Co-leaders

* Steve Berczuk, Corechange, Inc.
* Ralph Cabrera, AG Communication Systems, Inc.

Goal and Accomplishments

The goal of this workshop is to develop a pattern language for software configuration management (SCM) practice.

By software configuration management we mean the processes that are involved in making the development process proceed smoothly, including, but not limited to: source control, requirements tracking, and change management.

We will attempt to develop a guide for implementing good configuration management practices, which balance the forces imposed by:

* development process(es)
* social influences
* organizational structures

We will build on the small body of published patterns on SCM and on process, and use participants' experiences to try to connect and refine these patterns into something like a coherent whole. The goal is to integrate with existing patterns and build a language rather than a collection of freestanding patterns.

Issues we would like to address:

* How the SCM process can support piecemeal growth of a software system
* Difference in perspective between SCM implementers and SCM clients


(developers, managers, reuse people, etc.)


* Piecemeal growth of an SCM process

Our approach will be to:

* Discuss existing SCM patterns literature.
* Discuss SCM issues, and participants' thoughts on patterns.
* Develop a "Map" of what the patterns should look like (patlets).
* Document participants' common experiences as patterns.

Workshop Activities

The workshop leaders will help develop the SCM experiences into patterns and, eventually, a pattern language for SCM. During the workshop we will develop a map of patterns that identify patlets (rough-cut patterns) that can be workshopped during the conference. Further work on both the patterns and the language will be continued on the Wikki web or the organizational patterns wikki site.

Ideally, the patterns we develop should be supportable by something approaching empirical evidence, rather than instinct. Since there is often a connection between things that feel right and consistent solutions, where we can't obtain empirical evidence for the patterns, we'll discuss how we can discover the evidence.

Where possible we will connect our patterns with organizational patterns and (possibly) architectural patterns to show how these reflect the design of a SCM system. We will emphasize piecemeal growth (how we can incrementally add to an existing system), since we rarely start from scratch.