Abstract:
Emerging technologies such as the Internet, the World Wide Web,
JavaTM
technology, and software components are accelerating product life
cycles and encouraging collaboration across organizational
boundaries. The familiar coordination problems of large scale software
development reappear in a context where tools used by collaborators
must be less tightly coupled to one another than before. To the
traditional notion of scale, based on the size of software systems,
must be added a new dimension of scale: organizational
complexity. Designing configuration management systems that scale well
over both dimensions requires difficult trade-offs between reliability
and flexibility. At the heart of these trade-offs is the aggregate
information shared by collaborators: how it is represented,
maintained, and understood by the people and tools using it. While
designing a prototype development environment intended to scale in
both dimensions, we have revisited the role played by naming. A
proposed extension to the prototype's naming system addresses issues
such as which objects should be named and how the shared naming system
is constructed.
Ninth International Symposium on System Configuration Management (SCM-9) , Toulouse, France, September 5-7, 1999.
Published in Springer Lecture Notes in Computer Science, 15 pages (PDF)