This paper introduces a family of patterns that can be used to shape a new organization and its development processes. Patterns support emerging techniques in the software design community, where they are finding a new home as a way of understanding and creating computer programs. There is an increasing awareness that new program structuring techniques must be supported by suitable management techniques, and by appropriate organization structures; organizational patterns are one powerful way to capture these.
We believe that patterns are particularly suitable to organizational construction and evolution. Patterns form the basis of much of modern cultural anthropology: a culture is defined by its patterns of relationships. Also, while the works of Christopher Alexander [Alexander] deal with town planning and building architecture to support human enterprise and interaction, it can be said that organization is the modern analogue to architecture in contemporary professional organizations. Organizational patterns have a first-order effect on the ability of people to carry on. We believe that the physical architecture of the buildings supporting such work are the dual of the organizational patterns; these two worlds cross in the work of Thomas Allen at MIT [Allen].
There is nothing new in taking a pattern perspective to organizational analysis. What is novel about the work here is its attempt to use patterns in a generative way. All architecture fundamentally concerns itself with control [Carlin]; here we use architecture to supplant process as the (indirect) means to controlling people in an organization. Not only should patterns help us understand existing organizations, but they should help us build new ones. A good set of organizational patterns helps to (indirectly) generate the right process: this indirectness is the essence of Alexandrine generativity. In fact, organizational patterns might be the most generative approach to software architecture patterns. Alexander notes that architectures "can't be made, but only generated, indirectly, by the ordinary actions of the people, just as a flower cannot be made, but only generated from the seed." [Alexander]; A set of simple patterns together cause complex emergent behaviour. As with many of the principles in Timeless Way, this is curiously reminiscent of the Way of Non-Action (Yin) of the Tao Teh Ching; we also find it in the triple-loop thinking of Swieringa and Wierdsma's organizational learning model [Swieringa and Wierdsma].
At this writing, the work is speculative: only limited use has been made of these patterns in formulating new organizations. The "goodness" or "badness" of such patterns is difficult to test by experiment. First, any metric of organizational goodness is necessarily multidimensional and complex. Second, it is difficult to do large-scale social experiments with tight enough control variables that the effectiveness of a pattern could be verified. Third, such an experiment would take a long-term commitment (months or years), more than most software organizations are willing to expend in light of fragile and evolving markets.
For these reasons, the patterns fall back on case studies and on common sense. We look at recurring patterns of interaction in organizations, note recurring patterns between those patterns and some measures of "goodness," and then do analysis to explain the correlation. The patterns presented here all combine empirical observations with a rationale that explains them. The claim that the language as a whole captures essential characteristics of high-productivity organizations has been validated by the CEOs of many small, highly productive organizations who have read these patterns. [Gabriel].
The patterns meet other "standard" criteria emerging in the patterns movement. Each is stated as a problem or opportunity. Each is analyzed for the forces at play within it. Each instructs us to do something explicit. This form follows the pattern work of Alexander, whose books on architecture serve as a model for the modern software patterns movement. Some of these organizational patterns hark back to Alexander; where appropriate, a distinctive reference appears in the text of the form [Alexander pattern-number] to refer the reader to a pattern in Alexander's A Pattern Language [Alexander].
Though there are organizations that exhibit these patterns, combining them into a new organization, built from scratch, is a daunting task. The ideal organization envisioned by these patterns differs greatly from the state of the art in software development. These patterns are drawn from peculiar organizations with peculiarly high productivity. The patterns describe practices much different from those found in most project management texts.
This contains more introductory material on the language context.
This contains more introductory material on the language forces.
Using this language does not guarantee the Quality without a name.
This contains more introductory material on the language rationale.
This talks about notations used in the document.
Lots of great people contributed to this document--credits can be found here.
Next: Size the Organization