Requirements model evolution throughout the system lifecycle
RE (2008)
- DOI: 10.1109/RE.2008.11
Available from
Neil Ernst's profile on Mendeley.
or
Abstract
Requirements models are essential not just during system implementation, but also to manage system changes post-implementation. Such models should be supported by a requirements model management framework that allows users to create, manage and evolve models of domains, requirements, code and other design-time artifacts along with traceability links between their elements. We propose a comprehensive framework which delineates the operations and elements necessary, and then describe a tool implementation which supports versioning goal models.
Available from
Neil Ernst's profile on Mendeley.
Page 1
Requirements model evolution throughout the system lifecycle
Future work
● Work on traceability using incremental Latent
Semantic Indexing
● Examine model management ADT and less
formal requirements practices
● Implement slice/query operators
Implementation
● OpenOME is an extensible implementation
of most of the requirements model
management framework
● Integrated an Eclipse-based goal modeling
tool with Molhado, an object-based
configuration management backend
Requirements model evolution throughout the system lifecycle
Neil A. Ernst1, John Mylopoulos1, Yijun Yu2, Tien Nguyen3
1University of Toronto, 2Open University, 3Iowa State University
For more information:
Neil A. Ernst
Dept. of Computer Science
University of Toronto, Canada
nernst@cs.toronto.edu
The Problem
• Requirements change post-implementation
• Little research addresses this issue
• Requirements changes are the source for
system maintenance tasks (e.g., change
requests)
• Requirements models should persist post-
implementation just like code
Requirements model
management framework
● Define requirements and related models using
an abstract data type (ADT)
● Supports configuration of requirements
models at multiple points in time
● Links requirements and implementation
● Define elements: G – goal models, I –
implementation models, T – traceability links
between them
● Define operators on the models:
● Generic model management:
● merge, match (split), trace
● Model queries:
● slice, temporal query
● Configuration management:
● diff, version, temporal commit
● Goal model specific:
● evaluate
Evaluation
● Evaluated tool on three versions of Trac1, a
project management tool, by reverse
engineering the source code to goal models.
Scalability:
● Algorithms: O(N) or O(logN), N = # elements
● Changesets smaller than diff or XML diff
Minimality
● OpenOME works on artifacts in the problem
domain, not the model syntax, e.g.:
● From V1 to V2, OpenOME shows 871 added
and 339 deleted artifacts; whereas
● From V1 to V2, diff shows 3375 lines added
and 2309 lines deleted
Conformance
● Tests demonstrated OpenOME preserved
model elements throughout the process
● Change rationale recording algorithm
preserved entire edit chain from V1 to V2
1 trac.edgewall.org
The GMF editor of OpenOME, showing
the creation of an i* model
● Work on traceability using incremental Latent
Semantic Indexing
● Examine model management ADT and less
formal requirements practices
● Implement slice/query operators
Implementation
● OpenOME is an extensible implementation
of most of the requirements model
management framework
● Integrated an Eclipse-based goal modeling
tool with Molhado, an object-based
configuration management backend
Requirements model evolution throughout the system lifecycle
Neil A. Ernst1, John Mylopoulos1, Yijun Yu2, Tien Nguyen3
1University of Toronto, 2Open University, 3Iowa State University
For more information:
Neil A. Ernst
Dept. of Computer Science
University of Toronto, Canada
nernst@cs.toronto.edu
The Problem
• Requirements change post-implementation
• Little research addresses this issue
• Requirements changes are the source for
system maintenance tasks (e.g., change
requests)
• Requirements models should persist post-
implementation just like code
Requirements model
management framework
● Define requirements and related models using
an abstract data type (ADT)
● Supports configuration of requirements
models at multiple points in time
● Links requirements and implementation
● Define elements: G – goal models, I –
implementation models, T – traceability links
between them
● Define operators on the models:
● Generic model management:
● merge, match (split), trace
● Model queries:
● slice, temporal query
● Configuration management:
● diff, version, temporal commit
● Goal model specific:
● evaluate
Evaluation
● Evaluated tool on three versions of Trac1, a
project management tool, by reverse
engineering the source code to goal models.
Scalability:
● Algorithms: O(N) or O(logN), N = # elements
● Changesets smaller than diff or XML diff
Minimality
● OpenOME works on artifacts in the problem
domain, not the model syntax, e.g.:
● From V1 to V2, OpenOME shows 871 added
and 339 deleted artifacts; whereas
● From V1 to V2, diff shows 3375 lines added
and 2309 lines deleted
Conformance
● Tests demonstrated OpenOME preserved
model elements throughout the process
● Change rationale recording algorithm
preserved entire edit chain from V1 to V2
1 trac.edgewall.org
The GMF editor of OpenOME, showing
the creation of an i* model
Sign up today - FREE
Mendeley saves you time finding and organizing research. Learn more
- All your research in one place
- Add and import papers easily
- Access it anywhere, anytime
Start using Mendeley in seconds!
Readership Statistics
1 Reader on Mendeley
by Discipline
by Academic Status
100% Ph.D. Student
by Country
100% Canada