Automated Software Engineering - ASE'97
12th IEEE International Conference
Panel 1 - Large Scale Software Transformation: Lessons Learned from the Year 2000 Problem
Tuesday November 4, 8:30 - 9:30
The Year 2000 problem has posed significant challenges that are being
addressed using interesting automated software engineering (ASE)
tools, techniques, and methodologies. This panel will attempt to
relate and extrapolate the tools, techniques, and methodologies to
those for other large-scale software transformation problems.
The focus will be on comparing and contrasting the Year 2000 problem
with other interesting classes of problems and utilizing the lessons
learned from the Year 2000 problem for motivating further development
in ASE techniques applicable for the various phases of software
transformation lifecycles. The panel will attempt to leverage the
unprecedented scale required for solving the Year 2000 problem towards
the goal of scalable automated techniques with well-defined
characteristics.
The preliminary focus of the panel will be on tools and
techniques. Process and methodology issues may be considered in the
context of tools and techniques or separately if time permits and the
audience demonstrates a preference for them. The list of questions to
be addressed includes but is not limited to:
- Problem Assessment and Estimation
- What are the other interesting classes of problems, and
their similarities and differences, such as: currency changes, field
(e.g. part number) expansions, and legacy system transformation.
- What are the characteristics of the impact assessment
techniques used in solving the Year 2000 problem? What are the
tradeoffs involved?
- How could the techniques used for partitioning and
clustering "work units" of a transformation project be applied to
other transformation projects?
- Detailed Analysis and Transformation
- What are the roles of different analysis approaches? What are
the tradeoffs involved in applying syntactic, semantic, and slicing
techniques?
- What are the issues involved in separating analysis from
transformation, and in comparing user-assisted versus automated
transformation?
- Testing and Verification
- What are the key tools and techniques required for planning
and designing large scale tests? How could we quantify and evaluate
the tradeoff between testing intensity and correctness?
- Are there any interesting techniques for automated test-bed
building that can be generalized?
- Common Techniques
- What are the key features of repositories that assist in
various parts of the lifecycle? What are the consequences of store
versus recompute decisions on scale, consistency and usability.
- Are there any canonical representations for software
components that are critical for one or more phases?
- What are the key lessons learned from integrating analysis
tools, repositories, code transformation tools and testing tools?