10th International Software
Product Line Conference
(SPLC 2006)
21-24 August 2006
Baltimore,
Maryland, USA
Conference Panels
|
23 August 2006 |
|
|
|
Product Derivation Approaches |
|
P2 |
Testing in a Software Product Line |
|
|
|
|
24 August 2006 |
|
|
P3 |
Product Line Research |
Panel 1 (P1)
Panel on Product
Derivation Approaches
23 August 2006
Panel moderator: David Weiss, Avaya Labs
Panelists:
Danilo Beuche,
pure-systems
Charles Krueger, BigLever Software
Rob van Ommering,
Philips Research
Juha-Pekka Tolvanen, MetaCase
Abstract
This panel looks at
product derivation approaches and their differences, strengths and weaknesses
in different PLE situations. Each panelist will examine a common problem (the
Interactive Television Applications)
and provide an overview of their product derivation approach and how it was
used to solve the problem.
Overview
At some point, no
matter how wonderful your product line process is, you have to ship the
products. The panelists will each present a different approach to PLE,
concentrating on how actual products are derived from specifications. The
approaches presented include feature modeling, architecture description
languages, UML and domain-specific modeling languages.
A common product specification and derivation task will be given to all panelists, and they will show how their approach works on it. The audience can - and is warmly encouraged to - participate, ask additional questions, heckle, and hopefully laugh. A major goal is to identify the classes of PLE situations that best suit each approach.
Following are some of the questions and issues to be addressed by the panel.
- How large a portion of a product is automatically derived? Please answer in terms of some reasonably precise measure, such as percent of modules, classes, or KNCSL, or coverage in a feature model.
- How are new features and functionality developed? Give an example, if possible.
- What is the cost and time to create a new feature or change the application platform, e.g., in hours of effort as a fraction of effort needed to create the application engineering environment? Alternatively, how would you estimate the cost and time?
Panel 2 (P2)
Testing in a Software Product
Line
23 August 2006
Panel moderator: Klaus Pohl, Lero, The Irish Software Engineering Research Centre & University of Duisburg-Essen, Germany
Panelists:
Georg Grütter,
Robert Bosch GmbH, Germany
John D. McGregor, Clemson University, USA
Andreas Metzger, University of Duisburg-Essen, Germany
Tim Trew,
Philips Research, The Netherlands
Abstract
This panel is about
system testing of software product line artifacts. The panelist will present
different approaches for software product line testing. Together, we will
discuss their pros and cons. As a kind of benchmark, a common example of an
online store (The eShop Product Line) will be
used to ease the comparison of the different testing approaches.
Overview
Each panelist will
present an approach to test the domain and application artifacts in software
product line engineering. The decision whether to test the domain artifacts in
domain engineering or if testing is delayed to application engineering is left
to the panelists.
To facilitate a better comparison of the different test approaches, each panelist will illustrate his approach using a running example of an online store product line.
The discussions will, among others, cover the following questions:
- Should there be system testing in domain engineering, or should system tests be performed during application engineering only?
- Which test artifacts can be reused during product line testing?
- Is there an advantage of creating domain test artifacts which are reused during application engineering?
- Can application test cases be generated? And if so, should they be generated from domain test cases or just from application engineering artifacts?
- Does the model-based test case derivation offer benefits when compared with deriving test cases directly from natural language requirements?
Panel 3 (P3)
Product Line Research: Lessons
Learned from the last 10 years and Directions for the next 10
24 August
2006
Moderator: Liam O'Brien, Lero, The Irish Software Engineering Research Centre
Panelists:
Paul Clements,
Software Engineering Institute, USA
Kyo Kang, POSTECH, Korea
Dirk
Muthig, Fraunhofer IESE, Germany
Klaus Pohl, Lero, The Irish Software
Engineering Research Centre & University of Duisburg-Essen, Germany
Abstract
This panel is about past
and future research in software product lines. The panelists will look back at
the past 10 years to examine outcomes and lessons learned and will look forward
to the next 10 years and will give potential outcomes and directions for the
future of software product line research. The outcomes will be examined for
relevance to the practitioner community.
Overview
Each panelist will
present their lessons learned from the past and the directions for the next 10
years. Several industry judges will be asked to make a determination as to how
useful the outcomes have been or will be for practitioners. The panelists will
have an opportunity to respond to the judges comments and this will lead to a
general discussion.
The discussions will, among others, cover the following questions:
- What have been the main lessons and outcomes for research in software product lines over the past 10 years?
- What will likely be the major directions and outcomes for research in software product lines over the next 10 years?
- How relevant have the past outcomes been for practitioners and how relevant will future outcomes be for practitioners?
- What will be the next big breakthrough for software product line research?
