289 3.135 0.002 1.074
CF1 0.215 0.093 0.215 2.317 0.023 1.090
IF3 0.206 0.096 0.206 2.146 0.034 1.164
Standard regression equation V = 0.289 * CF2 + 0.215 * CF1 + 0.206 * IF3
On-the-job training (IF1) is significantly (p = 0.01) posi-
tively correlated with Capturing explicit knowledge, so is
information utilization (IF2) and Integrating documents
(CF1). Procedure standardization (CF2) is significantly
(p = 0.01) positively correlated with capturing explicit
knowledge, with the correlation coefficient reaching
0.341 (Table 12). F = 4.604 and P = 0.034 suggest sig-
nificant regression effect. And the variance inflation fac-
tor (VIF) is relatively small, indicating that the multicol-
linearity is not significant.
Adjusted R2 is 0.203, suggesting that the 3 indicators
of knowledge transfer can explain 20.3% of the total va-
riance. All the variables are significant with P < 0.05.
According to the sequence of independent variables en-
tering the regression equation, CF2 has the biggest in-
fluence on dependent variable, with CF1 and IF3 fol-

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
342
lowed (Tables 13 and 14).
5. Analyses and Results
Whether factors will be in regression equation depends
on whether it is significant in descriptive statistics and
whether it correlated with corresponding variables. It can
be found that communication factor (SF1) is significant
in descriptive statistics, correlation analysis and regre-
ssion analysis, and communication among project team
members and clients has significant effect on SPI. Cros-
sover collaboration factor (SF2) is significant in descrip-
tive statistics, correlation analysis and regression analysis,
and helping and discussing with each other on work
problems and divergence has significant influence. Inte-
grating documents (CF1) is significant in descriptive
statistics, correlation analysis and regression analysis,
and integrating documents from past projects to analyze
current projects and integrating knowledge in real time and
compiling written documents contribute significantly to
SPI (Table 15). The results can be illustrated in Figure 4.
The full line in the figure represents the biggest contri-
bution, with dashed line the second biggest and the dot-
Table 15. Test and statistic results of relate d hypotheses.
Factors ItemsDescriptive statistics Correlation analysis Regression analysis
S1 √
S2
Communication (SF1)
S3 √
◇ ☆
S4 √
S5
Crossover collaboration (SF2)
S6
◇◇ ☆
S7
Work results (SF3) S8 ◇ ☆
E1 √
E2 √
Requirements documentation (EF1)
E3
E5
Pair programming (EF2) E6 ◇◇
Recording in real time (EF3) E4 ◇
I1 √
On-the-job training (IF1) I2 √ □□
I5
Information utilization (IF2) I6
I3
Document discussion (IF3) I4 □ ☆
C1 √
C2 √
Integrating documents (CF1)
C4
□ ☆
Procedure standardization (CF2) C3 □□ ☆
Q3 √
Capturing explicit knowledge (QF1) Q4. Standard regression equation:
V = 0.289 * CF2 + 0.215 * CF1 + 0.206 * IF3
Q1 √
Capturing tacit knowledge (QF2)
Q2
Standard regression equation:
V = 0.433 * SF2 + 0.292 * SF3 + 0.238 * SF1
√ indicating that descriptive statistic results are significant; ◇ suggests that this factor is correlated with capturing tacit knowledge, and ind◇◇i-
cates stronger correlation; □ suggests this factor is correlated with capturing explicit knowledge，□□ indicates stronger correlation; ☆ suggests that
this factor is in the regression equation.

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project343
Figure 4. Impact of each factor on the success level of SPI.
ted line the least. IF2 is not in the equation, whose insig-
nificant influence is represented with dotted line. Mem-
bers can directly and face to face perceive the intention
of peers, especially the tacit knowledge which is difficult
to express, in communication and collaboration, and this
can seldom be achieved with other approaches. Then,
acquire tacit knowledge that is not codified in text and
files through practical work. Pair programming among
members can facilitate the disclosure of invisible tacit
knowledge in mutual check, and then members absorb
the experience. Making the invisible tacit knowledge
visible has great effect on the success of SPI.
6. Conclusions
The two conclusions can provide guidance to decision-
making in software organizations. Communication among
members, frequent crossover collaboration during the
practical work and integrating related documents have
great contribution to the success of SPI. And the follow-
ing suggestions are proposed: 1) Encouraging personnel
to communicate. Motivate personnel to communicate by
all means, and staff in different levels should commit to
crossover collaboration. Some physical or mental com-
pensation should be offered to drive experienced per-
sonnel to share their knowledge and their knowledge
achievements should be respected, because exchange and
assistance would take time and energy. 2) Integrating
documents in real time. Integrate the information on the
project in real time and generate new logical knowledge
which can be reused and taken for reference. This will be
useful in guiding and enhancing efficiency.
Our main finding is that practices used in SPI projects
to extract team-member knowledge can be quite valuable
for SPI project success. Exploration of additional nuances
of relationships between knowledge-creation approaches
and SPI requires further research. Future investigation
should better understand the state of evolution, or matur-
ity, of a firm’s SPI initiative and the resultant impact on
knowledge-creation practice selection and effectiveness.
In addition, the scales we created for capturing explicit
and tacit knowledge represent a contribution, and can be
used to address questions on the types of projects for
which it would be more beneficial to focus on one type
of knowledge capture over another. The approach illus-
trates that process improvement in general, and SPI in
particular, benefit from perspective and analysis at the
project level of observation. The insights from this study
on the role of knowledge creation in process improve-
ment provide practical guidance for SPI project leaders
and other managers of SPI projects and initiatives, espe-
cially in regards to the importance of capturing tacit
knowledge.
7. Acknowledgements
Thanks for the helpful discussion with Mr. Jianzhang Li
and Mr. Chuanbo Zhang, and my student Qingjing Liu
hard working.
REFERENCES
[1] I. Nonaka, “The Knowledge-Creating Company,” Harvard
Business Review, Vol. 69, No. 6, 1991, pp. 96-104.
[2] M. Polanyi, “The Tacit Dimension,” Doubleday, Garden
City, New York, 1966.
[3] W. S. Humphrey, “Managing the Software Process,”
Addison-Wesley, Reading, 1989.
[4] A. Fuggetta, “Software Process: A Roadmap,” Proceed-
ings of the Conference on the Future of Software Engi-
neering, Limerick, 4-11 June 2000, pp. 25-34.
[5] J. P. Wan and J. M. Yang, “Knowledge Management in
Software Process Improvement,” Application Research of
Computer, Vol. 19, No. 5, 2002, pp. 1-3.
[6] M. Blanco, P. Gutiérrez and G. Satriani, “SPI Patterns:
Learning from Experience,” IEEE Software, Vol. 18, No.
3, 2001, pp. 28-35. doi:10.1109/52.922722

Research on Explicit and Tacit Knowledge Interaction in Software Process Improvement Project
344
[7] J. D. Herbsleb and D. R. Goldenson, “A Systematic Sur-
vey of CMM Experience and Results,” Proceedings of
the18th International Conference on Software Engineer-
ing, Berlin, March 1996, pp. 323-330.
[8] R. S. Pressman, “Software Engineering: A Practitioner’s
Approach,” 5th Edition, McGraw-Hill Companies, Inc.,
New York, 2001, p. 19.
[9] J. P. Wan, “Research on Software Product Support Struc-
ture,” Journal of Software Engineering and Applications,
Vol. 2, No. 3, 2009, pp. 174-194.
doi:10.4236/jsea.2009.23025
[10] G. Anand, P. T. Ward and M. V. Tatikonda, “Role of
Explicit and Tacit Knowledge in Six Sigma Projects: An
Empirical Examination of Differential Project Success,”
Journal of Operations Management, Vol. 28, No. 4, 2010,
pp. 303-315. doi:10.1016/j.jom.2009.10.003
[11] I. Nonaka and G. V. Krogh, “Perspective—Tacit Knowl-
edge and Knowledge Conversion: Controversy and Ad-
vancement in Organizational Knowledge Creation The-
ory,” Organization Science, Vol. 20, No. 3, 2009, pp.
635-652. doi:10.1287/orsc.1080.0412
[12] J. P. Wan, Q. J. Liu, D. J. Li and H. B. Xu, “Research on
Knowledge Transfer Influencing Factors in Software
Process Improvement,” Journal of Software Engineering
and Applications, Vol. 3, No. 2, 2010, pp. 134-140.
doi:10.4236/jsea.2010.32017
[13] J. P. Wan and R. Wang, “The Exploratory Analysis on
Knowledge Creation Effective Factors in Software Re-
quirement Development,” Journal of Software Engineering
and Applications, Vol. 3, No. 6, 2010, pp. 580-587.
[14] J. P. Wan, H. Zhang, D. Wan and D. Y. Huang, “Re-
search on Knowledge Creation in Software Requirement
Development,” Journal of Software Engineering and Ap-
plications, Vol. 3, No. 5, 2010, pp. 487-494.
doi:10.4236/jsea.2010.35055
[15] J. P. Wan and R. Wang, “Empirical Research on Critical
Success Factors of Agile Software Process Improve-
ment,” Journal of Software Engineering and Applications,
Vol. 3, No. 12, 2010, pp. 1131-1140.
doi:10.4236/jsea.2010.312132
[16] R. J. Jensen and G. Szulanski, “Template Use and the
Effectiveness of Knowledge Transfer,” Management Sci-
ence, Vol. 53, No. 11, 2007, pp. 1716-1730.
doi:10.1287/mnsc.1070.0740