The aim of the ICSE 97 panels is to stimulate discussion about ideas and issues of crucial importance to the software engineering community. Panels may emphasize complex technical issues, but they may equally provide an opportunity to propose views and address controversies through the medium of informed debate. Suitable topics for such discussion include pressing issues in software engineering theory and practice, emerging industry trends and enabling technologies, and professional, organizational, and social issues associated with software engineering.
A panel is an interactive discussion. It is essential for the moderator and panelists to design it in advance, for them to prepare fully, and for the moderator to exercise control over the discussion. Too many promising panels turn out to be dreary paper sessions in disguise. We want the panels at ICSE 97 to stick (positively) in people's minds.
Panels last 90 minutes and can be organized in many formats. Several suggestions are provided below. However, we strongly encourage proposals in original formats that will engage the panelists and audience in a lively and substantive discussion.
An analytic panel analyzes and synthesizes current practices in the various fields of software engineering. The panelists aim at introducing the audience to new ideas and should therefore be technical experts selected from distinct fields but addressing a single problem. Panelists should address the following questions:
The biggest challenge in an analytic panel is to ensure it does not deteriorate into a paper session. It is therefore essential that panelists' ideas are explicitly interrelated and synthesized.
In a comparative panel, panelists compare distinct approaches, techniques and models to a particular problem. Panelists should explain why their approach is better than everyone else's (and not merely assert that it is). The panel should aim at identifying the common philosophy among the approaches represented in the panel, and on what crucial distinctions (as opposed to accidental details) their differences depend. A comparative panel should be controversial, but the panelists should treat each other's approaches with respect. Panelists should be technical experts within the same field.
The biggest challenge with a comparative panel is to ensure that the panelists understand and respect each other's approaches and do not see the panel as an opportunity for a sales job. One technique is to have each panelist present "someone else's" approach as positively as possible.
In an exercise, the aim is to have panelists solve a small but representative problem, using their chosen approach and with little preparation. In suitable circumstances, the problem scenario could be revealed to the panelists at the start of the session itself. The problem scenario will be selected in advance by the moderator of the session.
The biggest challenges with an exercise are to ensure that the panelists are prepared and that the audience goes away seeing the forest, not just the trees. For an excercise panel to achieve any closure, it is essential that the moderator be able to summarize at the end of the session the key similarities and differences among the panelists' approaches.
In a predictive panel, panelists from different fields discuss the ways in which emerging techniques and technologies will affect the practice of software engineering in the future. Panelists should be experts in their relevant fields and should have clear ideas about how work in their fields is influencing the future.
The challenge with a predictive panel is to avoid idle speculation and waffling. The key idea is careful selection of the topic and panelists.
Panels may address any topic of relevance to software engineering research and practice. Panels are an opportunity to stimulate thinking about how software engineering is evolving. This means that topics need not be limited to the conventional boundaries of the field.
Each proposal will be reviewed by the members of the panel committee, who will consist of academics and practitioners. We are looking for stimulating and timely proposals that will be debated by well-informed and engaging panelists and that will form a diverse, controversial, and well thought out collection. Please feel free to contact either panel co-chair in advance to discuss your proposal.
Our criteria will include:
In summary, we will ask ourselves whether the average audience member will go away from the panel with new and valuable ideas.
A proposal should contain the following:
A summary of each panel will appear in the conference proceedings. The panel organizer is required to collate the half-page position papers (or later revisions of them), together with an introduction into a summary description of the panel. Panelists must agree to the half-page position papers being published in the conference proceedings.
Authors will be notified of acceptance or rejection by the end of November 1996.
Summaries of accepted panels will be published on the World Wide Web in the pre-conference publicity. Panel organizers will be sent an author kit for preparing camera-ready or electronic materials ready for publication. These materials are due 24 Feburary 1997.
Panel organizers are expected to help panelists prepare for participation and coordinate the contributions of the panel.
Please follow the steps in this checklist to ensure completeness of your submission.
College of Computing
Georgia Institute of Technology
801 Atlantic Drive
Atlanta, GA 30332-0280, USA
20 Maguire Road
Lexington, MA 02173, USA