process representations are essential to allow for better communication about the processes, thus serving as a reference throughout measurement and change. However, there is a considerable lack of success stories about software process technology being used in industrial improvement programs. The reasons for this lack might be a mixture of the following: • Software process technology is used in industry less widely than expected and consequently it still needs to be transferred into industrial practice. • Software process technology is used “behind the curtain” which means that no value is seen in distributing messages about their usage in industrial projects. • Software process technology is substituted by approaches from other domains (e.g., business process engineering) or just home made applications. • Software process technology does not meet process engineers’ needs because it was developed in purely academic environments, apart from industrial experience and people directly involved in the processes. • Software process technology as a field is still in an immature state where powerful and widespread tools can be realized as soon as basic fundamentals are detected and manifested into components (e.g., frameworks for user agendas, or standard product database systems). MVP-E can be seen as an example process modeling environment for building large industrial software development processes. The use of MVP-L in industrial projects has helped to build a system that supports the process engineer in improvement programs. We see MVP-E as considerably different to many other system builds in the area of software process technology because it does not focus on processes described on a fine-grain level (i.e., process implementations) but allows for the management of process models on a coarse-grain level (i.e., process design). What is remarkable about MVP-E is not a single feature solely but the aggregation of the following issues: • MVP-E is a highly integrated environment which allows tools focusing on multiple aspects of software process technology to work together. From descriptive modeling to simulation and enactment many aspects of project support are covered. This tight integration demonstrates the feasibility of comprehensive support for process engineers. • MVP-E was developed especially for software process modeling. The tools share some knowledge about these processes which makes them more powerful than tools adapted for that purpose. • MVP-E was validated in industrial improvement programs. Although still a prototype it has demonstrated its suitability and usefulness for the process engineer. Incremental builds helped to recognize feedback from projects which drove the development in those parts of the system that were most promising. For example, a graphical interface was developed late in the project whereas checks that were hard to perform manually were automated first. • MVP-E addresses the need for specifying measurable properties of software development processes and products. Language constructs allow for an adequate representation of relationships among levels of refinement. One can define control flow by using entry and exit criteria but the primary reason for these constructs is the definition of valid states for process performance.
[1]
Martin Verlage,et al.
Tailoring Large Process Models on the Basis of MVP-L
,
1997
.
[2]
James A. Whittaker,et al.
Software Technology for Adaptable, Reliable Systems (STARS) Program
,
1991
.
[3]
W. H. Davidow,et al.
The Virtual Corporation: Structuring and Revitalizing the Corporation for the 21st Century
,
1992
.
[4]
Nazim H. Madhavji,et al.
A Systematic, View-Based Approach to Eliciting Process Models
,
1996,
EWSPT.
[5]
Martin Verlage,et al.
A Synthesis of Two Process Support Approaches
,
1996,
SEKE.
[6]
Victor R. Basili,et al.
A Methodology for Collecting Valid Software Engineering Data
,
1984,
IEEE Transactions on Software Engineering.
[7]
M. Marquardt,et al.
The global learning organization
,
1993
.
[8]
Alfred Bröckers,et al.
The role of software process modeling in planning industrial measurement programs
,
1996,
Proceedings of the 3rd International Software Metrics Symposium.
[9]
M. Verlage,et al.
Formalizing software engineering standards
,
1997,
Proceedings of IEEE International Symposium on Software Engineering Standards.
[10]
Øystein Haugen,et al.
Engineering real time systems - an object-oriented methodology using SDL (2. pr.)
,
1993,
BCS practitioner series.
[11]
Martin Verlage.
Ein Ansatz zur Modellierung großer Software-Entwicklungsprozesse durch Integration unabhängig erfaßter rollenspezifischer Sichten
,
1998
.
[12]
C.D. Klingler,et al.
A case study in process representation using MVP-L
,
1992,
COMPASS `92 Proceedings of the Seventh Annual Conference on Computer Assurance.
[13]
Martin Verlage,et al.
MVP-L Language Report Version 2
,
1995
.
[14]
Alfred Bröckers.
Modellbasierte Analyse von Software-Projektrisiken
,
1997
.
[15]
R. E. Miles,et al.
Managing 21st century network organizations
,
1992
.
[16]
Christopher M. Lott,et al.
Measurement-based feedback in a process-centered software engineering environment
,
1996,
Interner Bericht / Universität Kaiserslautern / Fachbereich Informatik.