Convention Center — Ballroom B-C
Session A
Panel: Is “Software Engineering” the Wrong Metaphor? And Why Should We Care?
Moderator: Alistair Cockburn, Humans and Technology
In an industry that is supposed to be driving the future, we see many recurring patterns of dysfunction and self-destructive behavior in the planning and execution of software development projects. Many have questioned typical practices of hiring, educating, and managing software-related professionals. Many of these problems seem to be rampant. Dilbert and many other parodies illustrate the recurring symptoms, but do they really identify the problem?
The response to undisciplined software development has been to search for a metaphor that helps the world envision a more disciplined approach to software that will produce reliable software at a predictable cost. It seems that “software engineering” is the metaphor that has stuck. Some argue that the plethora of problems we face in our industry stem from the use of the “software engineering” metaphor. Others argue that the problems stem from the misapplication of the metaphor.
Is it time for a new metaphor?
This panel explores the positive and negative roles the “software engineering metaphor” plays in various types of software development efforts and discusses the practicalities of other metaphors as a cure to the problems we face. The panelists offer widely diversified perspectives on the topic.
Panelists:
Pete McBreen, McBreen Consulting, Canada
Ken Auer, RoleModel Software, Inc.
Eugene Wallingford, University of Northern Iowa
Tom Ball, Microsoft Research
David Smith, Timeline Computer Entertainment
Convention Center — Ballroom A
Session B
Papers: Garbage Collection
Chair: Satoshi Matsuoka, Tokyo Institute of Technology
Automated storage management in the form of garbage collection (GC) is becoming an integral part of object-oriented system implementation. Indeed, superior GC technology improves robustness, speeds execution, and makes better use of available memory. The first paper here shows that pretenuring (i.e., selectively bypassing generational collection by directly allocating objects in the “old” space) is effective when guided by profiling information. The second paper reexamines the classic heap exhaustion problem and its customary solutions—whether to grow the heap or to garbage-collect—in a modern Java setting, demonstrating a strategy that results in surprisingly efficient execution. The third paper presents a novel algorithm for reference counting in a multithreaded environment, virtually eliminating expensive inter-thread atomic operations through clever per-thread bookkeeping of pointer updates.
Pretenuring for Java
Stephen M. Blackburn, University of Massachusetts
Sharad Singhai, University of Massachusetts
Matthew Hertz, University of Massachusetts
Kathryn S. McKinley, University of Massachusetts
J. Eliot Moss, University of Massachusetts
Controlling Garbage Collection and Heap Growth to Reduce the Execution Time of Java Applications
Tim Brecht, University of Waterloo and HP Labs
Eshrat Arjomandi, York University
Chang Li, York University
Hang Pham, York University
An On-the-Fly Reference Counting Garbage Collector for Java
Yossi Levanoni, Microsoft
Erez Petrank, Technion |
